eac3to - audio conversion tool

[madshi]

I'm trying to get eac3to running, I'm running nero 7 and I keep getting the message
Disabling DRC for Nero (E-)AC3 decoding... then lines go across the screen and then it exits. It does not continue.
is there a specific version of nero 7 I need?

I tried a nero lite version and it said getting nero audio decoder 2 instance failed.
Can you please give me the full eac3to output with the Nero 7 (non lite) version? Did you properly buy the HD DVD plugin?

[madshi]

1) there are 3 types of channel order for 5.1 MLP files:

ID 12: Group 1 (Lf, Rf), Group 2 (C, LFE, Ls, Rs)
ID 17: Group 1 (Lf, Rf, C), Group 2 ( LFE, Ls, Rs)
ID 20: Group 1 (Lf, Rf, Ls, Rs), Group 2 (C, LFE)

and they are all valid channel order for 5.1 MLP, decoded to WAV, they all produce the same WAV file with the valid channel order for 5.1 WAV files: Lf, Rf, C, LFE, Ls, Rs.
I know that there are multiple channel ordering variations for MLP which are all "correct". I didn't know the details, though.

but eac3to 2.06 reports:

ID 20 order is wrong for WAV files, but not for MLP, so "5.1 wrong order channels" doesn't seem right to me.
:) I know. I just wanted to post something other than "5.1" to indicate that this is not the normal 5.1 channel order, but something different. I didn't expect that this other ordering would really be used in real life, so I just named it "wrong order" just to give it a name.

2) with Nero decoder all of the samples: mlp51_id12.mlp, mlp51_id17.mlp, mlp51_id20.mlp are decoded, but the output is different for mlp51_id20.mlp, probably it produces WAV with channel order Lf, Rf, Ls, Rs, C, LFE instead of the correct one Lf, Rf, C, LFE, Ls, Rs. it's important, because most of the DVD-Audio discs use ID20, ID12 and ID17 are very rare.

3) with -libav switch all of the above samples produce:

and no output at all, i don't know why because i have ffmpeg patched with one of the first MLP patches and it decodes all of the samples without complaining about the checksums, so maybe with the newest ffmpeg MLP patches, something is broken and previous versions are better to use at the moment, because there is no problem with the checksums according to Surecode MLP:

here are the samples: http://xkodi.svobodno.com/mlp51/
Thanks for the samples! Will check this out. May take a few days, though.

[madshi]

Another problem here.

I converted X-Men: The Last Stand DTS-HD Master Audio track to FLAC, and the playing didn't go well
The reason for that is that you most probably don't have a "FLAC source" filter installed which could handle standalone FLAC files. ffdshow can only handle FLAC files which are part of a container. ffdshow cannot handle standalone FLAC files.

so I put the FLAC file in mka container and now it is playable (If I play the flac file as it is I get nothing but static rain/snow noise).

Anyway, now FFDShow decodes and it is fine, but FFDShow reports 16bit while it should be 24bit
That's another problem of ffdshow. Actually most of ffmpeg (which ffdshow is based on) doesn't support 24bit audio at all. I strongly do not recommend using ffdshow for FLAC decoding. I have an alternative suggestion for you which should fix all the problems you had. See here:

http://forum.doom9.org/showthread.php?t=130498

[madshi]

to build avcodec.dll, the newly build avcodec.dll still doesn't work with the samples above and eac3to produces the same error, but ffmpeg works without complaining.
Ah, you little :devil: you've recompiled avcodec.dll and replaced the version which is shipping with eac3to. Did I allow you to do that? :angry:

Just joking, of course.

When compiling ffmpeg with the default options, you won't get 24bit TrueHD tracks decoded properly. Instead you'll get them only as 16bit. Because of that I've compiled ffmpeg with special config flags. Basically after ./configure you have to edit config.h and add the following two switches:

#define USE_HIGHPRECISION 1
#define CONFIG_AUDIO_NONSHORT 1

Afterwards you can use "make" as usual. This config change makes sure that the MLP/TrueHD decoder outputs the data as 32bit instead of as 16bit. This is necessary to get full 24bit from 24bit TrueHD tracks.

BTW, if you do recompile ffmpeg once in a while with always the latest MLP/E-AC3 decoder patches, I'd be thankful if you could send me the dlls, so that I can update them in my eac3to distribution. I hate recompiling C++ projects, so I'll only do it rarely.

[nautilus7]

Thanks, I'll forward this to the MLP/TrueHD decoder author.

I can only reproduce the problem with the end of the file, though. Your "start.thd" sample decodes identically with Nero/ffmpeg for me. Can you recheck the start, please?
Thanks. I 'll check again and make another sample if needed.

[puppydg68]

Can you please give me the full eac3to output with the Nero 7 (non lite) version? Did you properly buy the HD DVD plugin?

C:\HDTV-Tools\eac3to>eac3to.exe "C:\FEATURE_1_MERGED.mpa" "C:\Feature1.ac3" -640

E-AC3, 5.1 channels, 2:23:27, 1536kbit/s, 48khz, dialnorm: -27dB
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...

C:\HDTV-Tools\eac3to>

I did not buy the plug-in. I thought it only needed the nero audio decoder 2, to go from a demuxed file to ac3. Maybe this is the problem?

[madshi]

Thanks. I 'll check again and make another sample if needed.
That would be good. Thanks.

[madshi]

C:\HDTV-Tools\eac3to>eac3to.exe "C:\FEATURE_1_MERGED.mpa" "C:\Feature1.ac3" -640

E-AC3, 5.1 channels, 2:23:27, 1536kbit/s, 48khz, dialnorm: -27dB
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
Disabling DRC for Nero (E-)AC3 decoding...

C:\HDTV-Tools\eac3to>
Well, I don't see any complaint by eac3to. But there's also no "Done". That's a bit strange...

I did not buy the plug-in. I thought it only needed the nero audio decoder 2, to go from a demuxed file to ac3. Maybe this is the problem?
Yes, that's most probably the problem. You do need to buy the plugin and register it properly.

[nautilus7]

That would be good. Thanks.I can't reproduce the problem with a sample. Only the original (uncut) thd gives differences in the beginning of each wav channel. There are 4 different bytes at the first 48 bytes of each channel.

Ah, it just came to me!
It has to do with other problem (at the end of the file), i think. I believe these different bytes in the beginning have to do with the track runtime, which is different when i decode the original (uncut) .thd (by 1 ms) and exactly the same when i decode the sample.

So, nothing to worry for now, at least. If it's not fixed when you fix the "end" problem, then i 'll bring it back. :D

[madshi]

I can't reproduce the problem with a sample. Only the original (uncut) thd gives differences in the beginning of each wav channel. There are 4 different bytes at the first 48 bytes of each channel.

Ah, it just came to me!
It has to do with other problem (at the end of the file), i think. I believe these different bytes in the beginning have to do with the track runtime, which is different when i decode the original (uncut) .thd (by 1 ms) and exactly the same when i decode the sample.

So, nothing to worry for now, at least. If it's not fixed when you fix the "end" problem, then i 'll bring it back. :D
Yes, these are the "WAV data chunk size" and "WAV total file size" fields in the WAV file header.

[nautilus7]

2 hints

1. Feature request: I would like you to add an option to choose which channels to decode, when the output format is .wavs.

2. Shouldn't that be the opposite order (the bold letters), like with nero decoding?

C:\Tools>eac3to input.thd output.wavs -libav
TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
Writing WAVs...
Removing dialog normalization...
...

C:\Tools>eac3to input.thd output.wavs
TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
DirectShow reports 5.1 channels, 24 bits, 48khz
Writing WAVs...
...

[madshi]

1. Feature request: I would like you to add an option to choose which channels to decode, when the output format is .wavs.
What would that be good for? You can throw away the channels WAVs you don't like. I don't really like the idea to add an option for this because it would just make the interface more complicated without too much real life benefit. The only thing you'd gain is a tiny little bit of performance. Or am I missing an important advantage of such an option?

2. Shouldn't that be the opposite order (the bold letters), like with nero decoding?

C:\Tools>eac3to input.thd output.wavs -libav
TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
Writing WAVs...
Removing dialog normalization...
...

C:\Tools>eac3to input.thd output.wavs
TrueHD, 5.1 channels, 48khz, dialnorm: -27dB
Removing dialog normalization...
Decoding with DirectShow (Nero Audio Decoder 2)...
DirectShow reports 5.1 channels, 24 bits, 48khz
Writing WAVs...
...
Yeah, it should probably be the other way round. But dialog normalization removal, WAVs writing, DirectShow decoding etc are all separate modules in eac3to v2.x which work more or less next to each other. Because of that the order of the log messages is sometimes "funny". No need to worry about that, though...

[nautilus7]

What would that be good for? You can throw away the channels WAVs you don't like. I don't really like the idea to add an option for this because it would just make the interface more complicated without too much real life benefit. The only thing you'd gain is a tiny little bit of performance. Or am I missing an important advantage of such an option?
Almost every time i do a conversion, i like to decode the input and output files to wavs, and compare them in a wave editor like audacity. One channel is enough for this job and my hdd will write it faster. That's the only reason. :p


Yeah, it should probably be the other way round. But dialog normalization removal, WAVs writing, DirectShow decoding etc are all separate modules in eac3to v2.x which work more or less next to each other. Because of that the order of the log messages is sometimes "funny". No need to worry about that, though...
Yeah, it was obvious...

[Furiousflea]

Quick question...

Finally got everything working for EAC3 decoding, however all my HD DVD EAC3 tracks that I'm using are recognised as 24bit, is this normal? Or is possibly Nero HD DVD plugin doing something to them that it shouldn't?

Thanks...:)

[nautilus7]

Normal. All e-ac3 tracks are 24bit.

[madshi]

Almost every time i do a conversion, i like to decode the input and output files to wavs, and compare them in a wave editor like audacity. One channel is enough for this job and my hdd will write it faster. That's the only reason. :p
Maybe I can add an undocumented option for that... :) Which channel are you using? Center?

[nautilus7]

Yes, center. That would be great!

[madshi]

Quick question...

Finally got everything working for EAC3 decoding, however all my HD DVD EAC3 tracks that I'm using are recognised as 24bit, is this normal? Or is possibly Nero HD DVD plugin doing something to them that it shouldn't?

Thanks...:)
E-AC3 tracks don't really have a defined bitdepth. They are stored in a way where bitdepth doesn't have much meaning, I think. The decoder usually always outputs 24 bit or even floating point.

[nautilus7]

Madshi, i 've kept you busy with some minor issues and you probably missed post #1700, which is burning me!

[hristoff2]

Is there a proper way to make a 7.1 -> 6.1 downmix (for DTS-ES 6.1 Discrete encoding)?
(got the encoder, so that's not the problem)

If nobody 'complains' I'm gonna make the Cs [Center Sorround] channel by mixing BL & BR channel.

/edit
6.1 PCM: BL = BR = CS :)

[Thunderbolt8]

That's good to know, thanks. So it seems that for now the mathematical delay calculation (based on first timestamps) works alright. The only remaining question is whether we need to multiply that value by "29.97 / 23.976" or not.
woudlnt this only apply in case the movie would typically be 1080i with its 29.97 fps?

[madshi]

Is there a proper way to make a 7.1 -> 6.1 downmix (for DTS-ES 6.1 Discrete encoding)?
(got the encoder, so that's not the problem)

If nobody 'complains' I'm gonna make the Cs [Center Sorround] channel by mixing BL & BR channel.

/edit
6.1 PCM: BL = BR = CS :)
I'm not sure how many PCM tracks are 6.1 and how many are 7.1. I've checked The Descent and it's stored as 7.1, but BL and BR are identical and the case also sais 6.1. But still it's stored as 7.1. Maybe PCM 6.1 storage is not allowed? Anyway, as long as BL and BR are identical, of course you don't need to mix them... :) You can ask eac3to to give you mono wavs. You can then use a file compare tool to check whether BL and BR are identical or not.

[madshi]

woudlnt this only apply in case the movie would typically be 1080i with its 29.97 fps?
I had to do "*29.97/23.976" for Bourne Supremecy which clearly is not 1080i. So I'm not sure yet whether Bourne Supremecy was a strange case or whether we really always have to do "*29.97/23.976" or not.

[madshi]

Concerning trueHD sync issues...

I am very confused with The Matrix REV HD DVD. I have the original HD DVD and an encode of it which comes with dts sound. I didn't make this encode, but i know that the dts track was made from lossless trueHD track, with an eac3to version prior to trueHD Dialog Normalization fix. So i thought of making a flac track in order to replace the dts track.

The result is very confusing... The flac track is in sync in the 1st half (i guess the part that came from the first evo file), but de-sync (needs about 150 ms of delay) in the 2nd part. The really strange thing is that when i decoded both the dts and the flac tracks to wav and opened them in audacity, they had a constant 10 ms offset from beginning to end!!! Of course the dts track that was original included in the encode is in sync with the video.

Can you give me a logical explanation, please?
Sorry, missed this post.

Not sure what happened there. Let me guess: Probably the reencoder reencoded both EVO files separately. And then he probably joined the resulting fully reencoded parts into one big piece. I could imagine that with such a joined movie there are some "funny" things going on at the join point. E.g. it could be possible that the timecodes let the DTS decoder skip or repeat some frames at the join point. Of course I'm only guessing...

[nautilus7]

Not exactly what i waited to hear. :p
I 'll do more tests.

[TripleH]

The reason for that is that you most probably don't have a "FLAC source" filter installed which could handle standalone FLAC files. ffdshow can only handle FLAC files which are part of a container. ffdshow cannot handle standalone FLAC files.

Actually, I did have it installed while getting the noise.


That's another problem of ffdshow. Actually most of ffmpeg (which ffdshow is based on) doesn't support 24bit audio at all. I strongly do not recommend using ffdshow for FLAC decoding. I have an alternative suggestion for you which should fix all the problems you had. See here:

http://forum.doom9.org/showthread.php?t=130498

Anyway, I tried your flac decoder and works like a charm.

[Furiousflea]

Normal. All e-ac3 tracks are 24bit.

Thanks for the quick answer :)

However now got another problem...bearing in mind that the program has worked fine on a good 10 or so HD DVDs.

I'm now having a problem with a DTS Master Audio track on "Rambo First Blood Part II"

I'm using this....

eac3to rambo.dtshd rambo.wavs

Its reporting everything correctly, except when the decoding starts it goes very slow and gets gradually slower until it stops about an eigth way through the the track.

Its just on this track, I can decode other MA stuff fine?

Thanks, Rob

[nautilus7]

I want to remove Dialog Norm. from an ac3 track.
I must do input.ac3 output.ac3. Right?

C:\Tools>eac3to input.ac3 output.ac3
AC3, 5.1 channels, 1:48:51, 640kbit/s, 48khz
Creating/writing file "output.ac3"...
Done.

Shouldn't i get a message whether the track had Dialog Norm on/off?

[madshi]

Actually, I did have it installed while getting the noise.
Then most probably (for whatever reason) your media player has not used madFlac but another (bad) source filter instead.

[madshi]

However now got another problem...bearing in mind that the program has worked fine on a good 10 or so HD DVDs.

I'm now having a problem with a DTS Master Audio track on "Rambo First Blood Part II"

I'm using this....

eac3to rambo.dtshd rambo.wavs

Its reporting everything correctly, except when the decoding starts it goes very slow and gets gradually slower until it stops about an eigth way through the the track.

Its just on this track, I can decode other MA stuff fine?
Not sure where the problem comes from. Please try using "eac3to rambo.dtshd rambo.ac3", just to make sure that it's not eac3to's mono wav output which is stumbling. However, I rather guess that the Sonic Audio Decoder doesn't like this specific track for whatever reason. You could try only decoding the core, but obviously that'd be bad for audio quality... :(

[madshi]

I want to remove Dialog Norm. from an ac3 track.
I must do input.ac3 output.ac3. Right?

