DGMPGDecNV 1.0.2

[neuron2]

WARNING: This is not for you if you do not have a supported Nvidia display adapter!

Here's Nvidia CUDA support for MPEG2 streams.

Please read the Release Notes for current limitations.

http://neuron2.net/dgmpgdecnv/dgmpgdecnv.html

Your feedback will be appreciated.

[laserfan]

Your feedback will be appreciated.First off, THANKS!

More to come... ;)

[neuron2]

Please re-download. I just fixed two bugs in DGMPGIndexNV. :)

[laserfan]

I re-downloaded and made a .dgm file from an MPEG2 Elementary Stream, but when I try to open my .avs file

LoadPlugin("c:\program files (x86)\DGMPGDecNV\DGMPGDecodeNV.dll")
MPEG2Source("d:\video\video.dgm")

I get an Avisynth error when trying to launch with VirtualDub:

Avisynth open failure:
MPEG2Source: The input file is not a D2V project file.

Tried changing the .dgm to .d2v but get the same error. Am I right to be using MPEG2Source?

EDIT: Never mind, on a hunch I changed MPEG2Source to MPGSource and it worked! :)

[neuron2]

EDIT: Never mind, on a hunch I changed MPEG2Source to MPGSource and it worked! :) You smart guy. I bow down.

Will add that to the Notes.txt file. Thank you for pointing it out.

[kebulek]

:thanks:

I tried one test encode with DVB stream (16:9 720x576i@3.5mbit) and this avs script:

MPGSource("D:\test.dgm", deinterlace=1).Crop(8,4,-8,-4).LanczosResize(624,352)

vs

MPEG2Source("D:\test.d2v").LeakKernelDeint(order=1).Crop(8,4,-8,-4).LanczosResize(624,352)

85-87fps vs 104-106fps, which is not bad. Can we expect some speedup in near future? :)

[neuron2]

What numbers do you get without deinterlacing? Don't forget, the PureVideo deinterlacer is a heavy load. It will be interesting to see if it is faster than similar Avisynth filters, such as yadif. LeakKernelDeint() is very simple by comparison. PureVideo produces much better quality deinterlacing.

[kebulek]

@neuron2

Without deinterlacing: 117-124fps vs 117-130. I expected a lot more without it. :confused:

It's XviD 1.2 1-pass fyi...

[neuron2]

Without deinterlacing: 117-124fps vs 117-130. I expected a lot more without it. You've not been reading the threads about this.

Did you compare PureVideo to yadif?

[kebulek]

I'm not familiar with yadif, but I'm learning how to use it right now. :)

[woah!]

when using the deinterlace=1 option, what are the purevideo settings used for the deinterlacing? or is what you get the best it can deliver for the type of material used?

so far it is working fine here :)

[neuron2]

I don't know the answer to that. All I know is I select the mode "Adaptive".

[woah!]

ok i seem to have missed something with purevideo then. how or where do you pick adaptive? i can see "smart/film/video/automatic" in this window and only "best available/combine fields/display fields separately"?

http://thumbnails11.imagebam.com/2615/4c882f26146564.gif (http://www.imagebam.com/image/4c882f26146564)

[neuron2]

I'm talking about the CUDA video API. It offers:

Weave
Bob
Adaptive

I guess that they correspond to:

Combine fields
Display fields separately
Best available

[woah!]

ah ok sorry to go off topic, if thats the case then it is doing the best it can with adaptive i suppose :)

[squid_80]

Since the source code hasn't been released I assume this program has been completely rewritten from scratch, even though the interface is very similar to the existing GPL licensed dvd2avi/dgmpgdec programs?

[jj666]

Thanks Neuron2, you're making my measly $15 donation stretch very far with the addition of VC1 and MPEG2 :-)

Cheers,

-jj-

[hajj_3]

is there any reason why dgmpgdec doesnt have .mpg in the allowed extensions to open by default? as lots of tv card software like hauppauge record .mpg mpeg2 files. i get the error "program streams not supported" when i enable all extensions to be opened and then try to open a .mpg file.

[kebulek]

MPEG2Source("D:\test.d2v").Yadif(mode=0,order=1).Crop(8,4,-8,-4).LanczosResize(624,352)
vs
MPGSource("D:\test.dgm", deinterlace=1).Crop(8,4,-8,-4).LanczosResize(624,352)

XviD 1-pass: 78-79fps vs 100-105fps (C2D E6600 + 9600GT)

I hope I used right Yadif settings. :)

[rack04]

Please read the file Notes.txt for current limitations.


I get the error "program streams not supported" when i enable all extensions to be opened and then try to open a .mpg file.

From the Notes.txt file.

Current limitations
-------------------

No MPEG1.

No program streams (e.g. .vob and .mpg).

No honoring of pulldown.

No Fusion-style audio.

[neuron2]

is there any reason why dgmpgdec doesnt have .mpg in the allowed extensions to open by default? Because I haven't implemented program stream support yet.

[nixo]

Hi,

Thanks for this. I've had no issues with dgmpgdecnv and you're right, the deinterlacer looks really good.

Initially though, I had problems with audio desynch. Turns out when demuxing with DGIndex I kept losing the first 10 frames of the vob. When using a different demuxer everything was fine. I'm using latest 1.5.3 so maybe there's still an issue with demuxing - I saw there was a recent entry in the changelog regarding this.

Here is a sample that illustrates the problem:

http://www.megaupload.com/?d=T3JSIH9Q

This probably belongs in the dgmpgdec thread but I figured that since people need to demux to use dgmpgdecnv... well it might be allright.

--
Nikolaj

[neuron2]

I figured that since people need to demux to use dgmpgdecnv That's needed only if you have a program stream for the moment. I'll add PS support.

I'll look at your stream. Thanks for pointing this out.

[hajj_3]

Because I haven't implemented program stream support yet.

ahh ok.

if i was to use purevideo to de-interlace would this mean that when i encode in gordian knot i wont have to choose "field de-interlace (blend)" and instead would choose not to de-interlace at all?

many thanks neuron.

[neuron2]

if i was to use purevideo to de-interlace would this mean that when i encode in gordian knot i wont have to choose "field de-interlace (blend)" and instead would choose not to de-interlace at all? That is correct.

[halsboss]

Er, can anyone please help ... I don't know what purevideo is, and I only download and install the driver for the 8800GT.

What else would I need installed to run this ? I'm in PAL territory.

You've not been reading the threads about this.

Did you compare PureVideo to yadif?Sorry ... anyone got a link ?

[hajj_3]

purevideo isnt free, you can buy it from nvidia, it allows your videocard to do some of the processing when playing dvds/blurays so less of your cpu will be used. players like cyberlink powerdvd allow purevideo support.

[halsboss]

Thankyou. Do I need to buy it to use this new DGMPGDecNV 1.0.0 ?

[neuron2]

Er, can anyone please help ... I don't know what purevideo is, and I only download and install the driver for the 8800GT. If you are talking about in the context of DGMPGDecNV, then it is already supported. All you need to do is enable it, either in the menu of DGMPGIndexNV, or using deinterlace=1 for DGMPGDecodeNV. I don't know what hajj_3 is talking about. If you are talking about some other context, then it's OT for this thread.

What else would I need installed to run this? Nothing.

Google is your friend.

[hajj_3]

i was on about purevideo, it isnt included in nvidia drivers is it? it didnt used to be atleast. NVIDIA PureVideo Decoder 1.02.233 is an installer i have which allows video players like powerdvd to use the gpu to playback videos so that it uses less of your cpu. Isnt this what DGMPGDec uses?

what are the differences between the 2 different options "single rate" and "double rate", which should we choose?

[neuron2]

i was on about purevideo, it isnt included in nvidia drivers is it? Yes it is.

it didnt used to be atleast. NVIDIA PureVideo Decoder 1.02.233 is an installer i have which allows video players like powerdvd to use the gpu to playback videos so that it uses less of your cpu. Isnt this what DGMPGDec uses? No.

what are the differences between the 2 different options "single rate" and "double rate", which should we choose? Single rate is deinterlacing. Double rate is bobbing.

[nixo]

A small thing: When I open DGMPGIndexNV.exe and select 'Disable Display', then open a vid, and then deselect 'Disable Display', I get a crash. Only seems to happen when I do things in that order, though.

--
Nikolaj

[neuron2]

Turns out when demuxing with DGIndex I kept losing the first 10 frames of the vob. OK, I have this fixed and will release it this evening.

And thanks for your "disable display" bug report too. I will fix it.

[rica]

neuron,
Are you too fast or am i too slow to catch you ? :)

[rebelvigilante]

I keep running into an error trying to run the included DGIndex- Missing License File... I'm attempting to use a .m2v I demuxed utilizing my regular DGIndex for testing purposes... am I just doing something wrong?

:edit:

Apparently I need to donate :o oops- I guess I'll be back this again once I find work and pay off the utilities

[Audionut]

Sweet, thanks Don.

[Inventive Software]

Yet another excuse for me to get an NVIDIA card. Won't happen for a while yet (poor student, money has different priorities). Nice work Donald. :)

[laserfan]

I noticed today that you include CUVID Server 1.04 with DGMPGDecNV, but still have zipped the older 1.03 with AVC and VC1 versions. Is this an oversight, or should I NOT be using 1.04 for all 3 of your apps (I currently have 1.04 in its own directory and launch it for any of the three)?

[neuron2]

Always use the latest version of the CUVID server and nvcuvid.dll.

I'll have to think about packaging some more.

[kebulek]

How important is updating that "win\system32\nvcuvid.dll" file? I forgot to update it with every new DGxxxNV versions... :/

[neuron2]

How important is updating that "win\system32\nvcuvid.dll" file? You could miss out on bug fixes. Try not to forget.

[neuron2]

Here is the second test version. I'll release this if there are no big problems found.

http://neuron2.net/dgmpgdecnv/dgmpgdecnv100b.zip

[~bT~]

^ i notice u have included dgdecode.dll with this release, is this intentional?

[neuron2]

Oops. Please re-download it. Thank you.

[kebulek]

From DGAVCIndexNV thread:
2. Fixed blocking on seeks with deinterlace=1.

@neuron: Did you fix this also in DGMPGIndexNV? I notice that in VDub few days ago, when I was seeking. And when i played same .avs in MPC, everything was fine. I'm in work right now, so I can't check new version. :)

[neuron2]

Everybody please re-download 1.0.0b. It had a bad bug in DGMPGDecodeNV.dll.

[blubberbirne]

Deinterlacing don't work for me. The encoded File stutter. Framerate is only 29,87 and not 29,97 after encoding.
If i use yadif for deinterlacing, everythink is fine.

[neuron2]

I can't help you without a source stream to allow me to duplicate your issue.

[lpm]

Using 1.0.0b (with NVidia GeForce 8500 GT on XP SP3 32-bit)

Your work rocks... the new tool seem to be working great functionally for me but there are glitches. Decoding works but is choppy and slow when frameserving with AVISynth... This is only with the new MPGSource, I get great performance when frameserving AVC with DGAVCDecNV. I will try to put together a reproducible test case but it looks like I'm not the only one experiencing so you might have one already.

One regret: so far MPGSource() doesn't take the "info" parameter and passes no hints by default so the ColorMatrix() filter can't automatically decide what to do if chained next.

A side suggestion: Today there are different command line syntaxes and AVISynth syntax variations between all the DG*Dec tools. How about merging all the tools (DGMPGDec, DGAVCDec, DGVC1Dec, DGMPGDecNV, DGAVCDecNV) into a single binary at some point - and a single AVISynth function - with the NV features enabled as a switch, and of course only when the required hardware/license are present? That would be awesome.... not a major requirement but a nice-to-have at some point.

[blubberbirne]

I can't help you without a source stream to allow me to duplicate your issue.

i will upload a small sample file later this evening (german time)

EDIT:

Here is the Sample: http://rapidshare.com/files/198874638/sample.rar

it also includes 2 Encodes. One with Cuda Deinterlace and the other one with Yadif.

i use this script for encoding...


loadplugin("DGMPGDecodeNV.dll")
#Load_Stdcall_plugin("yadif.dll")
mpgsource("sample.dgm", deinterlace=1)
#Yadif(2,1) #software-deinterlace
crop(2,0,1914,1076)#Temp
spline36resize(1280,720) #Temp

[blubberbirne]

I hope you noticed my edited post :)

[neuron2]

I tried with your stream and script and viewed the script in VirtualDub. Everything is fine while your MP4 is not.

Load your script into VirtualDub and single step. Is that OK? If OK, then something went wrong in your subsequent encoding. If not OK, then what is your OS, graphics card, and versions of DGMPGDecNV and CUVID Server?

[blubberbirne]

Encoding with Virtual Dub to xvid fails also.

I used the following Stuff here
- Vista x64 Sp1 and all Updates
- GeForce GTX260 with Driver 182.05 (but i will fallback to 181.22 for testing later)
- DGMPGDecNV and CUVID Server are both up2date.

[blubberbirne]

lol i found the problem :D

CUVID server needs to start in windows xp compatible mode on vista64. After this, everything works fine :D

So i think you need something to fix in the CUVID Server

[neuron2]

I don't have Vista 64 so unless you'd like to donate one, I'll just continue to pretend that it doesn't exist. But thank you for adding this useful tidbit.

[Sagekilla]

@neuron2: I just recently fixed my desktop computer, and I have a copy of Win Vista x64 on hand that I'm about to install. Would you like access to the desktop?

[neuron2]

You mean via Remote Desktop or some such thing?

Thanks, but I'd much rather bring up a dedicated machine.

[squid_80]

You can't use Direct3D over remote desktop so CUVIDServer wouldn't work anyway.

[Deinorius]

DGAVCIndexNV showed us, that quality is better than by decoding with DGAVCIndex. CRF-Encode was quite smaller.

Is there the same effect with DGMPGIndexNV? For example with DVDs?

[neuron2]

DGAVCIndexNV showed us, that quality is better than by decoding with DGAVCIndex. CRF-Encode was quite smaller. That's surprising and if it happens I can't see how it could possibly be due to just using DGAVCDecNV versus DGAVCDec.

Is there the same effect with DGMPGIndexNV? For example with DVDs? No such effect that I am aware of.

[halsboss]

Using 1.0.0b (with NVidia GeForce 8500 GT on XP SP3 32-bit) ... Your work rocks
I have an 8500GT 512Mb and am considering donating to get access to the NV tools... is there much quality/speed advantage of the NV versions ? Also, does one donate provide access to mpg&avc and can you do it via paypal ?

[Deinorius]

@neuron2
Well I was surprised too after I read that. I'm reffering to this post (http://forum.doom9.org/showthread.php?p=1185797#post1185797).

[Audionut]

I have an 8500GT 512Mb and am considering donating to get access to the NV tools... is there much quality/speed advantage of the NV versions ? Also, does one donate provide access to mpg&avc and can you do it via paypal ?

There is a slight speed increase as the decoding is moved to the GPU allowing more CPU time to encode. Depending on the encoder settings used, sometimes this can be a decent increase.

And the license works across all apps. So donating to 1 app will give you access to all apps. And yes to paypal.

[squid_80]

GPU MPEG2 decoding is wayyy slower than CPU decoding, unless you need to deinterlace as well.

[halsboss]

Oh. Thanks people. I'm mpeg2 only, so there's no advantage for me by the sound of that. Thanks anyway.

[ViRGE]

I've done a bit of digging and haven't found the answer, so I apologize in advance if this is a dumb question. Does this work on the G80-series GPUs (8800GTX/Ultra/GTS) and their bastardized VP1 hardware? I really want to get my hands on the PV deinterlacer, but I know this hardware doesn't have complete MPEG-2 decoding features on-board.

[neuron2]

Did you bother to read my web page about this tool?

P.S. You need VP2 or better.

[ViRGE]

Did you bother to read my web page about this tool?

P.S. You need VP2 or better.I did, and I did not find it at the time. Thank you for the quick response. :)

[Kurtnoise]

@Neuron2: any plan for VOB files support ?

[~bT~]

@Neuron2: any plan for VOB files support ?

i really want that too.

ps. looks like u got yourself an nvidia card after all :p good for us :D

[Kurtnoise]

ps. looks like u got yourself an nvidia card after all :p good for us :D
I'm still an ATI fanboy despite the fact that CUDA is a great thing...:cool:

[halsboss]

Does this work on the G80-series GPUs (8800GTX/Ultra/GTS) and their bastardized VP1 hardware? I really want to get my hands on the PV deinterlacer, but I know this hardware doesn't have complete MPEG-2 decoding features on-board.
What ? My 8800GT is crippled piece of rubbish ? But ... at the time, the specs looked sooo good in terms of items processed per sec even by comparison the later 9x00 series which appeared to have much lower throughput specs.

[squid_80]

What ? My 8800GT is crippled piece of rubbish ?
Relax, no-one mentioned the 8800GT. It should work fine.

[neuron2]

@Neuron2: any plan for VOB files support ? Program stream support is on my list, yes. I also need to support multiple input files.

[Chefkoch_ico]

Deinterlacing don't work for me. The encoded File stutter. Framerate is only 29,87 and not 29,97 after encoding.
If i use yadif for deinterlacing, everythink is fine.

Same here.

lol i found the problem :D

CUVID server needs to start in windows xp compatible mode on vista64. After this, everything works fine :D

So i think you need something to fix in the CUVID Server

Same here.

Can I help somehow?

[neuron2]

I don't have Vista64, so I can't do anything about this. Do what blubberbirne suggested.

[Chefkoch_ico]

I did, but it does still not work correctly in VDub and in MPC.

Normal Vista x64: deinterlace simply does nothing.
XP SP2 Compability mode, no deinterlace: OK
XP SP2 Compability mode, deinterlace: it does DeInterlace, but every few frames it shows some previously already showed frames, so the frames are "jumping".

Bye

[neuron2]

I'll ask Nvidia about it.

[Chefkoch_ico]

Thx, I am just uploading samples to rapidshare. Post Link soon.

Bye

[Chefkoch_ico]

I did, but it does still not work correctly in VDub and in MPC.

Normal Vista x64: deinterlace simply does nothing.
XP SP2 Compability mode, no deinterlace: OK
XP SP2 Compability mode, deinterlace: it does DeInterlace, but every few frames it shows some previously already showed frames, so the frames are "jumping".

Hi.

http://rapidshare.com/files/205239339/DGMPGDecNV_Vistax64_Deinterlace_Issue.rar.html

Contents:
a small ts recorded from DTV.
a AVS file
4 AVIs (XVid,mp3) encoded in VDub with various settings (see filenames)

Take a look to Frame 990 for example. And please dont download, before neuron2 has it.

Bye

[blubberbirne]

@chefkoch_ico

you talking about the same problem i have ;)

I think the cuvidserver is not 64bit compilant, but neuron can't test it, cause he has no 64bit os.

[woah!]

@chefkoch_ico

you talking about the same problem i have ;)

I think the cuvidserver is not 64bit compilant, but neuron can't test it, cause he has no 64bit os.

same thing here on xp with his sample he posted,so its not a vista issue ...

[neuron2]

I suspect that the blended fields are getting PureVideo confused. Can you make this happen on a clip that is not field-blended?

[lucassp]

I think there is a problem with the PureVideo deinterlacer. Here are the scripts and the results:

1. Single rate deinterlacing:
MPGSource("E:\Columbia\VTS_01_1.demuxed.dgm", deinterlace=1)
crop( 10, 4, -10, -2)
LanczosResize(640,528)
trim(0,2500)

Result: http://www.mediafire.com/?0i22dlmaj2n

2. Double rate deinterlacing:
MPGSource("E:\Columbia\VTS_01_1.demuxed.dgm", deinterlace=2)
ChangeFPS(25, linear=true)
crop( 10, 4, -10, -2)
LanczosResize(640,528)
trim(0,2500)

Result: http://www.mediafire.com/?1id02naju2a

The first video seems choppy to me and the second one smoother. Shouldn't those two videos look the same?

[agilpwc]

I use HDV cameras that output m2t files with mpg2 video and mpg1 audio.
When I open these files in the index program it comes up with an error about audio streams types not matching. Mentions ac3 and aac, which these files have neither of these types. This is preventing me from doing batch command line encodes and also this didn't happen with your non GPU version.

I either need a way to specify the audio pid or disable error messages or both I guess. Or make it work like it did in the non accelerated version.

Thanks

Here is a link to sample file

http://avs-scratch.s3.amazonaws.com/09-03-01-cooper-mjmp.m2t

[meatwad]

Here's a sample of the audio delay issue I'm experiencing if I use VideoRedo to edit my files:

http://www.mediafire.com/?sharekey=45a06ef0a0aeac78e62ea590dc5e5dbbe04e75f6e8ebb871

It's very similar to the one xopowo discussed in the DGAVCDecNV thread.

DGIndex= 32ms.
DGMPGDecNV =-619 ms

Thanks for wanting to fix the issue Neuron2.

[neuron2]

Okay, I've just released updates to DGAVCDec, DGAVCDecNV, and DGVC1DevNV, so I'm on the case over here now.

[UsedUser]

Works nicely. Using the PureVideo deinterlacer on 1080i TS -> 720p AVC encodes, I'm seeing 24% speed bump on first pass, and 12% on second pass. That's the fastest increase I've seen on anything, from swapping out any one piece in my encoding process. Thanks.

Question: How do I get audio to demux? The documentation makes it look like it's supported, but I'm assuming this just isn't implemented yet. I select Audio > Audio Demux. Set the one and only audio stream (which is already set anyway); close the window. Save the project. No audio demuxed.

[UsedUser]

I'm getting an error then a crash on drag & drop of multiple files in sequence. Something is not re-initialized between files.

Drag & drop a TS file into the app. Do something if you like, or don't do anything. Drag & drop another TS file into the app. Try to play or save the project. Try File > Close then drag & drop again.

On Play: "No video frames found!"
On Save: "Timed out looking for sequence header."
Or sometimes the attached image then a crash. Couldn't reproduce the exact steps on the crash.

[UsedUser]

Is it a requirement that the screen saver and Windows account locking be disabled?

When I lock my Windows account (i.e., on screen saver start), the CUVID server fails. I get two pop ups:

Server: cuvidMapVideoFrame() error.
Server: cuvidDecodePicture() error.

The cuvidDecodePicture() error repeats until I quit encoding and restart the CUVID server.

[neuron2]

Question: How do I get audio to demux? The documentation makes it look like it's supported, but I'm assuming this just isn't implemented yet. I select Audio > Audio Demux. Set the one and only audio stream (which is already set anyway); close the window. Save the project. No audio demuxed. It should be working. Can you upload a fragment of your source stream that I can use to duplicate your issue?

As clearly described in the release notes, multiple file opening is not yet supported.

I don't know anything about the screensaver issue. It may be an Nvidia driver strangeness that I have no control over.

[squid_80]

Probably the same as the remote desktop issue - direct3d apps get the boot when windows returns to the login screen.

[UsedUser]

It should be working. Can you upload a fragment of your source stream that I can use to duplicate your issue?
Linky (http://rapidshare.de/files/46049619/Late_Night_With_Conan_O_Brien-2009-01-15-0-short-clip.ts.html). Thanks for taking a look.

As clearly described in the release notes, multiple file opening is not yet supported.
Ah, I see that now. I did read all the documentation. Don't know what I thought it meant at the time, but it didn't occur to me that it meant that. In a future release, you might consider clarifying what the statement means for the user: "The application must be closed and reopened between opening each new file." Thanks.

[neuron2]

Linky (http://rapidshare.de/files/46049619/Late_Night_With_Conan_O_Brien-2009-01-15-0-short-clip.ts.html). Thanks for taking a look. Looking now.

Ah, I see that now. I did read all the documentation. Don't know what I thought it meant at the time, but it didn't occur to me that it meant that. In a future release, you might consider clarifying what the statement means for the user: "The application must be closed and reopened between opening each new file." Thanks. Oh, I see what you meant now. What you say should work. I'll look at that too.

[neuron2]

OK, I've got the audio demux issue fixed. It's an unusual case. But ouch, it affects all the other tools and I just released updates of them. :(

Looking at the other issue...

[UsedUser]

OK, I've got the audio demux issue fixed. It's an unusual case. But ouch, it affects all the other tools and I just released updates of them. :(
Sorry! At least DGIndex was able to handle it ok.

Can I ask what was unusual about this case? I'm curious about the details.

It was originally an ATSC broadcast transport stream, which I segmented with VideoReDo. Also wondering if I should contact my local broadcast affiliate or VideoReDo developers about something they could avoid doing.

[neuron2]

The stream is fine.

As I parse video transport packets during Save Project, I look for the first I frame right in the packet data so that I can start demuxing audio packets as soon as possible (I don't want to wait until the video decoder sees that it is an I frame because some audio packets could be skipped). But I was looking only in packets with payload_unit_start_indicator set. The problem was that the sequence header carried quant matrices and this pushed the I frame picture header into the next transport packet, which did not have payload_unit_start_indicator set. And so I never detected the I frames and audio demuxing never got started.

[neuron2]

I just put the fixed version up for you:

http://neuron2.net/misc/DGMPGIndexNV.exe

Let me know how it works for you.

Still have to look at the drag-and-drop issue.

[UsedUser]

I just put the fixed version up for you: http://neuron2.net/misc/DGMPGIndexNV.exe
Worked like a charm, thanks.

[meatwad]

neuron2,

I found a solution to the audio delay issue with VideoReDo. After reading some of the above posts I decided to look at VideoReDo's options. By default, VideoReDo is checked to convert all I frames to GOP. Once I saw that, I unchecked the box and now the audio delay issue is gone.

Edit: Nevermind. I forgot I put your new exe inside the folder last night. The results were the same with I frames being converted to GOP. I didn't think you'd fixed the issue I was having with VideoReDo files, just UsedUsers. Anyway, the new .exe seems to have fixed my issue as well. I must have missed something in the above posts.

[meatwad]

I'm having a problem with CUVID server when I run MediaPortal. If I watch a video while CUVID server is being used to encode a video, CUVID server will crash when video playback ends in MediaPortal. I don't have this problem with MPC or Windows Media Player (at least not that I'm aware of). I'll look at my settings in MediaPortal to see if there's something I can change to keep this from happening.

Edit: I believe the problem is with the VMR9 renderer. I switched to EVR (software mode) in MediaPortal and the crash didn't occur. Anyway, I'm not sure why VMR9 would cause CUVID server to crash when if I use another player with VMR9 as the renderer everything is fine? Oh well--as long as it works.

[UsedUser]

I'm having a problem with CUVID server when I run MediaPortal...
In my experience, you can't have two resources using DXVA simultaneously. It seems MediaPortal could be using DXVA, which will either prevent the CUVIDServer from using DXVA to process the video, or cause it to error outright.

neuron2, correct me if I'm wrong on this, but my experience and results from testing indicate any two applications that attempt to use DXVA simultaneously will cause an error. This is a driver / hardware / Direct3D limitation --- one or all of them aren't designed for multitasking.

Your decoder has to support DXVA to send it to the output pin, and your renderer has to accept DXVA on the input pin. In your situation, I believe MediaPortal allowed DXVA with VMR9, but switching to EVR in software mode precluded the use of DXVA. I'm assuming the decoders in your other apps, i.e., your default DirectShow decoder, as used by WMP, don't support DXVA as configured or at all, so you don't have an issue with them using software decoding.

[squid_80]

neuron2, correct me if I'm wrong on this, but my experience and results from testing indicate any two applications that attempt to use DXVA simultaneously will cause an error. This is a driver / hardware / Direct3D limitation --- one or all of them aren't designed for multitasking.
The fact that multiple copies of DGAVCIndexNV can operate at the same time shows this isn't true.

[UsedUser]

The fact that multiple copies of DGAVCIndexNV can operate at the same time shows this isn't true.
Then I may stand corrected regarding its possibility, but I do know the apps I commonly use - VideoReDo, BeyondTV, ZoomPlayer - cannot be used simultaneously if I'm using DXVA decoders. Perhaps it has to do with the way the resources are consumed and released by those applications or decoders.

In any event, because I've run into it with such apps, and can work around it by using software decoders, it remains that it may be the issue as described with MediaPortal.

[neuron2]

I'm no expert on D3D and I have no access to the relevant Microsoft or Nvidia code, so I simply don't know. For that reason, I DO NOT SUPPORT any playing of videos while encoding with my NV tools. If you choose to do that, you are on your own.

[woah!]

would a field order transition cause issues with this app ?? i have stuttering frames if i process with this app, but using your dgindex it catches the transition and corrects it and no stutters. i can supply a clip if you need it, i am asking first to be honest.

[neuron2]

"to be honest" ???

Yes, of course I need a clip.

[woah!]

"to be honest" ???

Yes, of course I need a clip.

i pm'ed you a ftp link to the file.

[neuron2]

I saw the jerkiness with deinterlace=1.

I set use_top_field=false and it seems OK.

I'll have to ask Nvidia what is going on, because IIRC it should work both ways.

[woah!]

you added this detection to dgindex and it can correct the problem i have, i take it that is a major rewrite to add to DGMPGDecNV .

[neuron2]

I don't think the transition is relevant when pulldown is not honored.

I am investigating the issue further so standby. Using the option I described should be a workaround. Is it working for you?

[UsedUser]

Sidebar: I'd like to perform IVTC on some content and perform the deinterlacing with DGMPGDecNV. How would this be done today, without pulldown honored? How will it change once pulldown is honored?

I guess I'm asking if I just throw my favorite IVTC filter into my AVS script, will it work with hardware deinterlacing, or is frame order, etc going to be screwed at that point? Once pulldown is honored, does NV have hardware IVTC?

[neuron2]

Sidebar: I'd like to perform IVTC as well as deinterlacing with DGMPGDecNV. How would this be done today, without pulldown honored? How will it change once pulldown is honored? Currently pulldown is ignored. This is equivalent to performing IVTC *if the pulldown is consistent 3:2 throughout*. If you want pulldown honored, as long as the pulldown is 3:2 consistent throughout, you can just apply pulldown with DGPulldown or the tool of your choice. It's not important which fields are pulled down in such cases. You only really need external IVTC processing when you have irregular pulldown. You also need it if there is hard pulldown, but then the DGIndexNV options are irrelevant anyway.

Things get complicated when the soft pulldown is not constant (irregular). That is the reason for the distinction between Force Film mode and Ignore Pulldown in DGIndex.

So, if you have a stream with irregular pulldown, then DGMPGDecNV will have problems with it. Since this tool is designed for BluRays and I haven't seen a BluRay with irregular pulldown, I don't (yet) consider it a big deal. At some point, if it's needed, I will add options equivalent to DGIndex.

