eac3to - audio conversion tool

[madshi]

This is report of xport
=================================
D:\[DVDISO] AIR BDBOX VOL.1\BDMV\STREAM>xport -h 00003.m2ts 1 3 4
xport Transport Stream Demuxer 0.98
program = 1, video channel = 3, audio channel = 4
Program Number = 0 (0x0000), Program Map PID = 31 (0x001f)
Program Number = 1 (0x0001), Program Map PID = 256 (0x0100)
program descriptor = 0x05, 0x04, 0x48, 0x44, 0x4d, 0x56
program descriptor = 0x88, 0x04, 0x0f, 0xff, 0xff, 0xfe
ES descriptor for stream type 0x02 = 0x05, 0x08, 0x48, 0x44, 0x4d, 0x56, 0xff, 0
x02, 0x44, 0x3f
ES descriptor for stream type 0x81 = 0x05, 0x04, 0x41, 0x43, 0x2d, 0x33
ES descriptor for stream type 0x81 = 0x81, 0x04, 0x08, 0x3c, 0x0e, 0x00
ES descriptor for stream type 0x80 = 0x05, 0x08, 0x48, 0x44, 0x4d, 0x56, 0xff, 0
x80, 0x31, 0x7f
ES descriptor for stream type 0x81 = 0x05, 0x04, 0x41, 0x43, 0x2d, 0x33
ES descriptor for stream type 0x81 = 0x81, 0x04, 0x08, 0x3e, 0x04, 0x00
Audio PID = 4355 <0x1103>, type = 0x80
ES descriptor for stream type 0x80 = 0x05, 0x08, 0x48, 0x44, 0x4d, 0x56, 0xff, 0
x80, 0x61, 0x7f
LPCM Audio Mode = 3/2+lfe
LPCM Audio Bits/sample = 16
LPCM Audio Sample Rate = 48000
ts rate = unspecified
============================
With the xport demuxed audio file everything seems to work for me. Just use "eac3to xportdemuxed.pcm xportdemuxed.flac". I think the audio track you demuxed with TsRemux was corrupt.

[madshi]

F:\>eac3to\eac3to.exe Innocence.LPCM16.8ch.pcm innocence.eac3to.ac3 -8 -16 -down6 -640
RAW, 7.1 channels, 0:01:00, 16 bits, 48khz
Converting the raw file to wav. Please wait...
Converting the wav file to ac3. Please wait...

Aften: A/52 audio encoder
Version 0.0.8
(c) 2006-2007 Justin Ruggles, Prakash Punnoor, et al.

input format: RAW Signed 16-bit little-endian 48000 Hz stereo
output format: 48000 Hz stereo (2/0)

SIMD usage: MMX SSE SSE2 SSE3
Threads: 2

progress: 100% | q: 306.6 | bw: 60.0 | bitrate: 640.0 kbps

Done.

innocence.eac3to.ac3 is just noise and 4 times bigger than the correct 5.1ch file.


tried converting the eac3to-generated wav file to ac3 using aften.exe.
F:\>eac3to\aften\aften.exe Innocence.LPCM16.8ch.wav innocence.aften.ac3 -b 640

Aften: A/52 audio encoder
Version 0.0.8
(c) 2006-2007 Justin Ruggles, Prakash Punnoor, et al.

input format: WAVE Signed 16-bit little-endian 48000 Hz 5.1-channel
output format: 48000 Hz 3/2 + LFE

SIMD usage: MMX SSE SSE2 SSE3
Threads: 2

progress: 100% | q: 255.5 | bw: 57.9 | bitrate: 640.0 kbps

innocence.aften.ac3 seems OK

converting to the other formats than ac3 works correctly.
That was a bug in eac3to. Thanks for reporting it. I'll upload a new build in a moment. Btw, "-down6" and "-640" are default when converting to ac3.

[madshi]

whats that:stupid:
You need to download GraphEdit. The file "dump.ax" is shipping with GraphEdit. Then run "regsvr32 dump.ax" in the command line to register the "dump" filter.

[madshi]

eac3to v1.15 released

http://madshi.net/eac3to.zip

This time there are only a few bugfixes.

[tebasuna51]

I think there are different kinds of corruption. The usual kind is probably that the length of the audio file is correct but that some of the data is corrupt. In that case "Silence" should maintain the sync. However, if the source container is corrupt (or if the demuxing routine has a bug) there could be additional invalid data in the source file. As a result the final repaired audio file could be too long. In that case "Skip" might maintain sync while "Silence" could lose sync.

Well, you are the author of last version I can't discuss with you :rolleyes: (I'm busy with ffmpeg to analyze your version)

But with previous DelayCut versions, see the post (http://forum.doom9.org/showthread.php?p=810264#post810264) all short frames are skipped and only full erroneus frames are replaced by silence then never the output can be long than input. But with 'Skip' option always the output is short than input because short and full frames are ignored.

[madshi]

Well, you are the author of last version I can't discuss with you :rolleyes:
Heh! No, actually I think you know more about delaycut than I do. I just copied every AC3 function and changed it to E-AC3 without really diving too deep into how it all really works. I definitely didn't change anything about how "skip", "silence" and "fix" work.

I'm busy with ffmpeg to analyze your version
That's good!

But with previous DelayCut versions, see the post (http://forum.doom9.org/showthread.php?p=810264#post810264) all short frames are skipped and only full erroneus frames are replaced by silence then never the output can be long than input.
The output can't be longer than the input. But it can be longer than the correct result would be (and that is what I meant). Let me give you an example:

Blu-Ray TrueHD tracks have AC3 frames + TrueHD blocks. delaycut in "Silence" mode will extract all correct frames just fine. But it might find the AC3 sync word in some TrueHD blocks. The CRC will not fit, but "silence" mode will then add silent frames instead. As a result you'll get an audio track which is too long. Agreed? So there are situations where "skip" is the better choice compared to "silence".

[tebasuna51]

Blu-Ray TrueHD tracks have AC3 frames + TrueHD blocks.

Then you are right one more time. :)

In standard ac3 streams only ac3 frames (goods or corrupted) can be present.

[Schotenhüter]

You need to download GraphEdit. The file "dump.ax" is shipping with GraphEdit. Then run "regsvr32 dump.ax" in the command line to register the "dump" filter.

i`am test it:thanks:

[madshi]

In standard ac3 streams only ac3 frames (goods or corrupted) can be present.
One thing I'm not sure about: Can a recorded TS stream be so corrupt that a video packet is mistakingly interpreted as an audio packet? If that's possible and if you then demux the AC3 stream, it might also contain too much data, I think.

[tebasuna51]

One thing I'm not sure about: Can a recorded TS stream be so corrupt that a video packet is mistakingly interpreted as an audio packet?

Well, is possible but improbable, not only the header must match but also the CRC frame.

[madshi]

Well, is possible but improbable, not only the header must match but also the CRC frame.
Ah, ok. Don't know much about TS yet. One of the things I still need to look into... :)

[Schotenhüter]

next prob :o

the audio decoder didn`t accept the input stream

:confused:

[madshi]

next prob :o
I don't have a telepathical connection to your computer. So if you don't give me any more details I can not help you. I need the full eac3to output and information about the source audio file you fed to eac3to.

[Icemaan]

Hallo

By me your great Tool works fine. But only with 1536kps Dts.When I want downsample to 768kps it not work.The Sound looks horrible with 1536 all ok
Any Idea because I am new may be I made somethinh not correct
Thanks for help

Icemaan

[VempX]

Sorry,I have to go to school Orz
I demux the audio track from m2ts by tsremux then I get a .pcm file.
I try to use eac3to to convert it to flac,but I get a strange .flac file which I gave you~
Does eac3to can't convert .wav file?
I demux the audio by gdsmux so I get a .wav file.But eac3to can't convert it =A=

[VempX]

Sorry,I have to go to school Orz
I demux the audio track from m2ts by tsremux then I get a .pcm file.
I try to use eac3to to convert it to flac,but I get a strange .flac file which I gave you~
Does eac3to can't convert .wav file?
I demux the audio by gdsmux so I get a .wav file.But eac3to can't convert it =A=

[VempX]

sorry
net hava a wrong
please delete this post

[markrb]

Strange problem and sorry if it's been discussed before. I did try and find it.

After converting an EVO to ac3 the file is small (165mb) and plays way too fast. I checked the wav file and it is 1.74gb, but it also plays way too fast.

There is no indication of anything wrong. The screen looks exactly like the successful screen shots I have seen here.
It even says 100% done at the end when aften is finished.

I did check the EVO with the single dd+ stream and that plays fine. I have tried it both command line and using the Gui with the same results.

2 Discs with the same issue. So I don't think it's the discs.

EDIT: I think my Nero was hosed. I reinstalled and I am testing it now.

Yep that was it. Un-install Nero. Use clean tool. Re-install. Seems to work fine.

Mark

[nautilus7]

You need to download GraphEdit. The file "dump.ax" is shipping with GraphEdit. Then run "regsvr32 dump.ax" in the command line to register the "dump" filter.

Where can i download graphedit?

[markrb]

Where can i download graphedit?

Try Google. Then look for the register.bat file and run that.

Mark

[Icemaan]

Hi all together

I have following Problem
When i merge the Video and Audio with MkvMerge i get this Message

audio/video go out of sync, but this stream is damaged.
Warning: dts_packetizer: skipping 3 bytes (no valid DTS header found). This might make audio/video go out of sync, but this stream is damaged

This only come bei Dts not by Ac3. Can anywhere help me.The Video File is 264 the Audio dts

Thanks a lot
Icemaan

[TheSof]

If I have just the TrueHD extracted via evodemux (.mpa), is there anyway I can get eac3to to accept it?

[honai]

@madshi

I've found a parsing problem with Pan's Labyrinth HD-DVD (EU release). For the English DTS-HD track EVOdemux reports:


DTS HD (DTS) audio stream 1 found!
First PTS = 0000D6D8
Substream id = 89
PCM core samples = 32
PCM sample blocks = 16
Primary frame byte size = 1023
Channel arrangement = LT + RT (left and right total)
Sampling frequency = 48 kHz
Transmission bitrate = 768 kbit/s
LFE channel = not present
Encoder software rev. = 7, Current
Source PCM resolution = 16 bits
PCM core samples = 32
PCM sample blocks = 16
Primary frame byte size = 1023
Channel arrangement = LT + RT (left and right total)
Sampling frequency = 48 kHz
Transmission bitrate = 768 kbit/s
LFE channel = not present
Encoder software rev. = 7, Current
Source PCM resolution = 16 bits


It's actually DTS stereo (2.0). eac3to complains:

The source file format is unknown.

[Roscoe62]

I've followed this thread as best as I can, but I still haven't figured out whether it will do this or not. I have a couple of questions and any help is appreciated!

1. Is it possible to demux a Dolby TruHD audio track from a BR disc? From what I've seen earlier in the thread it might be possible with xport...but where can I find xport?

2. Will EAC3to convert a Dolby TruHD audio track demuxed from a BR Disc into a multi-channel wav file? (or flac). If so, how?

3. If all the above IS possible how can I mux the newly created 5.1 wav file into a .ts container with the AVC video, while still staying in sync? If .ts isn't the way to go, could it be done with mkv? (I thought mkv had problems with AVC video)

I know there are a lot of questions here, but even getting answers for the first 1 or 2 would be great.

Spiderman 1 & 2 are being released very soon and they both have TruHD soundtrack which I'd like to backup without losing any quality. This is all pretty standard from HD-DVD, but I've not done anything like this with BR.

Again, any help is appreciated!

[madshi]

By me your great Tool works fine. But only with 1536kps Dts.When I want downsample to 768kps it not work.The Sound looks horrible with 1536 all ok
Any Idea because I am new may be I made somethinh not correct
Will need to check if I can reproduce the problem. Can you please try to use "eac3to source.whatever dest.wavs"? This will give you 6 wav files. Then please convert these 6 wavs to dts with Surcode manually. Is the resulting file ok?

[madshi]

Sorry,I have to go to school Orz
I demux the audio track from m2ts by tsremux then I get a .pcm file.
I try to use eac3to to convert it to flac,but I get a strange .flac file which I gave you~
Yeah, but you also gave me the PCM track demuxed by xport - and that works fine! So it seems that the PCM track taken from tsremux is corrupt. Use xport and it should work fine.

[madshi]

I have following Problem
When i merge the Video and Audio with MkvMerge i get this Message

audio/video go out of sync, but this stream is damaged.
Warning: dts_packetizer: skipping 3 bytes (no valid DTS header found). This might make audio/video go out of sync, but this stream is damaged

This only come bei Dts not by Ac3. Can anywhere help me.The Video File is 264 the Audio dts
Where is that dts file coming from? Was it created by eac3to?

[madshi]

If I have just the TrueHD extracted via evodemux (.mpa), is there anyway I can get eac3to to accept it?
Please read the first post of this thread. Short answer: No.

[madshi]

I've found a parsing problem with Pan's Labyrinth HD-DVD (EU release). For the English DTS-HD track EVOdemux reports:

It's actually DTS stereo (2.0). eac3to complains:
Can you please rebuild the DTS-HD track into a separate EVO file (with only that DTS-HD track in and nothing else) and then send me the first 5MB of that rebuilt EVO? Thanks!

[madshi]

1. Is it possible to demux a Dolby TruHD audio track from a BR disc? From what I've seen earlier in the thread it might be possible with xport...but where can I find xport?
Yes, it's possible with xport. You can download it here:

http://www.w6rz.net/

TsRemux should also be able to demux the TrueHD track.

2. Will EAC3to convert a Dolby TruHD audio track demuxed from a BR Disc into a multi-channel wav file? (or flac). If so, how?
Blu-Ray TrueHD tracks consist of an AC3 core with some additional TrueHD packets on top of that. eac3to can extract the AC3 core. Do "eac3to source.thd dest.ac3". This will give you the AC3 core of that TrueHD track. Obviously you'll only get 640kbps AC3 quality this way and not the full TrueHD quality. Currently to my best knowledge only PowerDVD can decode Blu-Ray TrueHD tracks and PowerDVD's Audio Decoder cannot be used outside of PowerDVD.

Spiderman 1 & 2 are being released very soon and they both have TruHD soundtrack which I'd like to backup without losing any quality. This is all pretty standard from HD-DVD, but I've not done anything like this with BR.
I'd suggest that you use TsRemux to rebuild the TrueHD tracks into a separate m2ts file. I think this way you should be able to back them up without wasting much space and with good hopes that we might find a way later to fully decode them. Of course you could also backup the whole movie m2ts. That would be the easiest way, but consume more space. Ideally, why backing them up right now? You should buy the Blu-Ray discs. So you have the data on your shelf, anyway... :)

[Roscoe62]

Ideally, why backing them up right now? You should buy the Blu-Ray discs. So you have the data on your shelf, anyway... :)

Thanks for the info Madshi. I will be placing my order for the discs tomorrow but I will not use PowerDVD for playback. I only playback via .ts or .mkv files so I can use my fav player - Theatertek. So I guess I will stick with the AC3 core for the time being, until someone smarter than me can figure out how to fully decode Blu-Ray TruHD tracks.

Here's hoping anyway! :)

[TheSof]

I only playback via .ts or .mkv files so I can use my fav player - Theatertek.

TT will happily play m2ts files.

[Icemaan]

Where is that dts file coming from? Was it created by eac3to?

The Dts File was created with eac3to. The File comes from the rebuild evo with only one audio stream inside

I have split the Evo with Evosplit may be this is the problem what you think

icemaan

[madshi]

The Dts File was created with eac3to. The File comes from the rebuild evo with only one audio stream inside

I have split the Evo with Evosplit may be this is the problem what you think

icemaan
I'm not sure. Does that warning come very often or just once? Is there a timecode for that warning? Could it be the timecode where you split the evo?

[Roscoe62]

TT will happily play m2ts files.

Perhaps,

but because of the TruHD soundtrack Theatertek won't play it unless I demux the BR M2TS, re-code the audio and then re-mux into a new container. And if I'm going to all that effort I'd rather re-mux to a container I'm familiar with.

[ACrowley]

Perhaps,

but because of the TruHD soundtrack Theatertek won't play it unless I demux the BR M2TS, re-code the audio and then re-mux into a new container. And if I'm going to all that effort I'd rather re-mux to a container I'm familiar with.


recode the Audio !? Just extract the DD core, no need to renencode it. Also maybe TT decodes only the DDcore from, Bluray TrueHD ?! i

[pimpMyHD]

I am interesting in using your software, but i have a question :

do i need nero hddvd/bluray plugin, if i already have sonic scenarist filters ?

I currently do not own nero 's plugin, but i own the sonic hd filters.

[TheSof]

I am interesting in using your software, but i have a question :

do i need nero hddvd/bluray plugin, if i already have sonic scenarist filters ?

I currently do not own nero 's plugin, but i own the sonic hd filters.

You need them if you are after the best sound quality. The sonic filters add nasty DRC.

[pimpMyHD]

You need them if you are after the best sound quality. The sonic filters add nasty DRC.
So, i should be able to use eac3to without the nero filters ?

Because it does not work on my system.

[TheSof]

So, i should be able to use eac3to without the nero filters ?

Because it does not work on my system.

If you just have the sonic decoders yes it should still work. What error message are you getting?

[Thunderbolt8]

I have a disc with 5.1 DD+ audio codec, what should I do now when I want to create a remux and this should include the best possible audio I can make out of it, which is currently possible to playback with a normal setup? I also guess that keeping the track as it is is also not a good solution (apart from the question if it was playable at all then?), because theres lot of additional junk in it like with dts-hs & dtscore ?
I think the solution would be something like converting to normal ac3 or DTS. but im not sure if dts conversion is possible then from a DD+ source, or if it does make sense at all?
another question I have is the removal of dialog normalisation, is such dialog normalisation included in every audio track? is the dialog normalisation somehow intended by the film makers, otherwise one wouldnt understand what people say in certain scenes? does it sound more natural & better without dialog normalisation?

[madshi]

I have a disc with 5.1 DD+ audio codec, what should I do now when I want to create a remux and this should include the best possible audio I can make out of it, which is currently possible to playback with a normal setup?
Define "normal setup".

If you have analog connection to your receiver or if you can transport multi channel PCM audio over HDMI to your receiver then the best way to handle DD+ audio is to keep it as it is. However, if you have SPDIF/coax connection to your receiver, you can only transport AC3 or DTS. So in that case you'd have to reencode the DD+ track to either AC3 or DTS.

I also guess that keeping the track as it is is also not a good solution (apart from the question if it was playable at all then?), because theres lot of additional junk in it like with dts-hs & dtscore ?
There's no junk in DD+ at all. And it's a good idea to keep it as it is. But see above.

another question I have is the removal of dialog normalisation, is such dialog normalisation included in every audio track? is the dialog normalisation somehow intended by the film makers, otherwise one wouldnt understand what people say in certain scenes? does it sound more natural & better without dialog normalisation?
The one and only purpose of dialog normalization is that it tries to save you from having to readjust volume during advertising breaks and when you switch from one movie to another. Basically dialog normalization turns the volume knob on your receiver so that the volume of speaking voices is the same in advertising and in every movie you're watching. Without dialog normalization advertising can sometimes be much louder than the movie audio track. Or it could happen that one movie is much louder mixed than another movie. So without dialog normalization you'll have to adjust the volume everytime there's an advertising break or everytime you switch to another channel/movie to keep volume at the same level.

Unfortunately, in real life most movies leave the "dialog normalization" information in their audio tracks at the default value. That means dialog normalization often simply doesn't work correctly. Furthermore many people argue that doing dialog normalization harms audio quality. Because practically it's not really the physical volume knob of your receiver that is changed by dialog normalization, but instead the digital audio track is processed to make audio volume lower. This effectively reduces bitdepth of the audio track (a very little bit).

Sony has decided to not use dialog normalization, anymore. This is based on feedback they received from home cinema consumers. But most other studios still leave dialog normalization activated (and most of the time on the default value).

[Thunderbolt8]

well, standart setup (should have said basic setup) means I just use my pc via ac3filter (or maybe ffdshow audio) to playback the movie/audio and currently just have 2 hi-fi boxes (non computer boxes), which are addressed via the rear green output. so basically nothing speacial, no external receiver or such :D
would the original, non converted track be able to played back via ac3filter? would I be able to listen to any improvements compared to a converted ac3/dts track? I guess the removal of the dialog normalization of the DD+ track then wouldnt change the fact that its still (un)playable with ac3filter?
in case I decide to convert to dts or ac3 and also would like to use dialog normalization, is there a fixed order I should apply to, e.g. first dialog normalization removal and then converting (or vice versa)?
when remuxing with mkvmerge, should the ending for the DD+ track kept .mpa or should/must it be changed into something else ?

[madshi]

well, standart setup (should have said basic setup) means I just use my pc via ac3filter (or maybe ffdshow audio) to playback the movie/audio and currently just have 2 hi-fi boxes (non computer boxes), which are addressed via the rear green output. so basically nothing speacial, no external receiver or such :D
Not even multi channel! :(

would the original, non converted track be able to played back via ac3filter?
Nope, at least not until ac3filter adds E-AC3 decoding support.

in case I decide to convert to dts or ac3 and also would like to use dialog normalization, is there a fixed order I should apply to, e.g. first dialog normalization removal and then converting (or vice versa)?
As long as you use eac3to for the conversion, you don't have to care.

when remuxing with mkvmerge, should the ending for the DD+ track kept .mpa or should/must it be changed into something else ?
Don't know how mkvmerge behaves. But I'd suggest renaming it to "ddp" or "eac3". Just try it out. If mkvmerge detects that it's E-AC3, you're done.

[Thunderbolt8]

well, since the track should be quite small compared to the video size of the remux I guess I'll do both, convert it, but also try to mux the original track into the .mkv to be on the save side for the future. when time comes, I only need to remux without the converted track then, instead of working with the HDDVD source again :S

edit: no changing of the ending of the demuxed audio at all needed for DD+, mkvmerge detects is as eac3 at once :)

[Thunderbolt8]

hm just read the first line of the OT again, wouldnt it make more sense for me just to convert the DD+ to flac? would it be completely lossless then, compared as when taking flac from a LPCM track? I guess that might be the best option then, because I can use it already now and im sure with better hardware it should still be possible to use it still in the future (wasnt there something about analog plug that allows flac use with external receivers?). or am I mistake here?

btw. what the difference between that flac121a, flac121b and the win.zip file from sourceforge, which one should I download?

[Taktaal]

Has anyone kept an old version of eac3to? It used to work very well but the versions in the last 2 months added only bloat and more dependancies. Unfortunately I overwrote my download from earlier this year...

[juzzy999]

I cant get eac3to encode a DTS-HD track.

I wont recognise the file.. How can i re-encode it to an AC3??????

[madshi]

hm just read the first line of the OT again, wouldnt it make more sense for me just to convert the DD+ to flac? would it be completely lossless then, compared as when taking flac from a LPCM track? I guess that might be the best option then, because I can use it already now and im sure with better hardware it should still be possible to use it still in the future (wasnt there something about analog plug that allows flac use with external receivers?). or am I mistake here?
Sure you can do that. But DD+ decodes to 24bit. And reencoding that to FLAC will consume a lot of space. Lossless compression works best for 16bit. The higher the bitdepth of the audio track is the lower is the compression efficiency of lossless compression.

btw. what the difference between that flac121a, flac121b and the win.zip file from sourceforge, which one should I download?
The FLAC which comes with eac3to is the only one which can handle big files without any problems.

[madshi]

Has anyone kept an old version of eac3to? It used to work very well but the versions in the last 2 months added only bloat and more dependancies. Unfortunately I overwrote my download from earlier this year...
Which dependancies do you mean? If you have a suggestion how to improve eac3to, just let me know.