C:\Tools>eac3to input.ac3 output.ac3
AC3, 5.1 channels, 1:48:51, 640kbit/s, 48khz
Creating/writing file "output.ac3"...
Done.

Shouldn't i get a message whether the track had Dialog Norm on/off?
eac3to sais: "AC3, 5.1 channels, 1:48:51, 640kbit/s, 48khz". There's no dialnorm listed there, so this AC3 track is already dialnorm free. Let me guess: Is it from a Sony Blu-Ray disc? Sony doesn't use dialnorm.

[Furiousflea]

Not sure where the problem comes from. Please try using "eac3to rambo.dtshd rambo.ac3", just to make sure that it's not eac3to's mono wav output which is stumbling. However, I rather guess that the Sonic Audio Decoder doesn't like this specific track for whatever reason. You could try only decoding the core, but obviously that'd be bad for audio quality... :(


C:\EAC3To>eac3to audio2.dtshd rambo.wavs
DTS Master Audio, 5.1 channels, 16 bits, 1536kbit/s, 48khz
Decoding with DirectShow (Sonic Audio Decoder)...
DirectShow reports 5.1 channels, 16 bits, 48khz
Writing WAVs...
Creating/writing file "rambo.L.wav"...
Creating/writing file "rambo.R.wav"...
Creating/writing file "rambo.C.wav"...
Creating/writing file "rambo.LFE.wav"...
Creating/writing file "rambo.SL.wav"...
Creating/writing file "rambo.SR.wav"...
------------- <- stops here

Thanks for the reply, here's a copy of what gets reported mate :)

I extracted the core with no probs and that decoded to 6 mono wavs no probs....hmmmm

[nautilus7]

eac3to sais: "AC3, 5.1 channels, 1:48:51, 640kbit/s, 48khz". There's no dialnorm listed there, so this AC3 track is already dialnorm free. Let me guess: Is it from a Sony Blu-Ray disc? Sony doesn't use dialnorm.
But eac3to wrote a new file, which has some changed bytes (periodically). So...
It's not from a blu-ray. It's an hd dvd e-ac3 --> ac3 track, done with eac3to (a very very old version, prior to DRC fix).

[tebasuna51]

Ah, you little :devil: you've recompiled avcodec.dll and replaced the version which is shipping with eac3to. Did I allow you to do that? :angry:

Just joking, of course.

But the problem decoding the mlp samples from xkodi is using your avcodec.dll. I make test using samples download from this thread:
eac3to 440hz.mlp 440hz.wav -libav

The libav decoder output an unexpected bitdepth.
---[mlp @ 68A442E0]Lossless check failed - expected 0, calculated d2
...
The WAV writer didn't receive any audio data.
====================================================================
eac3to mlp51_id12.mlp mlp51_id12.wav -libav

The libav decoder output an unexpected bitdepth.
---[mlp @ 68A442E0]Lossless check failed - expected 0, calculated 7
...
The WAV writer didn't receive any audio data.
====================================================================
eac3to Acapela1.mlp Acapela1.wav -libav

The libav decoder output an unexpected bitdepth.
...
The WAV writer didn't receive any audio data.
====================================================================
eac3to "God Save The Queen.mlp" godgood.wav -libav -0,1,4,5,2,3

OK

Only last sample work (with the needed remap for info).

Where:
440hz.mlp stereo, 16 bits, 44100 Hz
Acapela1.mlp stereo, 16 bits, 44100 Hz
mlp51_id12.mlp 6chan , 16 bits, 44100 Hz
God Save The Queen.mlp 6chan , 24 bits, 96000 Hz

The four samples (and the other samples from xkodi) are decoded well by ffmpeg_eac3_mlp_r11045 from Kurtnoise with the know issues:
- Bad channelmapping with multichanel ID 20
- 0.6 ms added at end of files in samples from xkodi, if sources included and mlp encode are right.

BTW, the message for ID 20:
"5.1 wrong order channels"
I think must be replaced by a correct remapping, you are supplying now a incorrect wav and the user don't know what is wrong.
A decoder must accept and know the input formats and supply the output in correct format.