I guess I'm asking if I just throw my favorite IVTC filter into my AVS script, will it work with hardware deinterlacing, or is frame order, etc going to be screwed at that point? Once pulldown is honored, does NV have hardware IVTC? You don't have to use an external filter. Since pulldown is ignored, IVTC is done automatically as explained above.

[UsedUser]

If you want pulldown honored, as long as the pulldown is 3:2 consistent throughout, you can just apply pulldown with DGPulldown or the tool of your choice.

You don't have to use an external filter. Since pulldown is ignored, IVTC is done automatically as explained above.
Sorry, I'm slightly confused, but I feel like I'm almost there with you.

If my content is 1080i 29.97fps and I am deinterlacing in hardware with DGMPGDecNV, and the output still ends up at ~29.97fps, what does that mean? My content is not consistent 3:2?

If I know the content is film, and needs to be IVTC'ed to ~23.97, then I need to use an external filter, which will work in an AVS script after MPGSource(...,'deinterlace=1')?

[woah!]

I don't think the transition is relevant when pulldown is not honored.

I am investigating the issue further so standby. Using the option I described should be a workaround. Is it working for you?

yes that did seem to fix the issue, weird that it does tho no?

[neuron2]

If my content is 1080i 29.97fps and I am deinterlacing in hardware with DGMPGDecNV, and the output still ends up at ~29.97fps, what does that mean? My content is not consistent 3:2? You appear severely confused. Please post an unprocessed sample of your source so we can use it to unconfuse you.

If I know the content is film, and needs to be IVTC'ed to ~23.97, then I need to use an external filter, which will work in an AVS script after MPGSource(...,'deinterlace=1')? No, you don't deinterlace and do IVTC! You're so way out there that I won't try to set you right without an actual sample clip to use as an example. Awaiting your unprocessed source sample...

[UsedUser]

You appear severely confused. Please post an unprocessed sample of your source so we can use it to unconfuse you.
Haha, that's very kind of you to put it so delicately.

20 sec sample: http://rapidshare.de/files/46198293/Smallville_-_s02e23_-_44_-___Exodus___-_1080i_-_HDTV.ts.html

I'm working from the assumption that outside of effects shots, this series is shot on film @ 24fps then telecined to 29.97fps for TV.

[neuron2]

That stream is hard telecined. You can tell that because it has the field pattern of 3:2 pulldown but when you preview it in DGMPGDecNV, the Field Rpts box stays at zero. Since there is no soft pulldown, there is nothing to honor or ignore, and those modes are irrelevant, as I said in my first reply.

Summary for those who may not know: Soft pulldown (soft telecine) means the repeated fields are done by simply setting a flag on the fields to be pulled down; the extra fields are not physically present in the stream. So you have the option of just ignoring the flag to get the output video without the repeated fields. Hard pulldown (hard telecine) does not use flagging; the repeated fields are physically present in the stream. Therefore, you must physcially remove them with an IVTC filter.

So, the right way to handle your stream is to use a filter(s) to perform the IVTC. Here is one possible script:

loadplugin("dgmpgdecodenv.dll")
MPGSource("Smallville.dgm")
telecide()
decimate()

You do not set deinterlace=1. That is for interlaced video only, not telecined video.

Haha, that's very kind of you to put it so delicately. Good to see that you are a seeker after truth, and value straight talk. :)

[UsedUser]

So, the right way to handle your stream is to use a filter(s) to perform the IVTC. Here is one possible script:

loadplugin("dgmpgdecodenv.dll")
MPGSource("Smallville.dgm")
telecide()
decimate()

You do not set deinterlace=1. That is for interlaced video only, not telecined video.

Good to see that you are a seeker after truth, and value straight talk. :)
Awesome clarification, thank you.

Follow up question: With further enhancement to the NV tools, can IVTC with hard pulldown be done in hardware?

I don't know the Nvidia drivers, but I know in ATI drivers, there is an option for 3:2 pulldown detection. It seems to me that this would only honor flags for soft pulldown content to display it at the intended framerate, as you wouldn't need an option to ignore the flags and have the result be IVTC'ed content at roughly the original framerate.

So, I'm assuming such an Nvidia option wouldn't affect content with hard pulldown anyway, but I can hope.

:thanks:

[neuron2]

You're baffling me again, so I will just answer the part I can understand.

There is nothing in the video API at this time to specify IVTC of hard telecined streams on the Nvidia GPU.

[UsedUser]

You're baffling me again, so I will just answer the part I can understand.

There is nothing in the video API at this time to specify IVTC of hard telecined streams on the Nvidia GPU.
You crack me up. Sorry, I'm very dangerous with a little information. I know there's a lot I don't know, but I'm trying.

You say there's no API for IVTC with hard pulldown. Do do you suspect that's because it hasn't been implemented in the API, or because the IVTC driver option is only for soft pulldown?

Nvidia driver setting has this description for IVTC:
Provides more accurate movie playback and superior picture quality by recovering original images from film-converted-to-video (e.g. DVDs, 1080i HD content).

[neuron2]

There is a low-level API for using video decoding called CUVID. This API does not allow for controlling the postprocessing beyond enabling the PureVideo deinterlacer. I can ask Nvidia if it will ever be augmented, but right now there is no way for me to enable anything else through the CUVID API.

[UsedUser]

There is a low-level API for using video decoding called CUVID. This API does not allow for controlling the postprocessing beyond enabling the PureVideo deinterlacer. I can ask Nvidia if it will ever be augmented, but right now there is no way for me to enable anything else through the CUVID API.
Thanks. If you have an opportunity to ask Nvidia, I'd be curious to know if it will be available at some point, and what type of pulldown PureVideo handles.

[Varies]

lol, i found decision of VOB processing :) - Demux video by DGIndex and work with .m2v :)
sad, but nvidia deinterlace can't cope with full interlaced video.
here (http://rghost.ru/160086?key=ad72b6078c8e8233a69518cbce07746d) sample of video. (mcbob good but slow)

[neuron2]

Is that sample the source or your encoded result?

I ask because it's progressive at 29.97, which would be very unusual.

You're not deinterlacing telecined content are you?

nvidia deinterlace can't cope with full interlaced video That's absurd FUD.

[Varies]

Is that sample the source or your encoded result?
demuxed VOB with "honor pulldown"
because it's progressive at 29.97
:eek: huh, its real ?
You're not deinterlacing telecined content are you?
I try few deint...TDeint, TFM, Telecine, FieldDeinterlace, Yadif...and only mсbob_0.3u made a normal picture...
I wrote "full interlaced" because one nub :rolleyes: wrote me "clear interlace |translated" ... and i think about 29.97 interlaced frames :D
:helpful:

[neuron2]

:eek: huh, its real ?
Did you try looking at it? Why do you want to deinterlace something that is already progressive?

[Varies]

Did you try looking at it? Why do you want to deinterlace something that is already progressive?
ofcourse, when i looked at this
http://i043.radikal.ru/0903/b5/05fca9788263t.jpg (http://radikal.ru/F/i043.radikal.ru/0903/b5/05fca9788263.png.html) http://i010.radikal.ru/0903/08/8b55778c96c5t.jpg (http://radikal.ru/F/i010.radikal.ru/0903/08/8b55778c96c5.png.html)
http://s44.radikal.ru/i103/0903/a2/a5e70a3d4ba9t.jpg (http://radikal.ru/F/s44.radikal.ru/i103/0903/a2/a5e70a3d4ba9.png.html) http://s44.radikal.ru/i103/0903/2f/f21bf3771d63t.jpg (http://radikal.ru/F/s44.radikal.ru/i103/0903/2f/f21bf3771d63.png.html) (without deint, last with decimate)
i think about interlace :)
...and not all video is progressive, part of the video is Interlaced...also when mcbob made normal picture i think about interlace.
and decomb VFR stats #VFR Stats: FILM cycles 5983 (70.4%) NONFILM cycles 2517 (29.6%) TOTAL cycles 8500
#VFR Stats: FILM frames 23930 (65.5%) NONFILM frames 12585 (34.5%) TOTAL frames 36515
maybe i must switch something in dgindex?
Forcing film got same result.
-----
last idea with m2v not very good :D I received A/V desync after encoding.
* add to DGMPGIndexNV start button for CUVID :)

[neuron2]

ofcourse, when i looked at this
i think about interlace That is not interlacing.

...and not all video is progressive, part of the video is Interlaced You ask about deinterlacing and then post a part that is progressive!? Post a part that is interlaced.

What is the title and region of your DVD that this VOB sample came from?

[Varies]

That is not interlacing.
so what is this ?
Post a part that is interlaced.
here (http://rghost.ru/171650) interlaced part. fielddeinterlace and nvdeinterlace gives normal picture, but TDeint and TFM gives worse result.
What is the title and region of your DVD
Region 2 Japan :) it's usually for this country :D
and here (http://rghost.ru/160819) @60 fps processing like mcbob but its not mcbob. also mcbob gives same result.

[neuron2]

You have subtitles rendered at 60i overlaid on video at 30p.

It looks fine to me with:

MPGSource("E:\tmp\Sekirei_vol3_s1.demuxed.dgm",deinterlace=1)

If you don't like the PureVideo deinterlacing here, then use something that you do like.

[Varies]

yeah, looks not bad :) but otaku evil peoples, they will kill me for this picture :D
also nvdeint better than tipical software deinterlacers, it' good :)

switchin "wait mode" for VOB processing :D

[neuron2]

Merger of AVC, VC1, and MPG comes first. I already have the DLL merged.

[QuadraQ]

I just donated to support your site and get the NV version of AVC, VC1, and MPG. I have an mpeg2 file I created with VideoReDo that works fine with DGIndex, but when I try to open it with DGMPGIndexNV I get this error:

---------------------------
File open error
---------------------------
Program streams not supported.
---------------------------
OK
---------------------------

I'm sorry but I don't understand what that means, please advise. Thanks.

[neuron2]

Program streams (VOB and MPEG) are not yet supported. It's described in the release notes. Only TS, M2TS, and elementary streams are supported so far. Program stream support will be added.

[woah!]

I just donated to support your site and get the NV version of AVC, VC1, and MPG. I have an mpeg2 file I created with VideoReDo that works fine with DGIndex, but when I try to open it with DGMPGIndexNV I get this error:

---------------------------
File open error
---------------------------
Program streams not supported.
---------------------------
OK
---------------------------

I'm sorry but I don't understand what that means, please advise. Thanks.

just save the edited file as a .ts instead of a .mpeg

[laserfan]

just save the edited file as a .ts instead of a .mpegNo, save out of VRD as Elementary Streams.

