DGMPGDec Development List

[gonwk]

Hi folks,:)

I am not sure if I am posting this in the right place ... but I searched and searched this forum and had no luck.

Q: Which version of the DGMPGDec 1.4.9 Final is correct? I downloaded one version from Free-Codecs site and one from Here and the only difference is that Doom9 has a newer date and different file size for "DGDecode.dll" & "DGVfapi.vfp" ... so I am curious which one is latest version?:confused:

Thanks,
G!:thanks:

[Guest]

Some people had a problem with the VC 2005 builds so I made later VC6 builds. They are functionally equivalent so you can use either one.

[gonwk]

Hi Neuron2,:)

THANKS for your answer ... and furthermore THANKS for this GREAT program ... really useful.

BTW, I am using the version that has "Latest" dates.

G!:thanks:

[kypec]

Hi neuron2:
I just want to point out that there is missing<a name="Top"></a> section definition at the top of the DGIndexManual.html file which is packed in 1.4.9 ZIP archive hence Table of Contents anchor is not working :(

[Guest]

I updated the file in the distribution. Thank you.

[dstarfire]

I really love DGMPGDec's template.avs feature.

It'd be even better, if it included an automatic script builder, similar to the one in megui, but oriented more towards creating general purpose templates (this would be most useful for long series, but also would speed up script creation for similar releases).

[Guest]

Thank you for your suggestion, but you'll need to give me a much better specification of desired functionality if you hope for that to become reality. I don't know anything about Megui script creator and I have no intention of ever using it.

[dstarfire]

Well, the most useful features are that it detects whether a given video is interlaced, telecined, or otherwise mangled by a formal mangling protocol, and generates appropriate de-mangling code. Also includes simple gui for resizing (choose target dimensions, and preferred filter), as well options to apply colour correction, and deblocking (simple yes or no for both).

Mostly, it's just the auto de-interlace/ivtc and resizing that make it worthwhile. And since, dgindex seems to already generate most of this info on it's own, I'd think it'd be fairly simple to add-on (especially since the source for a working example is freely available). Heck, you could almost just borrow that whole bit of code (giving approriate credit, of course).

[josey_wells]

I have found that DGMPGDec reports a file as interlaced when it may not be. I believe it is reporting the state of the last few frames found to be whether the file is progressive or interlaced. Until the count is added as a feature, I have written a program to parse the d2v file.

Attached is a simple program which parses the d2v file created by DGMPGDec and reports the number of I-Frames which are progressive and interlaced.

Usage is ParseD2v D2VFilename.d2v

[Guest]

I have found that DGMPGDec reports a file as interlaced when it may not be. DGIndex does not report a file as interlaced or progressive. This is a fantasy that exists only in your head because you refuse to read the documentation.

[josey_wells]

At the request of some users I have made some slight modifications to ParseD2V. It will now return the number of I-Frames, B-Frames, and P-Frames in the d2v file. It will also list the GOP segments that are interlaced and progressive according to the I-Frame flags.

Usage is still ParseD2v D2VFilename.d2v

Have fun!

[josey_wells]

Found a mistake so uploading correction

[Fallen]

Any chance of setting Location parameters via cli

I'm pulling about 15sec from up to a hundred 1GB .vobs >>d2v>>avisynth (creating motion menus)
Creating full d2v's for each .vob is time prohibitive (40 Sec +)(but quite quick (<1 sec when I use the Location sets)

I'm currently altering the ini (Adding Location-a bit clunky)>>processing the clip>>then reprocessing the .ini location entry back to the original (so subsequent calls to DGIndex aren't limited
It works -- but it seems more elegant via cli

[Fallen]

Sorry my last post was all wet. I can't do it via the ini (mixing up output and input)
But still trimming via cli would be quite useful (and would save a lot of time if you only need a partial clip)
from studying your source (I can't get it to compile yet--nasm paths:stupid:) it seems straightforward
CLI Ex -Trim[file,lba,file,lba] (same format as in your D2V output)

[Fallen]

Or Optionally loading a d2v file via cli (tried -- it kicks through to the media load file)
I noticed if I write a partial d2v file through the location ... entry
Then Start dgindex >>Load Project it all loads properly (with the partial file)
then Save Project to create the full file)
Advantage to this is that you can set any header value programmatically) clip,location etc

I'd make the change -- But can't get dgindex to compile with my old VC++ 6.0 Standard edition (Masm isn't supported)

Thanx for A great proggie

[carrot691]

Would it be possible to add the MPEG2 compressed frame size in Info()?

Thanks.

[Guest]

Would it be possible to add the MPEG2 compressed frame size in Info()? It's possible but you could tell me the purpose of it.

[Monamona]

The below linked MPEG-TS file contains a 2ch AAC audio stream whose beginning part is irregular.

http://rapidshare.com/files/279743773/SampleTS.zip.html

It sometimes happen because of timing of recording.
DGIndex incorrectly recognizes the audio stream as Mono MP1 and demuxes the whole stream as MP1.

Is it possible to modify DGIndex to discard the first sample of audio streams if it is irregular? (also recalculate audio delay to synchronize)
Or/And, could you add a command-line option like "-Dn: discard the first 'n' GOP" to start generating d2v and aac after 1st GOP?
Generally, for MPEG-TS, the first GOP tends to contain incorrect streams, so it is useful for encoding with batch files if DGIndex has a command-line option to discard the first several GOP.

BTW, I know existing "-RG" option has a similar function, but calculation of values is complicated.

[Guest]

Sorry, but I'm not interested in hacking up my code to allow for your bad edits. You'll have to step in by a GOP and start your project from there.

[Monamona]

Thank you for prompt reply.

Currenly, when 'offset2' of '-RG' option is set to 0, DGIndex does not process any frame.
How about modifying it to process to the end, since it is impossible to know the position of the last frame in advance.

[Guest]

That sounds reasonable. I'll do it but can't say when. Thanks for the suggestion.

[Monamona]

Thank you for your consideration!

[Selur]

from: http://forum.doom9.org/showthread.php?p=1427060#post1427060
Is it possible to use DGAVCDecNV directly as decoder (outputting: YUY2|IYUV|UYVY|YV12|NV12)?
(and pipe the output to ffmpeg/x264/ some other encoder X with YUY2|IYUV|UYVY|YV12|NV12 input support, without the need to add avisynth input support)
If not, that would be a nice feature to have.
It's something that could be done, but I don't see a usage case that would justify the effort. Anyway, that's off topic for this thread.
I would mainly use it to feed VC-1/H.264 to ivfenc, ffmpeg, mencoder (for webm creation), a badaboom like encoder (got no name atm.; cl h.264/vc-1 encoder using the nvidia libraries) and wanted to avoid using avisynth.
Just thought about using DGAVCDecNV as decoder and wanted to check if this is possible atm. so, if it needs much effort and nobody besides me wants/needs such a feature: don't implement it. :)

Cu Selur

[szabi]

When I save a project the bitrate informations are missing.

http://www.abload.de/img/vglap-2sdddo.png

Version 1.5.7.
Is it normal or it should be displayed after finish?
bye
szabi

[Guest]

It's normal. Those fields are populated only during play/preview. The idea is to be as fast as possible when saving a project, so minimal parsing is performed.

[szabi]

Ok, thanks for reply. :cool: