DGMPGDec 1.4.9 Final

[ACrowley]

ok, heres a sample from Kingdom Of Heaven BlueRay with
DTS HD Audio

http://www.megaupload.com/?d=ICEJJ421

Perhaps the Problemis in the Stream and its not DGindex

[Guest]

Perhaps the problem is in the Stream and its not DGindex Then before I go wasting time on a possibly bad stream, tell me how you know that m2ts file contains DTS audio and what player can play the audio from that file. Or what demuxer can extract good DTS audio from it.

[ACrowley]

Then before I go wasting time on a possibly bad stream, tell me how you know that m2ts file contains DTS audio and what player can play the audio from that file. Or what demuxer can extract good DTS audio from it.

because its the stream untouched from Blueray and containes DTS HD 1536kbps

[Guest]

Since you won't answer my questions, I will ignore you.

[guada2]

http://www.megaupload.com/?d=ICEJJ421

Your stream have some errors.
However, i don't found DTS HD file (strange!!!!).

tell me how you know that m2ts file contains DTS audio and what player can play the audio from that file. Or what demuxer can extract good DTS audio from it.

* Perhaps PS3.
*it was interesting to know this demuxer or the ........

[MichalHabart]

Since you won't answer my questions, I will ignore you.

That m2ts file contains DTS-HD because drmpeg utility xport.exe is able to demux DTS-HD and mpv track from it.
I am not able to do the same with dgindex, it can not see any audio stream inside m2ts.

At the moment, there is no other utility/program that can handle DTS-HD streams in m2ts.
Xport can be found here http://www.w6rz.net/xport.zip

[guada2]

Are you sure that xport.exe can import DTS-HD?

Drmpeg could you confirm that, please?

[Guest]

That m2ts file contains DTS-HD because drmpeg utility xport.exe is able to demux DTS-HD and mpv track from it. I tried that before but it didn't seem to work. Please tell me the exact command line you gave and then what you play the result with to see that it is OK. Thank you.

This is what I got:

-------
E:\Don\Programming\C++\Xport>xport -h sampleDTS.m2ts 1 1 1
xport Transport Stream Demuxer 0.94
program = 1, video channel = 1, audio channel = 1

packets for pid 6929 <0x1b11> = 1, first = 1, last = 1
coded frames = 0, video fields = 0

E:\Don\Programming\C++\Xport>
-------

The resulting video and audio files had length 0.

[MichalHabart]

I tried that before but it didn't seem to work. Please tell me the exact command line you gave and then what you play the result with to see that it is OK. Thank you.

This is what I got:

-------
E:\Don\Programming\C++\Xport>xport -h sampleDTS.m2ts 1 1 1
xport Transport Stream Demuxer 0.94
program = 1, video channel = 1, audio channel = 1

packets for pid 6929 <0x1b11> = 1, first = 1, last = 1
coded frames = 0, video fields = 0

E:\Don\Programming\C++\Xport>
-------

The resulting video and audio files had length 0.

Seems to me that sample provided by ACrowley was broken. I tried it but it has no output. This sample with command "xport -h 1 1 1" demuxes video file and DTS-HD track. Is it enough or should i provide bigger file?
http://rapidshare.com/files/22606202/sample2.m2ts.html

[Guest]

Thanks, Michal, that's just fine.

Here is beta 17, which can now get the DTS out of that stream. You wouldn't happen to know what the other three mystery PIDs are, would you (1200, 1201, 1202)?

Also, DGIndex was built with Visual Studio 2005 Express, rather than my old standby VC++ 6.0. So let me know if you see anything funny with it. That "dinosaur" crack really got to me. :)

http://neuron2.net/dgmpgdec/dgmpgdec149b17.zip

[yonta]

You wouldn't happen to know what the other three mystery PIDs are, would you (1200, 1201, 1202)?


subtitles, most probably.

[guada2]

Super neuron :)

[MichalHabart]

Thanks, Michal, that's just fine.

Here is beta 17, which can now get the DTS out of that stream. You wouldn't happen to know what the other three mystery PIDs are, would you (1200, 1201, 1202)?

Also, DGIndex was built with Visual Studio 2005 Express, rather than my old standby VC++ 6.0. So let me know if you see anything funny with it. That "dinosaur" crack really got to me. :)

http://neuron2.net/dgmpgdec/dgmpgdec149b17.zip

This is technical specification of that BR disc:

Video Resolution/Codec
1080p/MPEG-2

Aspect Ratio(s)
2.35:1

Audio Formats
English DTS HD Lossless Master Audio 5.1 Surround

Subtitles/Captions
English SDH
English Subtitles
French Subtitles
Spanish Subtitles

Supplements
Theatrical Trailer

So i assume, those PIDs are subtitles. But at the moment, there is nothing that can extract them out :(

Btw, how about adding also support for LPCM in m2ts files? If you want, i can give you sample of such file.