[neuron2]

Both ways will work.

[laserfan]

Only TS, M2TS, and elementary streams are supported so far. Program stream support will be added.I missed this, I'd been thinking a "program stream" was anything not a pure video-only elementary stream. Sorry for my mistake, woah! and neuron2 (and QuadraQ). :o

[QuadraQ]

Program streams (VOB and MPEG) are not yet supported. It's described in the release notes. Only TS, M2TS, and elementary streams are supported so far. Program stream support will be added.

That was pretty stupid of me. I think I was just braindead after work. I should know what elementary streams are by now! I don't think I hear about "Program streams" as much, but I figured it out on my own after I posted, and felt silly.

I'm getting about 9 fps in meGUI after I demux the mpeg2 file and use DGMPGDecNV instead of about 6 fps. Not as much of a bump as I was hoping for, but every little bit helps. (On a 9600 GT card BTW)

[woah!]

That was pretty stupid of me. I think I was just braindead after work. I should know what elementary streams are by now! I don't think I hear about "Program streams" as much, but I figured it out on my own after I posted, and felt silly.

I'm getting about 9 fps in meGUI after I demux the mpeg2 file and use DGMPGDecNV instead of about 6 fps. Not as much of a bump as I was hoping for, but every little bit helps. (On a 9600 GT card BTW)

well to be honest thats a 50% increase in frames...

[vucloutr]

Hi

I have a HDTV transport stream here where I get 2 video glitches with DGMPGDecNV/DGMPGIndexNV.

Here's a sample: http://www.mediafire.com/file/inwyxtinz3z/sample.ts (~33MiB)
Frame 438-454 and 1038-1054.

[neuron2]

Ah, good time to roll out my merged DLL for testing.

http://neuron2.net/misc/DGDecodeNV.zip

This DLL supports AVC, VC1, and MPG. It seems to work fine with your clip.

The script:

loadplugin("dgdecodenv.dll")
DGSource("sample.dgm")

[neuron2]

I also fixed the DGMPGDecodeNV.dll in the 1.0.0 distribution if you prefer to use that. Just redownload 1.0.0.

[vucloutr]

Yes, both are working fine. Thank you!

PS: how about putting the CUVID Server right into the package ?

[UsedUser]

I also fixed the DGMPGDecodeNV.dll in the 1.0.0 distribution if you prefer to use that. Just redownload 1.0.0.
Can I ask, if you have one, your philosophy on major version, minor version, revision, and build numbers?

Down the road, when I've forgotten what I've updated where, I'm concerned about being able to tell what version of your package I have installed. The file version information is the easiest way on an assembly/PE level to tell, but you don't seem to be iterating even the build number.

[neuron2]

I'm not as rigorous about it as I could be, sorry. It's a lot of work to do it right and I have all these different versions. Maybe after I combine them all, I'll get back to some rational system.

[Varies]

Honor pulldown flags work? ^ ^'

[neuron2]

Honor pulldown flags work? ^ ^' Read the release notes.

[laserfan]

Ah, good time to roll out my merged DLL for testing. http://neuron2.net/misc/DGDecodeNV.zip
Not sure why there isn't more excitement about this (though I've just now gotten 'round to trying it myself :o). Just want to say "thanks"! and looking forward to the rest of your consolidation efforts! Just tried this on a VC1 program and it worked great--my AVS template is getting much simpler:

#First use of new DGDecodeNV!!! Located already in C:\Program Files (x86)\AviSynth 2.5\plugins directory.

#For h264
#DGSource("video.dga")

#For MPEG2
#DGSource("video.dgm")

#For VC-1
DGSource("video.dgv")

and it used to look like this (yeah I know this is uglier than it need(ed) to be):

#For h264 use DGAVCindex or DGAVCindexNV
#LoadPlugin("c:\program files (x86)\DGAVCindex\DGAVCDecode.dll")
#LoadPlugin("c:\program files (x86)\DGAVCDecNV\DGAVCDecodeNV.dll")
#AVCSource("d:\video\video.dga")

#For MPEG2 use DGMPGindex (.d2v) or DGMPGindexNV (.dgm)
#LoadPlugin("c:\program files (x86)\DGMPGDecNV\DGMPGDecodeNV.dll")
#MPEG2Source("d:\video\video.d2v")
#LoadPlugin("c:\program files (x86)\DGMPGDecNV\DGMPGDecodeNV.dll")
#MPGSource("d:\video\video.dgm")

#For VC-1 use DGVC1indexNV
LoadPlugin("c:\program files (x86)\DGVC1DecNV\DGVC1DecodeNV.dll")
VC1Source("D:\video\video.dgv")

#Or for VC-1 make a graph w/Haali Splitter and WMV DMO filters
#DirectShowSource("d:\video\video.grf", fps=23.976, audio=false, framecount=135119)
:thanks:

[Sharc]

Not sure why there isn't more excitement about this ....

I wasn't aware of its existence. Sure I will try it. It should greatly simplify the process. Thanks neuron2.

[CruNcher]

Don is there no way to use the Deinterlacer independent from the input ?

[neuron2]

Don is there no way to use the Deinterlacer independent from the input ? Sorry but I don't understand the question. Can you clarify it please?

[CruNcher]

independent of the supported Video formats for any loaded Video :) something like nvdeint()

[Varies]

yeah :D and something like nvaaa() for nv hardware antialliasing :)

[UsedUser]

independent of the supported Video formats for any loaded Video :) something like nvdeint()
It would be useful in creating a generic AVS script, particularly for use with MeGUI, where it would be easier to create an AVS template with separate <input> and <deinterlace> sections.

yeah :D and something like nvaaa() for nv hardware antialliasing :)
See below.

There is a low-level API for using video decoding called CUVID. This API does not allow for controlling the postprocessing beyond enabling the PureVideo deinterlacer. I can ask Nvidia if it will ever be augmented, but right now there is no way for me to enable anything else through the CUVID API.

[Blue_MiSfit]

Hi neuron2,

Any plans for supporting 4:2:2 MPEG-2?

I'd imagine it's up to nVidia, but just thought I'd check :)

I have lots of 4:2:2 content, and would love to be able to run it through DGMPGIndexNV, and then decode with hardware deinterlacing!

~MiSfit

[neuron2]

No plans until the API supports it.

[UsedUser]

That stream is hard telecined. You can tell that because it has the field pattern of 3:2 pulldown but when you preview it in DGMPGDecNV, the Field Rpts box stays at zero. Since there is no soft pulldown, there is nothing to honor or ignore, and those modes are irrelevant, as I said in my first reply.
...
Good to see that you are a seeker after truth, and value straight talk. :)
Seeking more truth. :)

I have another clip that shows 59% video in the d2v. The Field Rpts value does increment in DGIndex and in DGMPGIndexNV with PureVideo deinterlacer off. If I enable the single rate PureVideo deinterlacer, the Field Rpts value does not increment.

1) How do you determine the field pattern?

2) Do I want to enable the PureVideo deinterlacer when I create the dgm, if I intend to enable it in my AVS script, or doesn't it matter?

3) What else do I need to know to determine whether I should A) rely on the IVTC done by DGMPGDecNV ignoring the pulldown flags, and that's enough, B) use an external IVTC filter, or C) deinterlace (with PureVideo)?

Clip: http://www.mediafire.com/?sharekey=e967b615389dce0b391d7d881749d3a7e04e75f6e8ebb871

[neuron2]

1) How do you determine the field pattern? http://neuron2.net/faq.html questions 1 and 2. This is an interesting clip. If you save project in DGIndex with Honor Pulldown Flags, you get the classical 3:2 field pattern, but it is achieved with a mix of hard and soft telecine. Therefore, this works:

telecide()
decimate()

Note that I said DGIndex! This clip cannot be handled properly with DGMPGDecNV (yet). Honoring of pulldown is not yet supported, so that won't work. And Ignoring won't work either, because there is a mix of hard and soft. And Force Film isn't implemented. So I need to add the proper modes as in DGIndex. I'll bump that up in priority as I see that DGMPGDecNV is being used for TS in addition to the BluRay M2TS that I intended it for.

[2) Do I want to enable the PureVideo deinterlacer when I create the dgm, if I intend to enable it in my AVS script, or doesn't it matter? Never deinterlace clean progressive material! Anyway, the setting in DGMPGIndexNV does not alter the DGM file. You would need to enable it in your DGSource() involation, but as noted, not for this case.

[UsedUser]

This is an interesting clip. If you save project in DGIndex with Honor Pulldown Flags, you get the classical 3:2 field pattern, but it is achieved with a mix of hard and soft telecine.
Thanks. I know it uses soft pulldown because of the Frame Rpts value; how can I know it's also using hard pulldown? Create another d2v, but with Ignore Pulldown Flags, and I should see the 3:2 field pattern disappear?

Never deinterlace clean progressive material! Anyway, the setting in DGMPGIndexNV does not alter the DGM file. You would need to enable it in your DGSource() involation, but as noted, not for this case.
Knowing this clip is progressive 3:2, it needs IVTC, not deinterlacing - you taught me that lesson last time. ;) Because I was seeing different field repeat values depending on the PureVideo setting, I just wanted to know if it affected the outcome.

[neuron2]

how can I know it's also using hard pulldown? Use the method at the link I gave you. If you see the constant 3:2 field pattern but not all the duplicated fields are done with flags, then the only way for the remaining duplicates to be created is by hard copying.

Because I was seeing different field repeat values depending on the PureVideo setting, I just wanted to know if it affected the outcome. I'm not seeing that with your sample, and there's no reason for it to happen. Please check again.

[Fotis_Greece]

I am a bit confused. I have donated for a license, can I use this license also for your other tools DGAVCDecNV and DGVC1DecNV?
And since DGDecodeNV works with all three programs (or am I wrong?) why don't you just "merge" them into one powerful application?
BTW nice work. Would it be possible to allow multiple selections of the loaded video so we could just ommit the commercials and then just deinterlace and save as mpeg2 ready for dvd authoring?

[neuron2]

I have donated for a license, can I use this license also for your other tools DGAVCDecNV and DGVC1DecNV? Yes. It says that at the web site. But feel free to make multiple donations if you prefer. :)

And since DGDecodeNV works with all three programs (or am I wrong?) why don't you just "merge" them into one powerful application? It's an ongoing process. I did the DLL first. I'm working on combining the indexers.

