Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion. Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules. |
24th November 2009, 23:41 | #601 | Link | |
Aging Video Hobbyist
Join Date: Dec 2004
Location: Off the Map
Posts: 2,461
|
Quote:
I dunno how Ripbot processes .avs files except that something called avs2avi.exe appears to be employed cuz I saw it running in the background using Task Manager. This is just an FYI and no problem for me at present, as I can simply use instead DGSource w/CUVIDServer and get what I want out of RipBot. Then I run x264 using DGMultiSource as usual with my own custom cmd files. |
|
29th November 2009, 16:14 | #602 | Link |
Guest
Join Date: Jan 2002
Posts: 21,901
|
Version 2.0.0 beta 4
* Full pulldown handling (Honor Pulldown, Force Film, Ignore Pulldown, frame repeats) implemented for all video types: AVC, MPEG2, and VC1.
* Prevent crash due to stream corruption erroneously causing the decoder to try to create a frame size bigger than 1920x1088. * Fixed a bug where if a stream has a coded size of 1088 and you respond to the crop warning with "yes", cropping is not properly set. That happened only if the display rectangle was also 1088. * Fixed bug that added an extra extension to the project file name if one was already present. * A more informative error message is printed if an MPEG1 program stream is loaded. * Removed a check that caused some valid audio streams to be rejected. http://neuron2.net/dgdecnv/dgdecnv.html Note that my web pages for AVC, MPG, and VC1 have now been consolidated into one page. |
30th November 2009, 03:08 | #604 | Link |
Registered User
Join Date: Apr 2006
Posts: 225
|
Neuron2
I have a question about how your indexing tools identify frames in transport streams. I am working on a tool to correct audio sync caused by errors during recording. It currently uses a .dgm index file to determine when frame or field repeats are in the stream. I came across a recording where my program lost synchronization with the index file. After some investigation, I discovered that this recording has packets marked with the Transport Error Indicator Flag, 1197 of them to be exact. Of these, 22 have the Payload Unit Start flag, and another 5 have a valid PES start code. When I adjusted my program to ignore packets with the Transport Error Indicator flag set, it still picks up one more frame than your indexer does. I know this because at the end of the scan through the file, I can see that my program is expecting a field repeat one frame early. So my question is, what does your indexer look for to decide that a new frame is starting? The clip I'm working on is MPEG2, and I am using the NVBeta10 MPEG2 indexer. Thanks! Matt |
30th November 2009, 05:18 | #605 | Link |
Guest
Join Date: Jan 2002
Posts: 21,901
|
In some previous versions I used to reject packets with transport_error_indicator set. The latest version ignores it.
For frame detection: Wait for a picture coding extension and then: Code:
if (picture_structure == 3 || (top_field_first && picture_structure == 1) || (!top_field_first && picture_structure == 2)) { // New frame. } Last edited by Guest; 30th November 2009 at 05:24. |
30th November 2009, 06:46 | #606 | Link | |||
Registered User
Join Date: Apr 2006
Posts: 225
|
Quote:
Quote:
Quote:
One more thing, does the non-NV indexer detect frames the same way? I don't use it much anymore, but I would like my program to be able to use those indexes too someday. |
|||
30th November 2009, 14:02 | #607 | Link | |
Guest
Join Date: Jan 2002
Posts: 21,901
|
Yes, of course. It's part of the MPEG2 video syntax. You can't reliably find frames just using the transport layer and PES syntax.
Quote:
If you need any help finding the cause of discrepancies between our methods, let me know. I do think you'll need to parse the video syntax, though. It's not so hard. |
|
30th November 2009, 14:45 | #608 | Link |
Registered User
Join Date: Oct 2009
Posts: 212
|
Hey neuron2 thanks for the tool it looks very good. I'm going to test it, is there any visual quality boost compared to Dgindex? I was thinking of using it with DVD sources if this is supported? I need to update my Geforce 9500Gt drivers, the latest are 195.62 version, is this ok? Thanks
Last edited by SilaSurfer; 30th November 2009 at 14:47. |
30th November 2009, 15:12 | #609 | Link | ||
Guest
Join Date: Jan 2002
Posts: 21,901
|
No. The benefit over DGIndex is a) offloading some CPU load to the GPU for transcoding, b) resizing/deinterlacing performed on the GPU.
Quote:
Quote:
|
||
30th November 2009, 18:27 | #612 | Link |
Guest
Join Date: Jan 2002
Posts: 21,901
|
The capabilities of the tools are documented in their manuals. If you don't see a feature described then it is not supported. As mentioned several times, I can support only the features available in the NVCUVID API, and currently it does not support the things you mentioned.
|
30th November 2009, 21:35 | #614 | Link |
Deblender
Join Date: Nov 2007
Posts: 18
|
There seems to be an issue with this version when decoding interlaced VC-1 content.
First I thought I had [] when it crashed for all 5 blurays I have here, but in fact all 5 are interlaced VC-1 (don't ask me where I get all that [] from^^) So I had a friend test some files for me: Non-interlaced VC-1 was fine with him, interlaced crashed right after starting the indexing process. (Same as for me) I had him test this file: http://www.mediafire.com/?zjm4jyywmdl which happens to be the same you wrote the stream corruption fix for, so I suppose the bug was introduced after this fix Anything else I can provide to help you spot and annihilate the bug?
__________________
This post has been created out of recycled electrons only Last edited by Guest; 30th November 2009 at 21:37. Reason: 4 |
30th November 2009, 21:49 | #617 | Link |
Deblender
Join Date: Nov 2007
Posts: 18
|
Oops, sorry for the vocabulary
Yes, the sample is the same because I had it handy. It happens for ALL 5 interlaced VC-1 blurays I own though. In beta 3 I can index that sample, but it crashes while encoding. In beta 4 it crashes right after I start indexing and does so for all interlaced VC-1 as it seems. By "don't ask me where I got all those from" I was just saying that I don't know why out of 5 blurays I have 5 interlaced VC-1 as they are not *that* common after all and have always given me trouble
__________________
This post has been created out of recycled electrons only |
30th November 2009, 21:59 | #619 | Link |
Deblender
Join Date: Nov 2007
Posts: 18
|
Ah, missed that one.
It's from a King of Queens Bluray that was only released in Germany so far if I'm not mistaken. When I noticed the issue I even got another copy from a rental store and tried to encode the episode from that disc with the same result. So I believe the issue should not be a corruption on my side.
__________________
This post has been created out of recycled electrons only |
Thread Tools | Search this Thread |
Display Modes | |
|
|