Thanks a lot, Donald, for your work and testing. 1.I've noticed that on Test2 your plugin runs properly. But on Test3 always crashes. =============== Test2 =============== ------------------------------------ DGAVCDecodeDI+DiAVC+MT ------------------------------------ with SetMTMode(2,0) - 20-30 fps...

That's strange because you say it crashes at the beginning but the dump shows decoder_flush() in the call stack and that is not called until end of stream. I have no time to notice any video when it crashes. Can you test on a WinXP SP3 system? I can try under virtual machine only. But I don't think...

admin wrote: Can't you provide a text dump?

How it can be done? I'll do it, of course.

admin wrote: Also, what happens with your script if you just load it into VirtualDub and play it?

It crashes in very beginning. I've attached text file with crash info.

I've made this dump with Windows Task Manager. It can can be viewed with Windows debugger which comes with Windows SDK 7. As far as I know there is no Dr.Watson debugger in Windows 7. But maybe I am wrong and you can tell me how I can run it in Windows 7. Also I have diagnostic message from OS. I do...

There is a dump. http://www.filefactory.com/file/b4g7c95/n/x264.zip MD5: 5B24798EFD065F78523CA09F45308FD4 Size: 19'002'233 Avisynth script: SetMemoryMax(2000) SetMTmode(5,0) LoadPlugin("D:\arh\media\edit\video\_Filters\DGAVCDecDI\2025_20110106\DGAVCDecodeDI.dll") DGSource("g:\video1.dgi") return las...

I am not lecturing you. I'm just voicing my thoughts aloud. Also I clearly see the difference of CPU utilization between DGSource and DSS2 in my script. You can easily check that I am right. Just use Microsoft Process Explorer tool. It can show how many threads actually in use within one application...

When I run my script (DiAVCDecode), It crashes all time SetMTMode(5,4) DGSource() SetMTMode(2,4) MVTools processing part ... If I am running following script (DiAVCDecode), it works but with one thread only. It is because only 1 thread actually load video source as I wrote early. DGSource() SetMTMod...

Not quite. Firstly you need to install usual Avisynth and after that you should copy dll-s from SEt distributive over dll-s from official Avisynth.

I have usual Avisynth (last 32-bit version from SourceForge) and MT-libraries from SEt version 2009.07.12. I've got Avisynth MT here http://forum.doom9.org/showthread.php?p=1301410#post1301410 There are my Avisynth files http://www.filefactory.com/file/b4g40cf/n/my_Avisynth.zip MD5: 869C8001ACF98537...

I know, I know... But DiAVC decoder is better than it analogues. So I tried to use it. What makes you think that DSS2() cannot use the DiAVC decoder? At present I not quite understand how to make it work with DiAVC. One more time: DiAVC runs multithreaded distributed on all your CPUs. There is no n...

You need to ask DiAVC about it. I have confirmed that DGSource() itself is threadsafe. I can open a script several times in simultaneous VirtualDub instances and they play fine and independently. As far as I know possibility to run several applications simultaneously is not equal to multi-threading...

This is certainly not what I expected. DSS2() can work stable with multi-threading in all these scripts. But... That's something at least.
Thanks.

Thank you. I've downloaded last version. And what I have. This code works properly but only with 1 thread (please pay your attention on comments within script). DGSource("my_file.dgi") SetMTMode(2) # or SetMTMode(2,0) or SetMTMode(2,3) or SetMTMode(2,4) super = MSuper(pel=2, sharp=2) backward_vec2 =...

I don't understand how it could be reached within Avisynth (and with trial version of DiAVC)? Do you mean something like that? DGSource("my_file.dgi") super = MSuper(pel=2, sharp=2) backward_vec2 = MAnalyse(super, isb=true, delta=2, overlap=4) backward_vec1 = MAnalyse(super, isb=true, delta=1, overl...

Ok. I think I understand you.

Can you propose some test case regarding stable multi-threading with DGSource+DiAVC? I mean some avisynth script. Because you've told that in your computer everything works fine.

I know that DiAVC is multithreaded and I see it using multiple threads on my PC.

Well. May be you are right regarding decoder misconfiguration. The matter is that trial version has no configuration tool at all. :) Only file diavc.ax. I know this is not your area. It's just my think aloud. BTW. May be you will be interested. I've asked regarding my problem in DiAVC thread elsewhe...

Ok, thank you. May be you know some other way how to reach multi-threading processing in your application? I read a little about decoder you are using in your application. It is multi-threaded in its nature. But I see that only one core used when I am processing my video. It is not very effective I ...

Thank you very much, Donald. It works fine now. But... When I am using SetMTMode in my Avisynth script it crashes. There is no such problem when I am opening source file with DSS2() plugin instead of DGSource(). DSS2() comes with HAALI media splitter - avss.dll. Source file is big enough (about 18Gb...

It has the same problem with illegal frame_num value. What are you using to encode this? Is it an official version of x264.exe? If so, you need to ask the x264 guys why they are generating illegal files. Yes, it is official version. I took it from x264.nl. Thank you for your analysis. I doubt wheth...

There is an example.

http://www.filefactory.com/file/b4f9h39 ... seless.264

MD5: 178CB2406C737728C6AD2A2CA16DE9AD
Size: 28'586'623 b


By the way, I can't play this file with your soft. But the same file wrapped in mkv container by x264.exe encoder - I can.

Anyway, I demuxed the 264 out of the MKV and ran a parser on it. I get the dump below. You can see that the first IDR slice has a frame_num of 1 (must be 0 by spec). can it be error in coder (x.264.exe) or in muxer? This is ILLEGAL and so your stream is invalid. I am willing to try to support these...

So there is proof in attachment. Please take a look at it.

And there is MD5 hash for riget34_1_nv_progressive_loseless_mkvtools-001.mkv

8DFE95E6A9FA492A370022D667078B33

Its original size is 28'609'892 b. An its size on disk is 28'610'560 b (as Windows 7 shows).

I've downloaded your soft and decoder (trial version) yesterday (03 jan 2011). And I can play the file, at least one time. Within two hours I can prove it. And for any case I can provide you with MD5-hash for the file. It is possible that you did not downloaded it completely. Even if your soft can't...

DGAvcDecDI 2001 crashes when playing file specified in URL below. I am using trial version of last DiAVC decoder (1.2.2). http://rationalqm.us/board/viewtopic.php?f=8&t=69 I do following: 1. Load file to application 2. Play it (it plays correctly). You've lost me already. DGAVCDecDI cannot play the...

DGAvcDecDI 2001 crashes when playing file specified in URL below. I am using trial version of last DiAVC decoder (1.2.2). http://rationalqm.us/board/viewtopic.php?f=8&t=69 I do following: 1. Load file to application 2. Play it (it plays correctly). 3. Try to create project 4. Try to play file again ...