Would it be possible to allow multiple selections of the loaded video so we could just ommit the commercials and then just deinterlace and save as mpeg2 ready for dvd authoring? The problem with it is you could get orphaned frames that decode with macroblocks if you cut at open GOPs. I have an idea to make it work by replacing all the orphans at all the cut points with copies of the first decodable frame. Can't say when I might get to it though as these days half my time is going to theoretical physics simulations.

[Fotis_Greece]

Thanks neuron, and some last question (since I still wait for the license file so I cannot really test). When I load a m2ts (from HDV camcorder) file and select to enable deinterlacer and I click save and demux the resulting m2v video file is deinterlaced and there is no need to deal with dgdecode.dll and avisynth and cuvid server. Right?

So will the resulting video file be 1440x1080 25p (PAL country here)?
The original HDV is 1440x1080 50i (upper field first)

[tre31]

Thanks neuron, and some last question (since I still wait for the license file so I cannot really test). When I load a m2ts (from HDV camcorder) file and select to enable deinterlacer and I click save and demux the resulting m2v video file is deinterlaced and there is no need to deal with dgdecode.dll and avisynth and cuvid server. Right?

So will the resulting video file be 1440x1080 25p or 50p (PAL country here)?
The original HDV is 1440x1080 50i (upper field first)

No what dgindex basically does is index the source for frameserving too an encoder, so you still need avisynth too actually frameserve it, but yes using the cuda hardware deinterlacing will mean that within the avisynth script you will not need too deinterlace using other methods (not sure exactly how the cuda based dll works but I think it may have a deinterlacing parameter that is passed too it - neuron can correct me on this, also it should be in the manual, readme or faq - and if it isn't .. then ...).

[Fotis_Greece]

Well, I don't want to frameserve it. Only to deinterlace and demux.

[neuron2]

You cannot deinterlace without re-encoding using any tool.

You can remove pulldown so if the camera uses pulldown to achieve the video rate then you have a hope.

[Fotis_Greece]

So in other words when saving and demuxing (with deinterlace checked) the only operationtaken place is just saving project and demuxing the video-audio, right?

[neuron2]

Right, exactly as stated in the users manual.

[Varies]

I have a new problem with deinterlace :)
Here (http://rghost.ru/221487?key=e6bb6b21ede96cb8fec8a4d829566f5a) is the sample of video (source).

[neuron2]

What DVD did you rip that from? It's pathological.

[Varies]

It's TV translation from NHK-E japanese chanel ^ ^

I try other deint, best result with Decomb 3:2 pulldown remove. But same frames looks like incorrect :(

[neuron2]

It's TV translation from NHK-E japanese chanel ^ ^ How did you get it?

[Varies]

I download TS stream.
Don't ask me "how" :D This violates the rules :)

Inverse 3:2 Telecine with Pattern Guidance. If your telecined source material is NTSC 3:2 pulldown, you can enable pattern guidance, which can make the field matching more accurate for some clips. Proceed as follows:

AssumeTFF()
Telecide(order=1,guide=1)
Decimate(cycle=5)

[neuron2]

You've already been struck for a rule 6 so you're well aware of the rules. Struck for rule 6.

[hajj_3]

any eta for program stream support yet? would love to use this instead of dgmpgdec, should help with the speed alot of 1080i .mpg files.

[halsboss]

I know you're busy, so just adding +1 to that .... hoping you can get around to it in the list of things you're doing.

[Turtleggjp]

I'm having trouble with the hardware bob deinterlacer. When I played back my encoded video, it was stuttering and not smooth. Looking back at the served video frame by frame, I saw that I was getting some duplicate frames often followed by some combed (not deinterlaced at all) frames. I thought that the problem was with my video card (Nvidia GeForce 9650m GT, the m means it is in my laptop) since it often struggles playing 1080i content smoothly. When I tried to use the indexer to clip out a sample to upload, the sample no longer had any combed frames, but still had some of the duplicate frames. By duplicate frames, I mean that it is repeating a frame and then another frame is missing, so no extra frames are being added.

Thinking that my video card was to blame, I tried a similar transcode using some of my video camera's footage (Canon HF100, AVCHD 1080i). However, the hardware deinterlacer seemed to work perfectly with the clips I tried. This leads me to believe that it is some sort of bug in DGMPGDecNV's code. The problems also show up when doing a preview in the index tool.

When I get home tonight I'll try putting together some sample clips. I don't think there is anything more I can do until I have access to the original file again, as my first attempt at making a sample didn't reproduce all issues. Thanks.

Matt

[neuron2]

any eta for program stream support yet? would love to use this instead of dgmpgdec, should help with the speed alot of 1080i .mpg files. I have finally started it. I have DGMPGIndexNV working with video and AC3 and MPA audio. I plan to get DGDecodeNV working next and then add the other audio types. What audio types do you need most?

I don't plan to support MPEG1 video or MPEG1 program streams. Is that a problem?

[neuron2]

I'm having trouble with the hardware bob deinterlacer. Try both field orders with the use_top_field set as true and false. If it won't work either way then upload a failing source stream.

[hajj_3]

I have finally started it. I have DGMPGIndexNV working with video and AC3 and MPA audio. I plan to get DGDecodeNV working next and then add the other audio types. What audio types do you need most?

I don't plan to support MPEG1 video or MPEG1 program streams. Is that a problem?

Thats great i just cap ws.pdtv and 1080i mpeg2 so .ac3 and .mpa/.mp2 support is all i need and pretty much all anyone needs really, not many countries use mpeg2 with aac.

have u done any speed tests comparing it with the normal dgmpgdec? 1080i takes a while to make a .d2v file, SD is fairly quick, atleast for short tv shows.

No-one uses mpeg1 anymore, no point adding that.

[Turtleggjp]

Uploading a sample now. It is a .ts file trimmed with HDTV2MPEG 1.11.89.

In the indexer, you can see the repeated frames, but no combed frames (even with Full HD display enabled). If I index the file, and serve it with AVISynth, I can see repeated frames and the following combed frames. If I set deinterlace=1 with use_top_field=true, it looks smooth (no repeated/dropped frames). With use_top_field=false, I can see duplicated frames, followed by a noticeable jump caused by a missing frame.

Here is the sample:
http://www.mediafire.com/?sharekey=ffd20e881bdc4436e5c3dee5769931ece04e75f6e8ebb871

[halsboss]

I have fast sports field action HD sources with a lot of blocking (the channels mangle these, unfortunately, also upscaled SD due to inter-channel agreements). The non-NV version seems to provide deblocking natively and easily (can I make it stronger ?)
#ipp=true interlaced post-processing
#cpu=4 DEBLOCK_Y_H, DEBLOCK_Y_V, DEBLOCK_C_H, DEBLOCK_C_V
#cpu=6 DEBLOCK_Y_H, DEBLOCK_Y_V, DEBLOCK_C_H, DEBLOCK_C_V, DERING_Y, DERING_C
MPEG2Source("G:\HDTV\q2.d2v",info=0,ipp=true,cpu=6)
Does/will the NV version have the same facility ?

[neuron2]

Does/will the NV version have the same facility ? No, because the GPU delivers the decoded frame and my code has no access to the quants. You can use a deblocker in your script, such as BlindPP(), Deblock(), and others.

[neuron2]

1. When demuxing with the CLI, audio filenames no longer erroneously include the string ".dgm" in the filename.

2. Program streams are now supported. Currently, only MPA and AC3 audio is supported.

3. The machine ID field in the About box is now copyable.

4. Fixed a bug in PAT/PMT parsing.

5. Added a warning popup for cases of blank audio at the start of the stream that can cause incorrect reported audio delay. User is recommended to skip GOPs at the beginning of the project.

http://neuron2.net/dgmpgdecnv/dgmpgdecnv.html

Next I want to add pulldown handling and address the Purevideo deinterlacing issues that have been raised.

[hajj_3]

ooooo :)

The license file would work with DGAVCDecNV too right?

How is the purevideo de-interlacing doing, does it do a good visual job, is it better than leakkerneldeint or field de-interlace (no blend)? Is it faster than those?

If i wanted to use this with gordianknot what would i need to put in the .avs to be able to load the .dgm file and would there be any way for me to check the cropping in gordian knot or would i have to use a different app to find out the cropping then manually enter the cropping needed into a .avs file?

keep up the great work neuron!

[shon3i]

Something wrong when i select whole DVD or multiple vobs. Decoding stoped after first vob.

[neuron2]

Read the release notes.

[halsboss]

2. Program streams are now supported. Currently, only MPA and AC3 audio is supported.
etc
Nice. Thank you very much Neuron2.

[halsboss]

A hopeful query: if the GPU is doing the work of decoding and returning frames, and can deinterlace if you ask it to, are the cards/cuda smart enough to re-size (interlaced/deinterlaced) frames for you too to a specified size ? I guess I'm asking if that's possible, then would the NV based decoders may be able to resize something like HD1080i->SD576i. That'd be really really handy, if it could :) I ask about interlaced in particular since a lot of my footage is fast sports field action where the interlaced fields have quite large differences.

[squid_80]

Actually they can, the API allows the program to set the desired size of the decoded frames.

[neuron2]

Actually they can, the API allows the program to set the desired size of the decoded frames. Can you be more specific and/or provide sample code, please?

[squid_80]

From the cudaVideoDecode SDK sample app:
VideoDecoder::VideoDecoder(const CUVIDEOFORMAT & rVideoFormat)
{
// Validate video format. Currently only a subset is
// supported via the cuvid API.
cudaVideoCodec eCodec = rVideoFormat.codec;
assert(cudaVideoCodec_MPEG1 == eCodec || cudaVideoCodec_MPEG2 == eCodec || cudaVideoCodec_H264 == eCodec);
assert(cudaVideoChromaFormat_420 == rVideoFormat.chroma_format);
// Fill the decoder-create-info struct from the given video-format struct.
memset(&oVideoDecodeCreateInfo_, 0, sizeof(CUVIDDECODECREATEINFO));
// Create video decoder
oVideoDecodeCreateInfo_.CodecType = rVideoFormat.codec;
oVideoDecodeCreateInfo_.ulWidth = rVideoFormat.coded_width;
oVideoDecodeCreateInfo_.ulHeight = rVideoFormat.coded_height;
oVideoDecodeCreateInfo_.ulNumDecodeSurfaces = FrameQueue::cnMaximumSize;
// Limit decode memory to 24MB (16M pixels at 4:2:0 = 24M bytes)
while (oVideoDecodeCreateInfo_.ulNumDecodeSurfaces * rVideoFormat.coded_width * rVideoFormat.coded_height > 16*1024*1024)
{
oVideoDecodeCreateInfo_.ulNumDecodeSurfaces--;
}
oVideoDecodeCreateInfo_.ChromaFormat = rVideoFormat.chroma_format;
oVideoDecodeCreateInfo_.OutputFormat = cudaVideoSurfaceFormat_NV12;
oVideoDecodeCreateInfo_.DeinterlaceMode = (rVideoFormat.progressive_sequence) ? cudaVideoDeinterlaceMode_Weave : cudaVideoDeinterlaceMode_Adaptive;
// No scaling
oVideoDecodeCreateInfo_.ulTargetWidth = oVideoDecodeCreateInfo_.ulWidth; // Set this to the desired output width
oVideoDecodeCreateInfo_.ulTargetHeight = oVideoDecodeCreateInfo_.ulHeight; // Set this to the desired output height
oVideoDecodeCreateInfo_.ulNumOutputSurfaces = 2; // We won't simultaneously map more than 2 surfaces

// create the decoder
CUresult oResult = cuvidCreateDecoder(&oDecoder_, &oVideoDecodeCreateInfo_);
assert(CUDA_SUCCESS == oResult);
}
I haven't played with different values myself so YMMV.

[halsboss]

this does interlaced resizing, right ? sorry to harp on it, but that's important for some people (well, me) with 1080i captured sources with lots of movement and panning/zooming, and only a SD (PAL 576) DVD player to watch it with)

[neuron2]

Hold your horses there, bro'. We don't even know if it resizes at all yet.

[halsboss]

No worries :) I hope, but do not expect at all ! Imagine it, a GPU based sourcer/resizer. (I know, there's probably heaps of pragmatic issues which would rule it out). I wish upon a star that Nvidia had deblocking built-in as well, but that's for another decade or century :D All the best.

[Turtleggjp]

Next I want to add pulldown handling

Sweet. I'm unable to use your NV version for working with my TV shows right now, as most of them are using pulldown.

and address the Purevideo deinterlacing issues that have been raised.

Cool, so I can assume that you were able to repeat my problem on your end?

[neuron2]

Yes, and Nvidia is looking at it. It appears to affect MPEG2 only.

[laserfan]

Hey I notice you've slipped-in a newer DGDecodeNV.dll to this release package! I'm gonna assume I should use this latest version for my VC1 and AVC testing now, unless you tell me differently... :)

EDIT: Oops--now I looked at the Properties of the old vs. new and they appear identical anyway! Never mind then! :o

[neuron2]

@laserfan

It was a new version because it had to support program streams.

@all

Here is 1.0.3 beta 1 that supports multiple input files for MPG (e.g., load multiple VOBs). Do not try to use the DGDecodeNV.dll with AVC or VC1 (yet), because they need new versions of the respective indexers. Please test and report.

http://neuron2.net/dgmpgdecnv/dgmpgdec103b1.zip

[kemuri-_9]

Here is 1.0.3 beta 1 that supports multiple input files for MPG (e.g., load multiple VOBs).

worked fine to load multiple .vobs here...

[cweb]

I had tried a brief mpeg clip (recorded from satellite) with both the beta and release versions, and while it loaded in a complex script, the script would not encode to xvid in virtualdub without crashing..
I'm using NV beta 4 from your website.

Now I updated my graphics driver to nvidia's latest 186.18 as recommended in the beta, and virtualdub is getting stuck when I try to load a one line avs script to load a video (this time a dvd backup). DGMPGIndexNV is what I used to create the dgm file.

dgsource("sja102.dgm")

It just stays 'not responding'... both with PureVideo for deinterlacing and without it.
any ideas?
I had an nvcuvid.dll in system32 but deleted it now. Doesn't seem to have solved the problem.

-edit-
I think I found what's wrong.
I'm using MT avisynth, and setmtmode(2,2) before dgsource - which it doesn't like. Solution for now is to dgsource before calling setmtmode (and to kill any running cuvid server).

Also how can I move the software to a different directory on another drive on my machine without having to get another license file?

[neuron2]

I think I found what's wrong.
I'm using MT avisynth, and setmtmode(2,2) before dgsource - which it doesn't like. Correct.

Also how can I move the software to a different directory on another drive on my machine without having to get another license file? Just move it.

[yesgrey]

Just a small cosmetic thing in NVbeta4:
When I open the DGDecodeNVManual.html file the title in the web browser says "DGAVCDecodeNVManual" instead of "DGDecodeNVManual".

[prokhozhijj]

Thank you for your work, neuron2.

1. Can I use DGMPGDecNV with FFT3DGPU filter (http://forum.doom9.org/showthread.php?t=89941) safely? I've tried to do this with small video fragments and it seems everything looks fine. But I am not sure.
2. Do you planning to include noise reduction and edge enhancement features provided by NVidia?

I have NVidia 8800 GTS card and Vista32 OS.

[neuron2]

1. Can I use DGMPGDecNV with FFT3DGPU filter (http://forum.doom9.org/showthread.php?t=89941) safely? I've tried to do this with small video fragments and it seems everything looks fine. But I am not sure. I don't know for sure either. My guess is that it is OK.

2. Do you planning to include noise reduction and edge enhancement features provided by NVidia? If and when the NVCUVID API allows it.

[prokhozhijj]

I had tried DGMPGDecNV beta 4 with DVD (it contains 4 vob files, detailed info see below). Index has been created succesfully. But within avs-script decoding process hangs approximately in the middle of the film. For other film it was stopped near to end of third part of the film.

I am using last NVidia drivers (v.190.38 international for Windows Vista 32), AviSynth 2.5.8, Windows Vista 32-bit Ultimate, GeForce 8800 GTS, GOM player.

here is the script
DGSource(dga="G:\video\_new\Gold Globus\England\video.dgm",deinterlace=2)

here is the detailed info about first part of the film

Video
ID : 224 (0xE0)
Format : MPEG Video
Format version : Version 2
Format profile : Main@Main
Format settings, Matrix : Default
Duration : 24mn 8s
Bit rate mode : Variable
Bit rate : 5 245 Kbps
Nominal bit rate : 9 800 Kbps
Width : 720 pixels
Height : 576 pixels
Display aspect ratio : 4/3
Frame rate : 25.000 fps
Standard : PAL
Colorimetry : 4:2:0
Scan type : Interlaced
Scan order : Bottom Field First
Bits/(Pixel*Frame) : 0.506
Stream size : 906 MiB (88%)
Writing library : Jan 12 2008

- edit 1 -
Just discovered that with the deinterlace=1 everything works good. But I want to know what I am doing wrong?

- edit 2 -
I found in documentation: "Note that double rate deinterlacing requires Windows XP SP3!" Is there any possibility to use double rate deinterlacing on Vista SP2?

[Revgen]

^I have Windows XP SP3 and I have the same issue.using the same drivers. I'm using an 8400 GS.

[neuron2]

Is there any possibility to use double rate deinterlacing on Vista SP2? You have to ask Nvidia. That is the advice I have from them.

[neuron2]

I have Windows XP SP3 and I have the same issue.using the same drivers. I'm using an 8400 GS. I can't help you without a stream that I can use to duplicate the issue.

[Revgen]

I can't help you without a stream that I can use to duplicate the issue.

http://www.sendspace.com/file/3quomn

[neuron2]

Now tell me the steps to create your problem.

[Revgen]

^
1) Create .dgm file
2) AVS script
dgsource("espytest.dgm",deinterlace=2)
3) Open avs script in Virtualdub.
4) First half of video plays smooth. 2nd half doesn't play at all. 2nd half frames can be seeked, but cannot be played.

[an3k]

^
1) Create .dgm file
2) AVS script
dgsource("espytest.dgm",deinterlace=2)
3) Open avs script in Virtualdub.
4) First half of video plays smooth. 2nd half doesn't play at all. 2nd half frames can be seeked, but cannot be played.

have the same issue (with the sample file you provided). windows media player and other programs also have problems playing the 2nd half.
2nd half is frame 123 - 248.


WinXP Prof SP3 + AviSynth 2.5.8 + VirtualDub 1.8.8 + DGIndexNV beta 4 + GeForce GTX 295 + nVIDIA GeForce 168.18

[Revgen]

^When your using WMP, what player are you using. FFDShow works fine for me.

[neuron2]

^
1) Create .dgm file
2) AVS script
dgsource("espytest.dgm",deinterlace=2)
3) Open avs script in Virtualdub.
4) First half of video plays smooth. 2nd half doesn't play at all. 2nd half frames can be seeked, but cannot be played. I know what the problem is and a fix will be forthcoming shortly. :)

[Revgen]

^ :thanks:

[prokhozhijj]

Yet another issue with NVBeta5.

When I am opening 4 vob files in DGMPGIndexNV for index creating I am getting error with following message:

GPU decoder: failed to create video decoder [100]

After this error happened I can't to close DGMPGIndexNV.

I am using last NVidia drivers (v.190.38 international for Windows Vista 32), Windows Vista 32-bit Ultimate, GeForce 8800 GTS.

== Edit 1 ==
I can't to use created index in avisynth script at all. Neither with deinterlace=0 nor deinterlace=1 nor deinterlace=2. There is no video output on my player (GOM).

[Revgen]

^I'm not having this problem.

Tried NVBeta6?

Try uninstalling and reinstalling drivers?

[prokhozhijj]

^I'm not having this problem.


May be this error happens due to my OS (Vista)?


Tried NVBeta6?


I do not see NVBeta6 URL o neuron2.net.


Try uninstalling and reinstalling drivers?

Why? I had no such problem with NVbeta4.

== Edit 1==
With NVBeta6 the same problem. I found NVBeta6 there http://neuron2.net/dgavcdecnv/NVbeta6.zip

[neuron2]

Make sure you have the latest Nvidia drivers installed and try beta 7.

[cweb]

May be this error happens due to my OS (Vista)?



I do not see NVBeta6 URL o neuron2.net.
Just change the link you get from the site, replace the 5 by a 6 and you can download it..

[prokhozhijj]

Make sure you have the latest Nvidia drivers installed and try beta 7.

With NVBeta7 I have the same issue as with NVBeta5 and NVBeta6.

With NVBeta4 I have no such issue.

I've tried to reinstall NVidia drivers (190.38 International for Vista32, this is latest version for my video card 8800GTS) but it did not help.

[neuron2]

I'll ask Nvidia about it.

[prokhozhijj]

I'll ask Nvidia about it.

Any progress?

[neuron2]

You're the only guy with this problem. One other person had it and the cause was an old version of the DLL in the Avisynth plugin directory. Have you checked that? It's consistent with the fact tha beta 4 works for you.

[prokhozhijj]

You're the only guy with this problem. One other person had it and the cause was an old version of the DLL in the Avisynth plugin directory. Have you checked that? It's consistent with the fact tha beta 4 works for you.

Sorry, Neuron2, but I have two problems. You forgot probably about that. First problem is error generated by your application DGMPGIndexNV.exe after opening vob files - "GPU decoder: Failed to create video decoder [100]". Is this error happens due to old library in Avisynth plugins directory?

After this error arised your application can generate index file but after that I can not legally shutdown the program (through menu or window button "Close").

I've checked my Avisynth version. I wrote already that I have the latest stable official release. This is 2.5.8 AFAIK. But I've reinstalled Avisynth JIC. The second problem still occured when I do not see anything when using avs-script.

Are others guys have Windows Vista 32 SP2? May be problem in operating system difference (Xp SP3 and Vista SP2)?

== Edit 1 ==
I found old version nvcuvid.dll in this folder C:\Windows\System32\DriverStore\FileRepository\nv_disp.inf_df3f64cf. In others folders (C:\Windows\System32 and C:\Windows\System32\DriverStore\FileRepository\nv_disp.inf_12c1d22e) there is newest version of this dll. I think that old version of this library saved by OS for possibility to rollback in case when new drivers installed incorrectly.

== Edit 2 ==
Version of DGDecodeNV.dll in NVBeta7.zip is 1.0.2.0. The same version of this library in plugins Avisynth folder. I rewrote this dll in plugins folder from NVBeta7.zip archive.

[cweb]

I've checked my Avisynth version. I wrote already that I have the latest stable official release. This is 2.5.8 AFAIK. But I've reinstalled Avisynth JIC. The second problem still occured when I do not see anything when using avs-script.

Are others guys have Windows Vista 32 SP2? May be problem in operating system difference (Xp SP3 and Vista SP2)?


I use 2.5.8 (mt version), XP SP3 (32 bit). Im using the beta 6 version successfully.

[neuron2]

Make sure you have 190.38 or later for your Nvidia driver. Make sure you are not loading a rogue nvcuvid.dll from the past. It should be the latest one installed by Nvidia.

Does it happen with all VOB files?

[prokhozhijj]

Make sure you have 190.38 or later for your Nvidia driver. Make sure you are not loading a rogue nvcuvid.dll from the past. It should be the latest one installed by Nvidia.

Does it happen with all VOB files?

Neuron2, please read my previous posts. I've already reinstall all stuff which I can (drivers, Avisynth). Why repeat this question twice?

I did not load nvcuvid.dll from the past. I wrote which dll's from NVidia I have and where they located. Please read my posts carefully.

I've noticed that when I firstly close the project than I can to close DGMPGIndexNV.exe legally.

Issue happens with all VOB files.

[neuron2]

I don't know how to help you further.

[prokhozhijj]

I don't know how to help you further.

:thanks:

But did you test your soft on Vista 32 SP2? I did not found any restriction (except for deinterlacing) about Vista using.

And may be you'll make your application with extended logging? It may help. I see that you just do not want to investigate my problem. All advices comes to "just reinstall your soft". How you can make conclusion about anything without any logs?

[SomeJoe]

I believe somewhere on your system is an old version of something, probably something associated with the NVidia drivers.

Be aware that NVidia used to have a large control panel/control suite called nView that was bundled with older versions of their drivers. They removed nView support for GeForce cards in the drivers starting several months ago, nView is for Quadro cards only now.

If you had older nVidia drivers on the system when you upgraded to 190.38, nView may have stayed behind, and it will cause problems with the latest drivers. My recommendation:

1. Remove your nVidia drivers.
2. Remove nView if it is listed in the Add/Remove programs control panel.
3. Do a search on the Windows directory for all nVidia-related DLLs, especially nvcuvid.dll, and delete if any are found.
4. Delete C:\Program Files\nVidia if present.
5. Remove all versions of DG's NV tools from your system, including any DGNV-related DLLs present in C:\Program Files\Avisynth\plugins (like DGDecode.dll)
6. Restart and reinstall nVidia drivers 190.38.
7. Download and unzip the latest DGNV Beta 7 from neuron's site.
8. Try your steps again.

If this does not work, I'm not sure where else to go. Many other people here are running DGNV Tools beta 7 under XP SP3, Vista 32, Vista 64, and Win 7 with no issues.

Edit: Just saw in the other thread that you discovered that your GeForce card may not have sufficient hardware support for the NV Tools. If this turns out to be the case, please disregard.

[rica]

Hi neuron.
Sorry in advance if i've missed something since i've been away for a while.

I have DGMPGNV 1.0.2.0.

I have this system:

http://img193.imageshack.us/img193/6053/66006105.th.png (http://img193.imageshack.us/i/66006105.png/)

Here is the MediaInfo report of the original VOB:

http://img199.imageshack.us/img199/2848/vobinfo.th.png (http://img199.imageshack.us/i/vobinfo.png/)

And here is the sample rar including log, original VOB and dgm files:

http://rapidshare.com/files/273106157/DG.rar.html

1) Log is empty.
2) Created dgm is slow-motion.
3) Even MediaInfo says the original is progressive and even more "enable purevideo deinterlacer" is selected in "double rate", interlacing artifacts are very clear.

What's wrong with me?

[rica]

What's wrong with me?

Missing DGNVTools beta :D

[neuron2]

The logs were broken. I have just updated beta 9 with the fix for the logs.

Regarding your interlacing, the stream is hosed. Turn off deinterlacing, serve the video with SeparateFields() to look at the source fields. See what I mean?

I don't know what you mean by the dgm is slow-motion. Please redownload beta 9 and try with that, and if there is an issue tell me what you do to make it happen.

[rica]

neuron,

i've already done.

No problem left here.

Sorry, but i haven't been realized DGTools beta until now.

I was dreaming this morning a software which makes HW decoding-resizing on the fly like SimHD and now i found it. :)

Thanks neuron.

Here are my test results:

aviscrypt:

LoadPlugin("C:\users\rica\desktop\Tools\DGNVTools\DGDecodeNV.dll")

DGSource("C:\users\rica\desktop\Copy\DG_Newest\output_newest.dgm", resize_w=1280, resize_h=720)




WMP:

http://img42.imageshack.us/img42/9258/dgwmp.th.png (http://img42.imageshack.us/i/dgwmp.png/)

GraphEdit:

http://img132.imageshack.us/img132/3320/dgavs.th.png (http://img132.imageshack.us/i/dgavs.png/)

GraphStudio:

http://img132.imageshack.us/img132/8959/dgavs2.th.png (http://img132.imageshack.us/i/dgavs2.png/)

Thanks again:thanks:

[neuron2]

You've helped me a lot in the past, rica, so you know I'll get right on the case when you call. Thanks for pointing out the problem with the logs.

[rica]

Always pleasure :)

[cweb]

I believe somewhere on your system is an old version of something, probably something associated with the NVidia drivers.

Be aware that NVidia used to have a large control panel/control suite called nView that was bundled with older versions of their drivers. They removed nView support for GeForce cards in the drivers starting several months ago, nView is for Quadro cards only now.

If you had older nVidia drivers on the system when you upgraded to 190.38, nView may have stayed behind, and it will cause problems with the latest drivers. My recommendation:

1. Remove your nVidia drivers.
2. Remove nView if it is listed in the Add/Remove programs control panel.
3. Do a search on the Windows directory for all nVidia-related DLLs, especially nvcuvid.dll, and delete if any are found.
4. Delete C:\Program Files\nVidia if present.
5. Remove all versions of DG's NV tools from your system, including any DGNV-related DLLs present in C:\Program Files\Avisynth\plugins (like DGDecode.dll)
6. Restart and reinstall nVidia drivers 190.38.
7. Download and unzip the latest DGNV Beta 7 from neuron's site.
8. Try your steps again.

If this does not work, I'm not sure where else to go. Many other people here are running DGNV Tools beta 7 under XP SP3, Vista 32, Vista 64, and Win 7 with no issues.

Edit: Just saw in the other thread that you discovered that your GeForce card may not have sufficient hardware support for the NV Tools. If this turns out to be the case, please disregard.
My GTX has nview installed without issues on XP SP3. I don't use the desktop management features though.

[UsedUser]

I don't know where to start, so I'm starting here.

I'm trying to use DGMPGDecNV on Win 7 x64 with an x64 build of x264 (from x264.nl, built with AVS input support). I set up the job in MeGUI for DGMPGDecNV decoding and it runs fine with x264 x86 encoding, but can't find the AVS file with x264 x64.

Any known issues, i.e., with decoder/encoder bitness?

Job commandline: "C:\Program Files (x86)\megui\tools\x264\x264.exe"
--profile high --pass 1 --bitrate 3255 --stats "D:\Video\x.stats"
--level 4.1 --ref 1 --no-mixed-refs --b-adapt 2 --direct auto
--deblock -1:-1 --subme 2 --trellis 0 --partitions none
--vbv-bufsize 50000 --vbv-maxrate 50000 --me dia
--thread-input --sar 1:1 --output NUL "I:\Recordings\x.avs"

[Error] An error occurred: x264 [error]: could not open input file 'I:\Recordings\x.avs'

[UsedUser]

^ Nevermind - death by bitness - http://forum.doom9.org/showthread.php?t=148925&highlight=x64

[SomeJoe]

My GTX has nview installed without issues on XP SP3. I don't use the desktop management features though.

Yes, you're right, the latest 190.38 and 190.62 drivers seem to have put it back in.

I have had intermittent problems with nView's desktop management features, though, so I don't use that package like I used to.

[neuron2]

This thread is closed as DGMPGDecNV is superceded by DGNV tools 2.0.0.