Edit: I just tried newest dgindex but it still can not see DTS track inside m2ts even if it is file i sent you sample from ;(

[SeeMoreDigital]

Edit: I just tried newest dgindex but it still can not see DTS track inside m2ts even if it is file i sent you sample from ;(Are you referring to your "sample 2" or guada2's sample.... If so, DGIndex/DGMPGDec 1.4.9 Beta 17 de-muxed the DTS audio stream of both samples just fine for me :)

[MichalHabart]

Are you referring to your "sample 2" or guada2's sample.... If so, DGIndex/DGMPGDec 1.4.9 Beta 17 de-muxed the DTS audio stream of both samples just fine for me :)

Cool, i did not change anything but now it works :)
Thanks neuron!!
But this will only work when video is MPEG-2. I have also m2ts with AVC (H264) inside. Is there any possibility to demux it also with gdindex? i get only error message that video sequencer not found.

[Guest]

Btw, how about adding also support for LPCM in m2ts files? If you want, i can give you sample of such file. Yes, I can do that. Please provide a test file.

Is there any possibility to demux it also with DGIndex? No. You'll have to wait for me to complete DGAVCIndex.

[MichalHabart]

Yes, I can do that. Please provide a test file.

No. You'll have to wait for me to complete DGAVCIndex.

At the moment, i don't have such m2ts but i will provide you one as soon as possible.
And for GDAVCIndex, is there any progress?

Edit: I found one m2ts with LPCM inside. Here is sample: http://rapidshare.com/files/22713733/sampleLPCM.m2ts.html

[Guest]

And for DGAVCIndex, is there any progress? I don't understand your question. It is in development, do you need more details about it?

[MichalHabart]

I don't understand your question. It is in development, do you need more details about it?

I found thread for it so now i know that beta3 is the last one. And i also know the progress at the moment.

[Terranigma]

thanks for the update neuron2. :)

[MichalHabart]

neuron, is there any reason why d2v created by beta17 can not be load into gordian knot? structure changed?

[kbello]

For me works fine, may be you need replace .dll

[Boulder]

Don,

just a quick one to let you know that your website shows beta 14 as the latest one instead of b17.

[Atak_Snajpera]

Neuron2: I will be honest with you . Audio Delay detection is not working at all! Two examples: Casino Royale - Dgindex shows -192ms instead of -272ms. KungFu Hustle -32ms instead of -144ms. Somehow DVD2AVI and VirtualDubMod are able to detect correct value but not Dgindex. As far as I know Dgindex is based on DVD2AVI so I assume You must have broken something in the engine.
You did great job but unfortunately you forgot to fix the most annoying bug.:(

[Guest]

It's been discussed to death. DGIndex follows the MPEG2 specification to the letter. If some other app to which you are comparing it doesn't, then that is not my problem and, "to be honest", given your attitude, I'm not very interested in helping you to understand why things are the way they are.

[Zep]

would it be worth while to some how double check by comparing time stamps or some other method?

[Guest]

would it be worth while to some how double check by comparing time stamps or some other method? I don't know what you mean. Do you mean compare between DGIndex and DVD2AVI? I already know why their calculations differ.

And Atak hasn't bothered to give us a VOB that I can use to look at his issue.

[Atak_Snajpera]

I've uploaded 2 samples. Please take a look

Casino Royale
http://www.mediafire.com/?2hwcynfzwhw

Kungfu-Hustle
http://www.mediafire.com/?0z2i21zkzmz

[MichalHabart]

I've uploaded 2 samples. Please take a look

Casino Royale
http://www.mediafire.com/?2hwcynfzwhw

Kungfu-Hustle
http://www.mediafire.com/?0z2i21zkzmz


Casino Royale has H264 as video track, could this be that problem?

[Guest]

Neither VOB is long enough to contain content that has a clear correlation between a video event and an audio event, so I cannot objectively determine the "correct delay. All I can say is that I am following the MPEG2 spec. If you don't like that, then use something else (or just adjust by 80ms, see below).

Here is the analysis: The first PTS refers to the first access unit in coded order, but because it is IBBP in coded order, the first access unit in coded order (the I frame) is the third frame displayed:

coded: IBB...
display: BBI...

Thus the time of the first display frame must be adjusted by two frame times, which is 80ms in PAL. So:

272 - 80 = 192ms

If you have a different reading of the MPEG2 spec, please enlighten us, with section references. :)

Ask mpucoder about it if you don't believe me.

[Zep]

I don't know what you mean. Do you mean compare between DGIndex and DVD2AVI? I already know why their calculations differ.

And Atak hasn't bothered to give us a VOB that I can use to look at his issue.

Yes i realize you follow the spec so I'm merely wondering if there was something more you could do the make sure the spec check was indeed correct when a source stream is NOT following specs then maybe you could use an alternate method to get offset.

Interestingly enough projectX gets the sync fine on the samples and DGindex does not. projectX must be doing more than just a strict spec check. I think it drops the first GOP then shifts audio alignment to that complete GOP which you then end up with offset = 0 on the demux.

My ultimate DGindex dream :) is someday DGindex could edit out more then just the front and back end then fix sync like projectX does then we get edit->sync fix->d2v and much faster cause it would be a streamlined work flow and it is not slow java! Now that would be so sweet! :D

thanks for all your hard work on DGindex, it really is great!


.

[Guest]

Interestingly enough projectX gets the sync fine on the samples and DGindex does not. projectX must be doing more than just a strict spec check. I think it drops the first GOP then shifts audio alignment to that complete GOP which you then end up with offset = 0 on the demux. What samples are you referring to? Let's be specific and take one clear sample that has an audio/video event that will allow us to determine the real sync. I'm willing to follow up on a specific case but I'm not going to respond to unsupported claims like you just made.

[Guest]

Let's be specific and take one clear sample that has an audio/video event that will allow us to determine the real sync. I'm willing to follow up on a specific case... All I'm hearing is crickets.

[Zep]

What samples are you referring to? Let's be specific and take one clear sample that has an audio/video event that will allow us to determine the real sync. I'm willing to follow up on a specific case but I'm not going to respond to unsupported claims like you just made.

Unsupported? I sent you many samples **DIRECTLY TO YOUR FTP SERVER** in the past that showed this very same problem which I talked about on more than one occasion. You even told me in a PM to STOP sending you so many samples as it was your work server or something and had limited space. So i stopped.


What did you learn from all those samples that had the exact same problem we are talking about now???


Anyway, i will start using DGindex again and get you even MORE samples. In the mean time i suggest you simply raw some HDTV streams on your home PC and see for yourself.

Thanks

[Zep]

All I'm hearing is crickets.

You actually replied twice to me before I had the free time to get back in here even once. I hope you enjoy the sound of those crickets. :D

[Guest]

Please indulge me by providing a stream again that shows the problem and has a clear audio/video event that will allow us to determine the real sync. As I said, I am now willing to follow up on it until we are all satisfied with the result.

[Guest]

Please indulge me by providing a stream again that shows the problem and has a clear audio/video event that will allow us to determine the real sync. As I said, I am now willing to follow up on it until we are all satisfied with the result. Hearing crickets again...

I suspect the problem may be the need for a clear AV event, such as a door slamming. Without such an event, there is no way to objectively determine the correct delay. Then we're left with "Well, application XYZ says the delay is ABC milliseconds, and DGIndex says it is TUV milliseconds, so DGIndex must be broken."

I repeat again, I'm willing to take a clear case through to completion, if anyone can provide one.

[plokm]

I found what looks like a bug in the templates. If you have two __vid__ placeholders on the same line, only the first one is processed. I didn't check to see if __aud__ is doing the same thing.

[Guest]

That's right. Why do you have more than one per line? Just curious.

[plokm]

TFM(d2v="__vid__", ovr="__vid__ override.txt", output="__vid__ metrics.txt")

I would prefer to do that when ripping multiple episodes in a series, just so the related files don't all have the same name.

[Guest]

Fair enough. Thank you for bringing it to my attention. Next beta.

[plokm]

Thanks!

[3ngel]

@neuron2
I dont' know much about avisynth env, but can you tell me if MPEG2Source() works syncronously with avisynth?
In other words, MPEG2Source() render a frame each time avs ask for it?
In this case you can implement a customizable read-ahead (asynch multithreaded) decoding in order to fully use all the cpus power?
In other words while avs is processing frame, MPEG2Source() at the same time read the (customizable) ahead frames?

[Guest]

@neuron2
I dont' know much about avisynth env, but can you tell me if MPEG2Source() works syncronously with avisynth?
In other words, MPEG2Source() render a frame each time avs ask for it? Yes, MPEG2Source() decodes one frame when Avisynth calls the GetFrame() method.

In this case you can implement a customizable read-ahead (asynch multithreaded) decoding in order to fully use all the cpus power?
In other words while avs is processing frame, MPEG2Source() at the same time read the (customizable) ahead frames? I don't have a lot of time for such work right now, as I am busy with other exciting things.

[3ngel]

Ok :)
Because dual (or more) processors, are becaming more and more common i thought that was a nice idea to gain a little (speed improvements i read are always requested) ;).

[ACrowley]

1.4.9 Beta 17 works fine with DTS( HD) demuxing from m2ts Streams from BluRay. THX

LPCM Audio is not supported at the moment right ?
1.4.9 Beta 17 cant detect LPCM in all my m2ts

[Guest]

LPCM Audio is not supported at the moment right ? 1.4.9 Beta 17 cant detect LPCM in all my m2ts If you can provide a sample stream, I will try to support it. Do it fast, because I am preparing to release 1.4.9.

[uncertainty]

If you can provide a sample stream, I will try to support it. Do it fast, because I am preparing to release 1.4.9.


I just used your latest beta for a Blu-Ray M2TS file (Stealth) and found it selected the French AC3 track and not English (it used PID 1102 instead of the PID 1100 which was english AC3). Is there anyway to have it allow the user to select which track to use from within the MEGUI frontend?

thanks for a great product btw....once I can get it to where the AC3 track is selectable (within megui) for M2TS it will remove 1 step I have to do now atm, ie demux the mpeg2 stream and AC3 out.

[Guest]

You'll have to ask the megui guys about that one. My selection mechanism is to set the right PID.

[uncertainty]

You'll have to ask the megui guys about that one. My selection mechanism is to set the right PID.


Thanks , for now I'm just using your program seperately to create the d2v then I start up megui.....this way I can then choose the correct PID...