DG NV tools

[Guest]

Now I am trying to reproduce the strange "license not found" bug. I am unable to find the "culprit", I think I'll make a video while importing with Camtasia Studio that will show when that bug jumps out. You opened more files than the program supports and I guess that a resulting memory overwrite caused the license error problem. Does it ever happen when loading less than the supported 512 files?

And, please, consider an automatic output of chapters.txt for multiple file input. It could save a lot of time when creating a multiple scene video. I don't know what a chapters.txt file is or what it should contain.

One question: why, while importing vob video from a cam generated minidvd, some shows an audio delay in the output ac3 and some not? If you provide stream samples, I can tell you.

Could be possible to use dgindex to realign the audio with the video instead of creating a delayed audio? It's possible and it's already on my to-do list.

[Guest]

What is the maximum possible number of files to index? I tried loading 946 MTS files (yes, it's a very big project) and got it crashed. It seems that the limit is about 500 files or slightly more. Adding to the file list in batches (with the "Add" button) doesn't help, it still crashes. Can you please rise up the bar to say 3000? I use DGNV exclusively for working with a camcorder footage, not for ripping BD's, and often have a lot of very short segments to work with because the camcorder produces a separate file on every record/stop. The max is currently 512. I'll increase it to 5000 in the next build.

[Guest]

Ever since I started editing my h264 .ts files with VideoRedo4 (currently in beta) I get a couple of warnings on opening the files:---------------------------
Warning
---------------------------
Found NALU type 13, len 2 undefined, ignore NALU, moving on

Hit Yes to continue notifying this specific error type; hit No to disable it.
Hit Cancel to suppress all further errors.

Is this anything to worry about? Life seems to go on OK if I ignore the warning. Do the VideoRedo guys need to be told to change something? Possibly. Can you please provide a link to one of these files so that I can analyze it?

[Guest]

question why does the crop automatically do increments of 4 most others i have used use increments of 2 which is more precise when cropping It's on my to-do list already.

[nautilus7]

Why not increments of 1 (with the restriction the result is mod2)? There are some cases where i need to crop 1px from one side and 1px from the other.

[TheProfosist]

Why not increments of 1 (with the restriction the result is mod2)? There are some cases where i need to crop 1px from one side and 1px from the other.

yes yes good idea ive never sen a program allow me to do that it would be really nice

[Asmodian]

They way color information is stored (at 1/2 resolution) gives the 2 pixel crop requirement. Look up the YV12 (or I420) color space for more information.

[hajj_3]

I presume DG NV tools doesn't allow you to use it without hardware acceleration for free? I know there were arguments about not including sourcecode for GPL or something for the free version and neuron stopped updating the free (non-NV) version, is this still the case or is there a way for this version to work for free without GPU support?

p.s the interface looks nice in these newer builds neuron, you'd added alot of support for new things over the past couple of months, keep up the great work :)

[Guest]

I presume DG NV tools doesn't allow you to use it without hardware acceleration for free? The only decoder in there is the Nvidia CUVID one, so there is no way to do anything without that.

Stopping DGAVCDec has more to do with the serious deficiencies of libavcodec, such as its dropping of good frames (a showstopper for accurate random access), its poor performance, its lack of a native Windows build, and its lack of support, than it does with LGPL issues, which were resolved in any case.

It may come back working with Diavc, however. I gave up on considering CoreAVC because after promising an API, they now appear opposed to the idea.

p.s the interface looks nice in these newer builds neuron, you'd added alot of support for new things over the past couple of months, keep up the great work :) Thanks. I am trying hard to give people their $15 worth. :)

[TheProfosist]

whats the easiest way to tell if the material you have is 3:2 pulldown? because MeGUI sets fieldop as 1 but i dont know if thats correct because idk hoow it came to that conclusion. Would MediaInfo tell me? or what line in the informatio window od DGindexNV would? i can post screenshots if necessary

like how can i tell the difference between something being interlaced and 3:2 pulldown?

[TheProfosist]

Thanks. I am trying hard to give people their $15 worth. :)

its definatly worth it so far just for the VC-1 compatibility not to metion everything else

[Guest]

whats the easiest way to tell if the material you have is 3:2 pulldown? That's not easy to answer, because you have hard and soft pulldown. If it is hard pulldown, then only stepping through the video and examining the fields will be reliable. See the first two questions here for how to examine your video:

http://neuron2.net/faq.html

If it is soft pulldown, then look at the coded and playback counts in the Info dialog. If you have playback = 1.25 * coded, then you have pure 3:2 pulldown. You can also look at the Film % field and if it is close to 100% then you have 3:2 pulldown. But the Film % is not yet implemented for AVC.

like how can i tell the difference between something being interlaced and 3:2 pulldown? See above.