To see if the remap is make when output ac3 I try:
eac3to "God Save The Queen.mlp" god.ac3 -libav -resampleTo48000
MLP, 5.1 wrong order channels, 24 bits, 96khz
Resampling to 48khz...
Encoding AC3...
invalid sample rate
without success.

EDIT: If can help you to solve this last problem:
after the -resampleTo48000 I have a $temp subfolder with
$a23440.dll 219.136
$a232d4.dll 219.136
$a23168.dll 219.136
$a22efc.dll 219.136
All the same and with Properties: r8b.dll 4.0.3.17

[madshi]

But eac3to wrote a new file, which has some changed bytes (periodically). So...
It's not from a blu-ray. It's an hd dvd e-ac3 --> ac3 track, done with eac3to (a very very old version, prior to DRC fix).
AC3 files encoded by eac3to/Aften don't have dialnorm activated. Not sure exactly what bytes eac3to is changing. I guess maybe it rewrites dialnorm from -31dB to 0dB (which is basically the same thing).

[madshi]

But the problem decoding the mlp samples from xkodi is using your avcodec.dll.
Oh, then I misunderstood him.

I make test using samples download from this thread:
eac3to 440hz.mlp 440hz.wav -libav

The libav decoder output an unexpected bitdepth.
---[mlp @ 68A442E0]Lossless check failed - expected 0, calculated d2
...
The WAV writer didn't receive any audio data.
====================================================================
eac3to mlp51_id12.mlp mlp51_id12.wav -libav

The libav decoder output an unexpected bitdepth.
---[mlp @ 68A442E0]Lossless check failed - expected 0, calculated 7
...
The WAV writer didn't receive any audio data.
====================================================================
eac3to Acapela1.mlp Acapela1.wav -libav

The libav decoder output an unexpected bitdepth.
...
The WAV writer didn't receive any audio data.
====================================================================
eac3to "God Save The Queen.mlp" godgood.wav -libav -0,1,4,5,2,3

OK
Ok, thanks for the report. Looks like a "simple" bug in eac3to.

BTW, the message for ID 20:
"5.1 wrong order channels"
I think must be replaced by a correct remapping, you are supplying now a incorrect wav and the user don't know what is wrong.
A decoder must accept and know the input formats and supply the output in correct format.
Yes, I agree. I'll fix that in one of the next builds.

To see if the remap is make when output ac3 I try:
eac3to "God Save The Queen.mlp" god.ac3 -libav -resampleTo48000
MLP, 5.1 wrong order channels, 24 bits, 96khz
Resampling to 48khz...
Encoding AC3...
invalid sample rate
without success.
Looks like another "simple" bug... :)

EDIT: If can help you to solve this last problem:
after the -resampleTo48000 I have a $temp subfolder with
$a23440.dll 219.136
$a232d4.dll 219.136
$a23168.dll 219.136
$a22efc.dll 219.136
All the same and with Properties: r8b.dll 4.0.3.17
That's normal. The r8brain dll I'm using for resampling is not thread safe. So I'm loading multiple copies of that dll. That's the only way I can use r8brain from multiple threads at the same time without getting crashes. The file name of each dll copy is the id of the thread it's being called from... ;)

However, the $temp subfolder should be automatically deleted. Will check why that didn't work. Maybe it's only deleted after a successfull action? Probably...

[Beastie Boy]

Regarding movies where the audio drifts out of synch throughout the movie:
I originally had this problem with Batman Begin HD-DVD (vc1 video). The problem was caused by the Sonic video filter in the Graphedit chain. When I replaced this with with WMV Decoder DMO (or similar, I'm not at home at the moment), everything was OK.

There are a few posts on the forum regarding synch issues caused by the Sonic video decoder. Maybe this is the cause of some of these problems.

Cheers, Beastie.

[nautilus7]

AC3 files encoded by eac3to/Aften don't have dialnorm activated. Not sure exactly what bytes eac3to is changing. I guess maybe it rewrites dialnorm from -31dB to 0dB (which is basically the same thing).
It's possible because they both look the same in audacity.

