Auto Gordian Knot: current version 2.55

[Taurus]

Originally posted by len0x
Weird though - lanczos4 is softer that lanczos3


This was the biggest surprise I got from my comparison.
And why are there sometimes the numbers changing?
In a mathematical model they should be all in a straight order,
So lanczos, lanczos4, lanczos, Lanzos4,...... or vice versa.

I will do some more testing, following the suggestions made by the
removegrain author.
Normally we're just in it for the fun, but this is real work.
Even by using some batches and registry hacks, it's a time consuming
thing.
To me it is no wonder, over at the matrice comparision thread, only a few members got involved, caus it's a painstaking procedure to get all the numbers counted.
Anyway, hope this was on any help.

[len0x]

Originally posted by Corence
Would it be possible to have some sort of visual notation showing if ESS compatibility is enabled after the installation?

Well, that possible, however you can already check that by looking for .ess file in AutoGK folder. If its there then ess mode is enabled.

[Corence]

I must have missed that. Thanks :)

[therealjoeblow]

Originally posted by len0x
I'm playing with a new denoiser for next version of AutoGK (that would be RemoveGrain in mode 2). For speed reasons and actually compressibility I'd like to move denoiser after resizing (which will become sharper with new Lanczos4Resize as well). Anyone here thinks its not a good idea or everybody's happy? :)

I never tried RemoveGrain, but using AutoGK Tweaker, I did some experimentation with Deen() on Breakfast at Tiffany's, which is a *very* noisy/grainy source, and the results were very impressive. AutoGK 1.60 on it's own was only getting around Q=40% @ 640xZZZ (sorry, logs aren't handy so numbers are rounded off). I inserted the extra filter both before and after Undot() with the following ersults: With Deen() before Undot(), quality jumped to somewhere around 70% for the same resolution, and Deen() after Undot() was around 65%, both significantly better than without the extra filter, and the visual results looked really good too.

Q: If any particular denoiser (or other simple filter, for that matter) is added to the script, as long as it is *always* added to every script at *exactly* the same place (eg. comptest, first pass, second pass), does that have any bearing on AutoGK's builtin processing logic?

I guess what I'm asking, is, could you not allow for (at least a few) additional filters, particularly denoisers, to be user-specified as a hidden option, and add those to the chain in the processing scripts? That's been my only hangup about AutoGK for a long time - it doesn't deal with noisy sources very well, so I've always had to go for a much higher bitrate to compensate.

Adding the filter as noted above seemed to work very well for me with my tests on that one movie (I also tried Convolution3D and a couple of others, which did *not* have as good a result as Deen() did). My gut reaction is that the impact of consistently adding one additional filter should be largely transparent to AutoGK, it sould just receive a different result from the comptest than it normally would, and go on with it's regular decision making regarding resolution, etc based on that. I don't know for sure, you're the author and are privy to the inner workings, but it seems simple to me.

Whatever you do would be a welcome addition, but if you're looking for input, my vote would go to Deen() as an additional denoiser. It would also be nice to allow for flexibility in where to place it, but that would definately count as an expert feature, requiring multiple encodes to get it right.

[len0x]

@therealjoeblow

No, you might be just lucky, but normally if there are quite a few encoding parameters changed (like matrix/resize/bframes) then comp test results can be thrown out of the window if you put heavy denoisers (resolution is adjusted based on precalculated empirical values that are tied up to avs script/codec parameter: i.e. switching matrix can give you 15% gain with undot and 30% with Deen). So there will never be support for more than just one default denosiser (which i don't think of as a denoiser, but just as simple cleaner). So we should not really discuss it any further :)

My question was only: before or after resize...

[richarddd]

I have a movie on 2 dvd disks that I'd like to convert to 3 CDs (2100 Mb) with AutoGK. Can I do this (other than by converting 1 dvd to 2 CDS and the other DVD to 1 CD)?

There's a reference in the FAQ to putting all files in one directory and renaming to *_.1*; *_.2*; etc. If this is the way to do it, what do I do about the two stream info.txt and .ifo files (one each from each disk)? For example, if I have VTS_02 - Stream Information.txt, VTS_02_0.IFO, VTS_02_1.VOB, VTS_02.2.VOB, VTS_10 - Stream Information.txt, VTS_10_0.IFO, VTS_10_1.VOB and VTS_10.2.VOB, how would I rename?

Apologies if I missed a post where this was covered.

[SNP]

any plan to add selectable sub in AGK? As in adding a browse button to select pre-ripped subs. This will be VERY useful esp in encoding anime as many anime DVDs have multiple eps in one single VTS. I know I can do that now by slipping in my own sub during the comp test phase but with that I have to constantly babysit the encoding if I'm doing multiple eps.

Thanks.

[ydobon]

@richarddd:

The reference you mention is for "file mode", not "DVD mode". And about the non-VOB files I think I remember Len0x saying that the .IFO files are only needed for the subtitles. So if you don't want subtitles, you won't need them.

About the stream info file, you can use just one of them if you're lucky and the streams in both are the same. If not, I don't think you can do it.

Other than that you can always do two 1050 MB AVIs and split at will. I know, not very pretty, and you can't be sure if you'll be able to join the end of the first with the beginning of the second, but hey, it's doable.

@SNP:

Why do you pre-rip the subtitles? And if you do, why not simply use them as external ones?


Regards,

[ydobon]

Hi, Len0x.

Originally posted by len0x
_.avs file is used only for analysis, nothing to do with comp test and resolution.
A follow-up question: When AutoGK says "Found 63667 frames" Where does it get that number from? The .d2v file? Also, do you use that number of frames to determine the chosen resolution?


Thanks in advance.

[Taurus]

Originally posted by len0x
My question was only: before or after resize...

I surely missed that part of your question a few numbers ahead.
Will do some more testing on the subject to see the behaviour regarding size, quality and speed.
Are you still interested in the results?

Cheers

Taurus

[len0x]

Originally posted by ydobon
A follow-up question: When AutoGK says "Found 63667 frames" Where does it get that number from? The .d2v file?


not only, its resulting number of frames (i.e. after IVTC etc)

Originally posted by ydobon

Also, do you use that number of frames to determine the chosen resolution?


of course

[len0x]

Originally posted by Taurus
Will do some more testing on the subject to see the behaviour regarding size, quality and speed.
Are you still interested in the results?


Yeah, but I know size and speed factors already (its only obvious that putting denoiser after resize you get better compressibility and faster speed). I'm very interested in subjective opinion about quality of the output. I myself would like to preserve as much visible details as possible (even if its grain). That's why I'm also interested to see how much difference is between mode 1 and 2 of removegrain (mode1=undot) you can see with your eyes.

[Taurus]

Originally posted by len0x
(its only obvious that putting denoiser after resize you get better compressibility and faster speed)

That's the same I found out.
Will do some more testing on visual appearance in the evening.
Mode=1 and Mode=2 compared to undot.
But I have to look for a more suitable source first, caus the last one used
is so grainy and bad (Buffy2). I have to search for another one for that job.

Cheers

Taurus

[ydobon]

Hi,
Originally posted by len0x
of course
Agh. Then I'm busted. The tweak I've done to avoid cutting my .ts files results in lower resolution the bigger is the cut :( :( :(

Well, at least I'm able to cut individual frames...

On a side note, with so many tests I noticed that if you have "no audio track" selected, AutoGK selects the first audio track when you open a new source. Is it a bug or a feature? :)


Thanks for the patience :D

[therealjoeblow]

Originally posted by richarddd
I have a movie on 2 dvd disks that I'd like to convert to 3 CDs (2100 Mb) with AutoGK. Can I do this (other than by converting 1 dvd to 2 CDS and the other DVD to 1 CD)?

There's a reference in the FAQ to putting all files in one directory and renaming to *_.1*; *_.2*; etc. If this is the way to do it, what do I do about the two stream info.txt and .ifo files (one each from each disk)? For example, if I have VTS_02 - Stream Information.txt, VTS_02_0.IFO, VTS_02_1.VOB, VTS_02.2.VOB, VTS_10 - Stream Information.txt, VTS_10_0.IFO, VTS_10_1.VOB and VTS_10.2.VOB, how would I rename?

Apologies if I missed a post where this was covered.

This is the way I do it:

you started with:

[files in DIR1]
VTS_02 - Stream Information.txt --> leave as is in DIR1
VTS_02_0.IFO --> leave as is in DIR1
VTS_02_1.VOB --> leave as is in DIR1
VTS_02.2.VOB --> leave as is in DIR1

[files in DIR2]
VTS_10 - Stream Information.txt --> ignore this file
VTS_10_0.IFO --> ignore this file
VTS_10_1.VOB --> rename to VTS_02.3.VOB and copy to DIR1
VTS_10.2.VOB --> rename to VTS_02.4.VOB and copy to DIR1

so now you have:

[files in DIR1]
VTS_02 - Stream Information.txt
VTS_02_0.IFO
VTS_02_1.VOB
VTS_02.2.VOB
VTS_02.3.VOB
VTS_02.4.VOB

[files in DIR2, which you can ignore]
VTS_10 - Stream Information.txt
VTS_10_0.IFO


You should be good to encode from DIR1 now for most 2 disc movies (I understand from others that there are a few oddballs that will give you problems with this method, but all I've tried worked correctly).

NOTE: the only major issue is that you will not be able to encode subs automatically this way, because the stream info for the 2nd disc isn't used in the process, so the subs would be screwed up. If you want subs, BEFORE you do the rename/move -> you should run a 'bogus' AutoGK session on each disc individuallty with any arbitrary video and audio settings, but with the correct subtitle settings for what you want. After the subs are indexed and rar'ed, you can then abort that session (as I said, do for both DVD discs, to produce proper subs for each one). Then SAVE THE SUB FILES SOMEWHERE!, and now you can carry on and do the rename/move bit, run the full encode, setting your target file size to take into account the size of the subs. After the encode, you'll then need to use the directvobsub tools to:

1) recombine your 2 separate sub files into 1 to match the overall length of time for the whole 2 DVD movie, and the
2) split the recombined sub file into separate chunks for each of your 3 CD's

The subs are a bit of a pain for multi-DVD movies, no doubt.

[Taurus]

Originally posted by len0x
That's why I'm also interested to see how much difference is between mode 1 and 2 of removegrain (mode1=undot) you can see with your eyes.

Sorry for the delay. Daytime job.
Well after looking for a new source for my comparision, I finaly decided to take another Buffy episode. This time Buffy 5.
It's much cleaner than the older episodes. PAL 16:9 25fr
Made an AVS scipt using AutoGk. 720*416 Resolution. Xvid Q:2 as used by AutGk. No more filtering; degrainers at the end of the scipt.
Mode=8 just for testing(incredible)(lanczos_undot for comparision)
-----------------------------------------------------
Buffy_lanczos4_removegrain_mode8.avi----42,252 MBytes
Buffy_lanczos4_removegrain_mode2.avi----43,023 MBytes
Buffy_lanczos4_removegrain_mode1.avi----44,537 MBytes
Buffy_lanczos4_undot.avi----------------44,605 MBytes
Buffy_lanczos_undot.avi-----------------44,998 MBytes
Buffy_lanczos4.avi----------------------45,371 MBytes
Buffy_lanczos.avi-----------------------45,604 MBytes
-----------------------------------------------------
So here are my subjective conclusions:

On a phosphor display it's hard to see any difference between undot and removegrain Mode=1&2. Even strong zooming into the clip does not show remarkable difference. If you want to stay as conservative as possible, my opinion is: use Mode=1. If you want some extra encoding headroom,
Mode=2 is the one. As I wrote, to my eyes both look almost identical,
even at zooming into the pictures.
But I think still undot is doing a fairly good job compared to a non filtered file.

But the real headshaker is Mode=8. At this greasy source it dramatical changed necessary bitrate and quality(great).
But this was just for fun:D

Hope this is on any help.

Cheers

Taurus

[len0x]

Originally posted by ydobon
On a side note, with so many tests I noticed that if you have "no audio track" selected, AutoGK selects the first audio track when you open a new source. Is it a bug or a feature? :)


May be both in a sense that other behaviour was never requested :)

[len0x]

Originally posted by Taurus
But the real headshaker is Mode=8. At this greasy source it dramatical changed necessary bitrate and quality(great).


Really? What is so dramatical - size went down just 0.8M, when mode 2 gave 1.5Mb less than mode 1...

[Taurus]

Originally posted by len0x
Really? What is so dramatical - size went down just 0.8M, when mode 2 gave 1.5Mb less than mode 1...
You're right...
Just the visual appearance. A little wax on human skin but the overall
impression was real natural. It all depends on the source.
I would NEVER use a degrainer on clean sources. Purist.
But this is another story.
P.S.: Just compared a few samples lanczos versus lanczos4.
Strange, the filesize is less on lanczos4.
And I can't tell a difference in visuality:rolleyes:

Have a good day (night)wherever you are.

[len0x]

Originally posted by Taurus
P.S.: Just compared a few samples lanczos versus lanczos4.
Strange, the filesize is less on lanczos4.
And I can't tell a difference in visuality:rolleyes:


ok, lets proceed with lanczos4 and see where it gets us :)

[richarddd]

Originally posted by therealjoeblow
This is the way I do it:

Thanks. That seems to work.

As a bit of background, I was ripping Lawrence of Arabia, which is approx 2 hours on disk 1 and 1 hour on disk 2, and having sync problems on the second disk. After some testing and searching this forum, I figured out that the problem was caused by the intermission included at the beginning of disk 2 (blank screen & silence followed by music). If I unchecked all the intermission chapters (about 20 chapters of about one second each and about 5 chapters of about one minute each) before ripping, all is well. Doing everything as one job resulted in approximately equal quality for each CD, which is what I wanted.

[haubrija]

I have a quick question (apologize if its been asked before)

I've been using AutoGK to archive my .ts streams recorded from a Fusion3 card. I edit out commercials with HDTV2MPEG2 before.

I had been told that ProjectX would help me sorten out some Audio Sync errors I had been having.

I edit the stream with HDTV2MPEG2 first. I then loaded the .ts into ProjectX. I chose to save it to a .ts stream.

When ProjectX is done doing its thing, I attempt to load the resulting .ts into AutoGK. For some reason the audio is not read after running it though ProjectX.

It looks like the PID have been changed. What was once 0x21 and 0x24 (audio and video respectively) have become 0xE0 and 0xFFFFF. Is there something I've been doing wrong?

Thanks in advance.

[TreborPugly]

Putting the files on CD.

I've looked over the How-to guides for Auto GK, and some other DivX creation tools. Nowhere do I see anything about the format of CD's you make for stand-alone devices. Do you just burn data CD's, with the *.avi files on them and some stand-alone DVD players will recognize the AVI's? Or do you need to create something like VCD's?

And are there any stand-alones that will recognize DVD's with DivX files on them? For example, I like the idea of creating one DVD with several kid's DVD's worth of shows on them, reducing the number of disks I need to have available to the kids.

[btothec2]

Yes,

You just burn them as data cd's. And almost all f not all standalone Divx players will allow you to burn several avi's on on dvd.

B


Originally posted by TreborPugly
Putting the files on CD.

I've looked over the How-to guides for Auto GK, and some other DivX creation tools. Nowhere do I see anything about the format of CD's you make for stand-alone devices. Do you just burn data CD's, with the *.avi files on them and some stand-alone DVD players will recognize the AVI's? Or do you need to create something like VCD's?

And are there any stand-alones that will recognize DVD's with DivX files on them? For example, I like the idea of creating one DVD with several kid's DVD's worth of shows on them, reducing the number of disks I need to have available to the kids.

[Carraway]

Originally posted by Carraway
I recall a while ago that there was talk that AutoGK was producing dark encodes. [...] I have several just generally dark sources [...] how feasible (or infeasible!) would a .filters file be, in which you could add internal avisynth filters like tweak(bright=xx) or levels(1,1.x,255,0,255) that would then be appended to the end of the comptest and movie avs files?

Originally posted by len0x
wait for config file support (which is at least 1 month away from next stable release date).

Two questions:

First, len0x: any update on support for levels() and tweak() to brighten unusually dark Xvid encodes?

Second, anyone can answer, but len0x will probably know best: if avisynth filters are appended to the end of both the movie avs files as well as the comptest avs files, will that still throw off the values of the comptest?

EDIT:

About the second question, by "appended to the end of the avs file" I of course mean inserting them immediately prior to the resizing. Applying filters after resizing would change their use in the comptest since the comptest uses a different resolution than the movie.avs file would.

[len0x]

Originally posted by haubrija
It looks like the PID have been changed. What was once 0x21 and 0x24 (audio and video respectively) have become 0xE0 and 0xFFFFF.

that can be a problem. AutoGK and its applications only support four digits PIDs, so FFFFF may not work. You can assign another PID manually in options of ProjectX, I believe.

[len0x]

Originally posted by Carraway

First, len0x: any update on support for levels() and tweak() to brighten unusually dark Xvid encodes?


no

Originally posted by Carraway

About the second question, by "appended to the end of the avs file" I of course mean inserting them immediately prior to the resizing. Applying filters after resizing would change their use in the comptest since the comptest uses a different resolution than the movie.avs file would.

There is absolutely no difference for comp test where filters are...
Please read my reply to therealjoeblow on the previous page to see what is affecting comp test for AutoGK's purpose (basically everything, but noise filters the most).

P.S. by comp test I mean resizing algorithm based on comp test and used in AutoGK.

[len0x]

As seen in today's news in www.doom.org jdobbs came up with an interesting idea of distributing an application to the donators first and some time later to the public. That is very appealing idea for me - what does the public thinks?

[bourtzovlakas]

My opinion is,that AutoGK is your program,your creation,your "child"!Therefore,you are justified to distribute it however you like....
But,I have one question.....How can you prevent a donator to release it to the public,before you do....
Anyway,if you choose to distribute AutoGK like this....I will gladly donate an amount,for no other reason,but to akwnoledge and support your efforts....

[Mc|Atm0s]

How about new XviD-1.1? :cool:

[len0x]

only when it gets released

[len0x]

Originally posted by bourtzovlakas
But,I have one question.....How can you prevent a donator to release it to the public,before you do....


Its more of a moral issue though - why would someone who supports the software and donates money would try to piss off the developer by leaking it outside?

P.S. there are ways of tracking who's copy is being distributed, but its mainly a question of - is it worth going through all that trouble to find that out or not.

[len0x]

Very experimental and I might not have time to fix bugs quickly these days. I would be happy to get this stable in a month time, so that before christmas I manage to get DV/AVI support in (but I will only start that once everything else new is working OK). Only after that I will consider rewriting hidden options support that should conclude version 2.0. Only after that hacking of AVS scripts should be possible (i.e. post 2.0 feature). And I'm still not sure what kind of distribution policy I will have by then.

[ydobon]

Hi, Len0x.

Originally posted by len0x
[...] interesting idea of distributing an application to the donators first and some time later to the public. That is very appealing idea for me - what does the public thinks?
First, the software is yours and you can do whatever you want.

But I don't like the idea. Maybe I'm being picky but doing something like that would convert the donation scheme in a priority pricing scheme. Want to have it first? Gimme money :(

To be honest and to profile myself, I've never donated any money to any software author. I've contributed with opinions, testing, translations and even some chunks of code, but no money.

I completely agree with your source code control policy. Rip-off packages are disgusting and I think keeping the source and avoiding CLI are appropiated measures. But messing with the concept of donation... bad move.


Again, it's just my opinion.

[len0x]

You probably don't realize but I'm starting to think that the idea of "forcing" ppl to donate is only happening because quite a few of them are ready to do so, but don't because its not needed (its quite suprising but I saw many times ppl saying that they would gladly buy AutoGK if they could - amazing fact it is, especially in the light of not so many donations). Also it might be the case that donations are considered to be not worthy - buying is much better justification of money spending.

Personally I think there are better methods of giving attention to the fact that you can donate (like startup splash windows etc), but I'm just too nice to put that all that stuff in there. And the only reason I don't like "donate for first access" options is that its too difficult to develop something that you release very often and it becomes a mess until you have your own private forum with lots of ppl inside it to make quick bugfixes.

Anyway it all comes from the fact that motivation is these projects tend to go away over time. I think I reached the point when AutoGK is almost everything it is supposed to be originally (this is not saying there is not much to do though) and most ppl are happy with it. So one day might wake up with a question: "Do I really need all this? Don't I have something else to do? It's time to move on!". And financial motivation can be the only one that may postpone this morning... And I already see that moment from time to time :)

[Sharktooth]

Wow, the sharp matrix is now EQM V2.
Nice :)
Did you do any tests?

[SNP]

Hi len0x, I posted this a couple pages ago, any suggestion or comment?

Originally posted by SNP
any plan to add selectable sub in AGK? As in adding a browse button to select pre-ripped subs. This will be VERY useful esp in encoding anime as many anime DVDs have multiple eps in one single VTS. I know I can do that now by slipping in my own sub during the comp test phase but with that I have to constantly babysit the encoding if I'm doing multiple eps.

Thanks.

[Taurus]

Thanks for the new release.
cos of deleting accidentally cookies in IE I had to do my way back to the AutoGk thread manually:)
but now I'm back on the tracks.
Will give Sharktooth's EQM V2 a try..

Cheers

Taurus

[Corence]

With the ESS support enabled, does the change to sharp matrix EQM V2 make any difference? Or are those of us with ESS based standalones not receiving the benefits of such changes?

[len0x]

Originally posted by SNP
Hi len0x, I posted this a couple pages ago, any suggestion or comment?

I'm still thinking about it

[len0x]

Originally posted by Corence
With the ESS support enabled, does the change to sharp matrix EQM V2 make any difference?

No, ESS mode doesn't use any of custom matrixes.

[len0x]

Originally posted by Sharktooth
Wow, the sharp matrix is now EQM V2.
Nice :)
Did you do any tests?

A bit. I was looking for a better matrix than HVSBest, but having similar compressibility (although EQMV2 has lower compressibility it is being compensated by usage of RemoveGrain instead of Undot). Your matrix was highly recommended to me and current result that AutoGK produces has less block artifacts in high motion scenes that previous (ppl can test that - I'll be interested to hear opinions about quality 1.60 vs 1.70).

[Sharktooth]

Interesting.
I will test it ASAP.

[bourtzovlakas]

[20/10/2004 2:36:13 oLd] AutoGK 1.70b
[20/10/2004 2:36:13 oLd] OS: WinXP (5.1.2600).2
[20/10/2004 2:36:13 oLd] Job started.
[20/10/2004 2:36:13 oLd] Input dir: D:\**********\VIDEO_TS
[20/10/2004 2:36:13 oLd] Output file: D:\**********\gits.avi
[20/10/2004 2:36:13 oLd] Audio: Japanese
[20/10/2004 2:36:13 oLd] Subtitles: none
[20/10/2004 2:36:13 oLd] Codec: XviD
[20/10/2004 2:36:13 oLd] Target size: 700Mb
[20/10/2004 2:36:13 oLd] Started encoding.
[20/10/2004 2:36:13 oLd] Demuxing and indexing.
[20/10/2004 2:38:22 oLd] Processing file: D:\*********\VIDEO_TS\VTS_01_1.VOB
[20/10/2004 2:38:22 oLd] Processing file: D:\*********\VIDEO_TS\VTS_01_2.VOB
[20/10/2004 2:38:22 oLd] Processing file: D:\*********\VIDEO_TS\VTS_01_3.VOB
[20/10/2004 2:38:22 oLd] Processing file: D:\*********\VIDEO_TS\VTS_01_4.VOB
[20/10/2004 2:38:22 oLd] Source aspect ratio: 16:9
[20/10/2004 2:38:22 oLd] Source resolution: 720x480
[20/10/2004 2:38:22 oLd] Found NTSC source.
[20/10/2004 2:38:22 oLd] Source seems to be pure FILM.
[20/10/2004 2:38:22 oLd] Found 118894 frames
[20/10/2004 2:38:22 oLd] Encoding audio.
[20/10/2004 2:49:42 oLd] Audio size: 77,229,344 bytes (73.65 Mb)
[20/10/2004 2:49:42 oLd] Overhead: 4,755,712 bytes (4.54 Mb)
[20/10/2004 2:49:42 oLd] Video size: 652,018,144 bytes (621.81 Mb)
[20/10/2004 2:49:42 oLd] Running compressibility test.
[20/10/2004 2:49:42 oLd] Writing the following script to D:\***********\agk_tmp\gits_comptest.avs
===========================================================
LoadPlugin("D:\PROGRA~1\AutoGK\DGMPGDec\DGDecode.dll")
LoadPlugin("D:\PROGRA~1\AutoGK\filters\autocrop.dll")
LoadPlugin("D:\PROGRA~1\AutoGK\filters\RemoveGrain.dll")

movie = mpeg2source("D:\**********\agk_tmp\gits.d2v")
cropclip = autocrop(movie,mode=0,wmultof=4,hmultof=4,samples=10,aspect=0,threshold=34,samplestartframe=0)
fixed_aspect = 1.18518518518519
c_width = width(cropclip)
c_height = round(height(cropclip) / fixed_aspect)
input_par = float(c_width)/float(c_height)
input_par = input_par > 1.4 ? input_par : (4.0/3.0)
out_width = 640
out_height = round(float(out_width) / input_par)
hmod = out_height - (floor(out_height / 16 ) * 16)
out_height = (hmod > 4) ? (out_height + (16 - hmod)) : (out_height - hmod)
new_aspect = (float(out_width) / float(out_height)) / fixed_aspect
autocrop(movie,mode=0,wmultof=4,hmultof=4,samples=10,aspect=new_aspect,threshold=34,samplestartframe=0)
Lanczos4Resize(out_width,out_height)
RemoveGrain(mode=2)
SelectRangeEvery(300,15)
===========================================================
[20/10/2004 2:49:52 oLd] Duration was: 10 seconds
[20/10/2004 2:49:52 oLd] Speed was: 590,81 fps.
*************************************
EXCEPTION: Cannot open file "D:\********\agk_tmp\frames.log". The system cannot find the file specified
*************************************
[20/10/2004 2:49:52 oLd] Job finished.



What am i doing wrong????

[Thomas Davie]

Originally posted by len0x

Personally I think there are better methods of giving attention to the fact that you can donate (like startup splash windows etc), but I'm just too nice to put that all that stuff in there. And the only reason I don't like "donate for first access" options is that its too difficult to develop something that you release very often and it becomes a mess until you have your own private forum with lots of ppl inside it to make quick bugfixes.



I've already found Auto GK useful enough that I would purchase it. I also am lazy enough that I won't bother with a donation (I know, a fault on my part). I'm also unable to use Pay Pal. Do you have a physical address where I could send a money order?

thanks

Tom

[len0x]

Originally posted by bourtzovlakas
What am i doing wrong????

If you open *_comptest.avs in vdubmod you will probably see the exact error.

[Corence]

I'm getting the same error as bourtzovlakas. The VirtualDub Error is:

Avisynth open failure:
LoadPlugin: unable to load "C:\PROGRA~1\AutoGK\filters\RemoveGrain.dll"
(F:\Burn\Comedy\agk_temp\Eddie Izzard - Dress to Kill_comptest.avs, line 3)

I hope this helps. I've never had this error using previous versions.

[len0x]

Originally posted by Corence
LoadPlugin: unable to load "C:\PROGRA~1\AutoGK\filters\RemoveGrain.dll"


is the dll file there?

P.S. Is your CPU SSE-compatible (coz thats what is required for RemoveGrain...)

[Corence]

The dll file is there. I'm using an AMD Athlon XP 1700+ CPU.

[len0x]

Originally posted by Sharktooth
Interesting.
I will test it ASAP.

btw, what is the difference between V2 and V3HR?