[TheProfosist]

so from this picture i would use fieldop=1 instead of deintelace for TFF?

http://bayimg.com/NajBpaaCN

am i going to be more likely using fieldop=1 that deinterlacing? because previously i deintelace all content and didnt even know about 3:2 well i had heard about it but thats it.

and what would it look like if im supposed to leave it at 29.970 and deintelace instead


sorry about all the questions im knew to someof the ideas but so far yoiur programs is better and faster than what i was previously using

[tormento]

You opened more files than the program supports and I guess that a resulting memory overwrite caused the license error problem. Does it ever happen when loading less than the supported 512 files?
The same thing does not happen if I put all the files in a single directory. Strange enough!
I don't know what a chapters.txt file is or what it should contain.
It is something likeCHAPTER01=00:00:00.000
CHAPTER01NAME=Living Weapon
CHAPTER02=00:05:00.750
CHAPTER02NAME=A Better World
CHAPTER03=00:09:31.937
CHAPTER03NAME=Aboard Serenity (Main Titles)
CHAPTER04=00:14:49.088
CHAPTER04NAME=Going for a Rideand so on. It is used in many container muxer and, to me, every single file imported should give a new chapter, at least for the first release. The best thing should be the possibility to "collapse" many files in a single chapter at will, giving full edit capability.
If you provide stream samples, I can tell you.
I'm trying to catch the files you need.
It's possible and it's already on my to-do list.
Thanks! ;)

[Guest]

@TheProfosist

That's pure soft 3:2 pulldown, so you need fieldop=1.

You can see that the film % is 99.99.

It's all covered in the documentation.

[TheProfosist]

@TheProfosist

That's pure soft 3:2 pulldown, so you need fieldop=1.

You can see that the film % is 99.99.

It's all covered in the documentation.

ok thanks i just want to make sure because im am going to have to re-encode a ton a material that that should have been applied to but only a de-interlacer was applied

i will report back with results

thanks again

[Guest]

im am going to have to re-encode a ton a material that that should have been applied to but only a de-interlacer was applied Where did you get that material and what is it?

[TheProfosist]

Where did you get that material and what is it?

so the content is from DVD, BD, or RAW(high quality/bitrate encodes that use a range of formats)

for most 29.970i content i was just deinterlacing using TDeint and after indexing a few of the DVD i have found that I should have reversed the pulldown instead.

[Efenstor]

The max is currently 512. I'll increase it to 5000 in the next build.

Thanks! :)

[TheProfosist]

after reversing the pulldown there is stil some lines left over do i need to run a deinterlacer as well? if so can this be done using DG NV tools? a picture is below
http://bayimg.com/GAjdnaAcN and http://bayimg.com/gAJdoAacN

sorry about the weird resolution the program was acting weird the video is fine resolution wise 616x480 no black borders at all

[MOS-Marauder]

Somebody recently posted in a MeGUI thread using only 3 instead of 4 memory slots helped making NV tools work.

I'll have a card shortly, it will be:

G512P Sparkle (R) 9400GT 512MB DVI/TV

I'll test it on Win 7 Ultimate 32bit, I might also be able to test it on XP since I should still have my old XP partition. If it's related to the memory on the card, how about then testing source files with smaller resolution?

It was me ;) U had 4 x 2 GB Ram in my 4 Slots and alway got Driver Crashes. Removing one helped me.
(Win7 64 Bit)
The Resolution doesnt matter in my case...it happened on SD and HD Sizes.

Chris

[Guest]

well i encode anime. go to my website link should be in my sig.
so the content is from DVD, BD, or RAW(high quality/bitrate encodes that use a range of formats) I asked where you got the file and you point me to your site pushing torrents! Struck for rule 6 and signature link deleted. Do not restore it.

[Alf Bundy]

Hi,

I have a problem with DGindexNV 2003 (I don't know with previous versions, I only have my license since yesterday).

I tried my first encode using DGindexNV, but the result ended out of sync.
But not a constant or progressive delay.
At some points in the episode, dialogue is suddenly cutted and audio continues further, generating an instant few seconds delay with the video.
This problem is not present in the source files (MPG from PVR = Mpeg2 + MPA 48Khz stereo 320kbps .mpg).

So I tried using the free DGindex to index my source files and I discovered something strange :

With DGindex, demuxed audio is 128.519 Kb for 54 min 50 sec. (wich is correct and does not produce sync problem.)
With DGindexNV, demuxed audio is 127.650 Kb for 54 min 27 sec.

I have 3 different episodes, and the same problem happens all the time with DGindexNV.


Is it possible that DGindexNV 2003 wrongly demux MPA ?

[Guest]

Is it possible that DGindexNV 2003 wrongly demux MPA ? Not likely but possible. Please post a link to a sample file such that when demuxed with DGIndex and DGIndexNV audio streams of different lengths are produced.

I just did a test demux of a DVD and the audio streams demuxed by DGIndexNV and DGIndex were binary identical.

[TheProfosist]

I asked where you got the file and you point me to your site pushing torrents! Struck for rule 6 and signature link deleted. Do not restore it.

my site is just a blog i blog on it yes details of my release are on my blog because its my blog sorry. but i encode initially for my self and my personal use. and i had the link im my sig becuase its my site/blog no different than others having links to their sites or blogs.

i wasnt trying to push torrents on u just to show you what i had previously encoded and how it was done. ignore the dl links and i wont put the link in my sig its not all that important i get enough stuff daily already (not because of that exact thing but the overall situation)



but for source it comes mostly from DVD (usually R1 or R2J) or BD (usually RA or RAJ) and 99% of it is anime.

what i used to do was just deinterlace with TDeint(custom settings or cource) but after indexing a few more dvds i foiund that i should reverse the pulldown instead



after reversing the puldown usiong fieldop=1 after reversing there is stil some lines left over do i need to run a deinterlacer as well? if so can this be done using DG NV tools? a picture is below
http://bayimg.com/GAjdnaAcN and http://bayimg.com/gAJdoAacN

sorry about the weird resolution of the pics the program was acting weird the video is fine resolution wise 616x480 no black borders at all

AVS script:
LoadPlugin("C:\Program Files (x86)\megui\tools\dgindexnv\DGDecodeNV.dll")
DGSource("C:\TEMP\Lain\07 - Cvt Vid\01.track_224-002_Track1.dgi",resize_w=616,resize_h=480,fieldop=1)



once again sorry about the sig link im not trying to offend you or go against site rules mabye im just nieve? but neverthe less im greatful for your and the communities continued help

[Guest]

@Profosist

Stop asking about this stream or you'll get a rule 16 strike and be suspended. Also, followup about rules to PM or you can get a rule 3 strike.

[Alf Bundy]

Not likely but possible. Please post a link to a sample file such that when demuxed with DGIndex and DGIndexNV audio streams of different lengths are produced.

Here is a sample (100Mb - 1min 36") that produces different lenght and size audio when demuxed with DGindex or DGindexNV :

http://dl.dropbox.com/u/2137721/0120133335Composite_1.mpg

http://dl.dropbox.com/u/2137721/DGindex_vs_DGNV.png

Audio lenght when demuxed with DGindex = 1min 36" (like the source)
Audio lenght when demuxed with DGindexNV = 1min 29".

This sample was cutted from a 1Gb file with DGsplit 1.0
I can post more samples if you need.


I made 2 scripts to test this sample (scripts played with MPC-HC).

DGindex :
video=DGDecode_MPEG2Source("E:\Recode\DGindex.d2v",cpu=4)
audio=MPASource("E:\Recode\DGindex Tc0 L2 2ch 48 320 DELAY -80ms.mp2")

AudioDub(video,audio)
DelayAudio(-0.080)
ConvertToYV12() -> Sync is OK like in the source sample.

DGindexNV :
video=DGMultiSource("E:\Recode\DG NV.dgi")
audio=MPASource("E:\Recode\DG NV Tc0 L2 2ch 48 320 DELAY -80ms.mp2")

AudioDub(video,audio)
DelayAudio(-0.080)
ConvertToYV12() -> at 1min 21", audio "jumps" and the rest is out of sync.


About the video, both indexers always produce the same amount of frames.

[Guest]

@Alf Bundy

That's a great problem report. I wish all were as complete. I'll investigate and report back.

[meatwad]

@neuron2

i got good news to report DGIndexNV IS WORKING WITH StaxRip 1.1.4.1 (bit.ly/6FxoaN)

settings
http://img683.imageshack.us/img683/7853/ss20100114163647.th.png (http://img683.imageshack.us/img683/7853/ss20100114163647.png)
http://img683.imageshack.us/img683/4734/ss20100114163751.th.png (http://img683.imageshack.us/img683/4734/ss20100114163751.png)
http://img254.imageshack.us/img254/8746/ss20100114163029.th.png (http://img254.imageshack.us/img254/8746/ss20100114163029.png)
http://img69.imageshack.us/img69/7998/ss20100114163135.th.png (http://img69.imageshack.us/img69/7998/ss20100114163135.png)
http://img69.imageshack.us/img69/3491/ss20100114163228.th.png (http://img69.imageshack.us/img69/3491/ss20100114163228.png)

Hi, I followed all of the steps as indicated in the images, but StaxRip 1.1.4.1 crashes after the indexing is done. It will still work as long as I don't close the crash window. It's crashing the nvcuda.dll:

Problem signature:
Problem Event Name: BEX
Application Name: StaxRip.exe
Application Version: 1.1.4.1
Application Timestamp: 4b4b9949
Fault Module Name: nvcuda.dll_unloaded
Fault Module Version: 0.0.0.0
Fault Module Timestamp: 4b4c13da
Exception Offset: 0c8143e9
Exception Code: c0000005
Exception Data: 00000008
OS Version: 6.1.7600.2.0.0.256.1
Locale ID: 1033
Additional Information 1: 0a9e
Additional Information 2: 0a9e372d3b4ad19135b953a78882e789
Additional Information 3: 0a9e
Additional Information 4: 0a9e372d3b4ad19135b953a78882e789


Sometimes it doesn't say nvcuda.dll unloaded and instead gives the version number. Also, after it's done encoding the video it will crash again and you're forced to exit the program. My OS is Windows 7 Ult. 64bit and I wonder if that's why I'm getting this error? Have you tested it with a 64 bit OS yet?

[stax76]

@meatwad

DGSource support was added in test build 1.1.4.3 (http://forum.doom9.org/showthread.php?p=1364661#post1364661) so you need a newer build.

[meatwad]

Thanks! I'll give that a try.

Edit: It seems to work just fine. Great job!

[Lynx Pardinus]

They way color information is stored (at 1/2 resolution) gives the 2 pixel crop requirement. Look up the YV12 (or I420) color space for more information.

When dealing with MPEG2 4:2:2 streams you can crop vertically in 1 pixel increments and the resolution after cropping doesn't need to be even.

DGIndex supported this kind of streams, I don't know if DGNVTools does.

[Asmodian]

When dealing with MPEG2 4:2:2 streams you can crop vertically in 1 pixel increments and the resolution after cropping doesn't need to be even.

DGIndex supported this kind of streams, I don't know if DGNVTools does.

DGNVTools does support MPEG2 and yes you can crop vertically in 1 pixel increments in the YUY2 color space but NVTools are really for H.264 AVC & VC1 where you cannot. ;)

[Guest]

* Increase max files to 5000.

* Add DGMultiDecodeNV Manual item to Help menu.

* Fix bug in the parsing architecture. This fixes Alf Bundy's issue. Thanks for pointing that out, Alf!

* Fix film % reporting for MPG and VC1.

http://neuron2.net/dgdecnv/dgdecnv.html

[Guest]

Please re-download as I had a bogus version up there for a while.

[Alf Bundy]

Build 2004 works fine with my problematic files. Thank you for fixing that so fast. :thanks:


Now I have a probably stupid question :

May I watch a movie using MPC-HC's h264 DXVA internal decoder while DGmultisource() is in use for an encode ?
Could DXVA playback corrupt DG NV decoding process ?

[Guest]

May I watch a movie using MPC-HC's h264 DXVA internal decoder while DGmultisource() is in use for an encode ?
Could DXVA playback corrupt DG NV decoding process ? I don't support that.

[Guest]

* Replace GPU->CPU async memory copy with sync copy. Sync copy is faster because it does not require a wait loop with a sleep. You should see higher max fps figures when doing a play/preview in DGIndexNV (with playback speed set to maximum). Note that the decode DLLs already used sync copy so they were already serving video as fast as possible.

http://neuron2.net/dgdecnv/dgdecnv.html

[stax76]

@neuron2

Is it OK to kill the CUVIDServer process? If not how do I terminate it then? In case the window is created hidden on startup I could send a close message, is it? Is there stdout status yet?

[Alf Bundy]

Could DXVA playback corrupt DG NV decoding process ?
I don't support that.

I did try myself.

DGmultiSource() decoding VC-1 ES for an encode, and watching a movie using DXVA decoder.

No crash, nothing wrong.
No skipped or corrupted frames in the encode output and playback was OK.

(NVidia 9800GTX+ 512Mb)

So I answer my own question : No, DXVA playback doesn't corrupt DG NV decoding. :D


Big thanks, Neuron2.
With DG NV tools I get 5 to 20% speed up encodes compared to DGindex and DGAVCindex on my C2D 7400. http://img290.echo.cx/img290/9768/yahoo6lp.gif

[Clumpco]

With version 2005 I am getting random server crashes.
Opening video in the indexer gave me an error 100. Closed, restarted server and all was OK
Start of MeGui second pass got an error 400. Aborted pass, restarted server and the 2nd pass and all seems OK.
This is with Win7 32 bit and Nvidia drivers 8.17.11.9621 (196.21).

[TheProfosist]

i just started a new dvd since im waiting on help on Lain. I just picked up Moribito today but i can figure out what to do with it because MediaInfo says :
Frame rate : 29.970 fps
Standard : NTSC
Colorimetry : 4:2:0
Scan type : Progressive
Scan order : Top Field First

and i stick through DGindexNV and get a film percentage of 63.6%

what does this all mean because progressive film cant be interlaced and how do i intpret the 63.6%

log files from both mediainfo and DGindexNV
http://www.mediafire.com/file/wiidf0cdvfz/VTS_02_2.VOB_DGindexNV.txt
http://www.mediafire.com/file/yyygl253khl/VTS_02_2_MediaInfo.log

[Guest]

With version 2005 I am getting random server crashes. All streams or just this one?

[Guest]

@Profosist

This thread is really for bug reports and feature requests for my tools, not for asking noob questions about video processing. I'll help you if you post in the proper thread/forum.

[TheProfosist]

@Profosist

This thread is really for bug reports and feature requests for my tools, not for asking noob questions about video processing. I'll help you if you post in the proper thread/forum.

ok ill create a thread arround this program for noobs like me
if there already is one for your programs tell me and ill go there instead and you can close mine sorry about the questions then jswiting to this program is just hard for me for some reason

[Clumpco]

All streams or just this one?

I don't know yet. Re-running the same encode to see what happens second time around.

ETA: Seems to have gone fine, not a single error. BTW the computer hasn't been rebooted.
Very strange. I will keep an eye on things.

[Sharc]

@neuron2
Thank you for the excellent documentation which comes with DGNV tools.
Just a minor remark: In the DGIndexNV User Manual/Information Panel/Video Section [MPG] the explanation for 'Frame Coding' is missing (should come after 'Frame Type').

[Guest]

@Sharc

Thanks for pointing that out. I added this:

Frame Coding - Displays "Progressive" if the progressive_frame syntax element is set, otherwise "Interlaced" is displayed. Do not make the mistake of assuming that this tells you whether the video content itself is progressive or interlaced! What determines that is whether there is motion at the field rate. This Info dialog field tells you only how the video was encoded; it is quite common to encode progressive content as interlaced.

[sangwich]

I'm using 2005 with megui and I have strange problem. If i use this syntax:

LoadPlugin("DGDecodeNV.dll")
DGSource("movie.dgi", fieldop=0)
crop( 0, 140, 0, -140)
LanczosResize(1280,528)

everything works fine, however if I try to do resizing with DGDecode like this:

LoadPlugin("DGDecodeNV.dll")
DGSource("movie.dgi", fieldop=0, resize_w=1280, resize_h=720)
crop( 0, 96, 0, -96)

the processing gets stuck at 99.96% and hangs there indefinitely. Does anyone has same problem or is it just me?

[MuLTiTaSK]

@neuron2

might be a shot in the dark or feature request but here we go if i dont crop in DGIndexNV would it be possible to add crop values after saving the .dgi?

example:
DGSource("*.dgi", deinterlace=0, use_top_field=true, use_pf=false, l=0, t=140, r=0, b=140, resize_w=1280, resize_h=528, fieldop=0)