[nautilus7]

I am doing an eac3 --> ac3 conversion using version 2.06 and the aften version that comes with eac3to (input .eac3 output.ac3).

Why the output ac3 file is 10-11 ms delayed compared to the input eac3?

Here's a sample: http://rapidshare.com/files/72646698/sample.eac3.html

[Icemaan]

Hallo to all

With the new Version of eac3to i get following error when I want encode one DD+ track into Dts or Ac3

The Format of the Source File Could not be detected.This Error I get when I want make
1. Ac3
2. Dts

With Old Version and aften/sox I can convert to ac3
But I want to encode to Dts

Whats going wrong .Please Help
Thanks
icemaan

[shambles]

i've got a truehd track that's 2.13gb and using nero filters, the flac i encoded from it turned out to be 3.26gb. using -libav, the flac is 1.88gb

i guess it's a 20bit track.. here's a 5mb sample http://rapidshare.com/files/72692697/001.thd.html

[tebasuna51]

I am doing an eac3 --> ac3 conversion using version 2.06 and the aften version that comes with eac3to.

Why the output ac3 file is 10-11 ms delayed compared to the input eac3?

How do you determine this delay?

I decoded your sample to wav and encoded to ac3.

The ac3 decoded to a new wav and is 5.333 ms delayed over the previous wav like was predictable.
Always Aften (and other commercial ac3 encoders) delay the output 5.333 ms, unless you use the parameter -pad 0.

[nautilus7]

How do you determine this delay?

I decoded your sample to wav and encoded to ac3.

The ac3 decoded to a new wav and is 5.333 ms delayed over the previous wav like was predictable.
Always Aften (and other commercial ac3 encoders) delay the output 5.333 ms, unless you use the parameter -pad 0.eac3 --> ac3 --> wav
eac3 --> wav

Then i compared the 2 wavs.

[tebasuna51]

eac3 --> ac3 --> wav
eac3 --> wav

Then i compared the 2 wavs.

Same procedure using:
eac3to sample.eac3 sample.ac3 -libav
eac3to sample.eac3 sample1.wav -libav
azid -d3/2 -L0 -l1 -ol,r,c,lfe,sl,sr sample.ac3 sample2.wav

Sample2.wav delayed 5.333 ms

If:
eac3to sample.eac3 sample1.wav -libav
aften -pad 0 sample1.wav sample.ac3
azid -d3/2 -L0 -l1 -ol,r,c,lfe,sl,sr sample.ac3 sample2.wav

Delay 0 ms.

Maybe Nero decoder (I haven't) introduce a new delay?

[nautilus7]

Possible. I don't know what to say.

What is the frame length that delaycut reports for every stream? I get 16 ms for eac3 and 32 ms for ac3. I got 5.333 ms for another eac3 track with a 1536 kbps bitrate.

[mutha88]

Please excuse my stupidity, but how can i encode to DTS? I have buyed and installed SurCode, but when I launch "eac3to" i can't see the DTS rate, filters and so one. I can't select them... where can i read a FAQ/guide for this wonderfull tool?

:stupid::scared:

[nautilus7]

What you need is in 1st post.

You can use: eac3to input.track output.dts -your bitrate

[mutha88]

What you need is in 1st post.

You can use: eac3to input.track output.dts -your bitrate


Well... this is a little light in the dark... but i will try...

By the way... i am using the GUI :rolleyes:

[shambles]

madshi, earlier in the thread you posted that you converted pirates of the caribbean 24bit pcm to flac and that the filesize was bigger than 2gb.. was that dead man's chest or curse of the black pearl?

the reason i'm asking is because i just converted the curse of the black pearl pcm track and the bitrate is 1395 kbps, filesize 1.39gb despite both the input pcm and the output flac track being 24bit. i sort of hope something has gone wrong at my end because having a 16 bit track in a 24 bit file would be amazingly stupid when it's pcm...

sample http://www.sendspace.com/file/4730s7

[mutha88]

Well... this is a little light in the dark... but i will try...

By the way... i am using the GUI :rolleyes:

Nothing happens. How can i make the GUI to find the surcodedvd.exe ?!:eek: