eac3to - audio conversion tool

[tebasuna51]

I wonder, does eac3to recognize AC3 4.0 channels correctly as such? what does 4.0 mean, 2 front and 2 rear channels? because eac3to recognized it as 3.1 so there might be a difference?

Yep, eac3to can difference 4 channel ac3 with:
AC3, 2/2 channels (FL,FR,BL,BR)
AC3, 3/1 channels (FL,FR,FC,BC)
AC3, 2/1.1 channels (FL,FR,LF,BC)
AC3, 3.1 channels (FL,FR,FC,LF)

for more info see this post:
http://forum.doom9.org/showthread.php?p=1134710#post1134710

[Thunderbolt8]

k, I guess that 3/1 thing should be correct then ;)

thanks!

[pascalwil]

Hi

Anyone to help me use SurCode DVD DTS Encoder alone (ie outside of eac3to) in Command Line mode?
I would like to batch convert the 6 wav files I create with eac3to into one DTSWAV file using a script instead of the Surcode GUI.
An example of the script to use would be most welcome to automate the process of creating DTS audio CDs from my 5.1 DVDs.
Thanks for your help.

Cheers

[leeperry]

am I going nuts, or do DTSHD(MA and HR)>FLAC conversions always end up w/ the dialogs center channel being too low?

TrueHD or LPCM>FLAC is perfectly fine(same dialogs volume as core DTS), and I've tried both Sonic 4.3 and Arcsoft(they give identical CRC btw).

I've encountered this problem many times, DTSHD>FLAC is unusable for me...I use ffdshow to decode if that means anything(DRC is disabled).

I've also tried to enable -keepDialnorm in eac3to, but it doesn't chance anything(same CRC for the FLAC).

[Thunderbolt8]

does it change when you use madflac to decode?

[leeperry]

yep, center is still too low compared to the core DTS...even w/ madflac.

this time it was a DTS HR 24bit track that I'm downsampling to 16 :

eac3to track.dts N:\eng_16.flac -sonic -down16
DTS Hi-Res, 5.1 channels, 1:51:33, 24 bits, 3840kbps, 48khz
(core: DTS, 5.1 channels, 1:51:33, 24 bits, 1509kbps, 48khz)
Decoding with DirectShow (Sonic Audio Decoder)...
DirectShow reports 5.1 channels, 24 bits, 48khz
Reducing depth from 24 to 16 bits...
Encoding FLAC with libFlac...
Creating file "N:\eng_16.flac"...
eac3to processing took 15 minutes, 19 seconds.
Done.

I'll try to come up w/ some proper sample, was just wondering if anyone else encountered this problem :)

[tebasuna51]

am I going nuts, or do DTSHD(MA and HR)>FLAC conversions always end up w/ the dialogs center channel being too low?

TrueHD or LPCM>FLAC is perfectly fine(same dialogs volume as core DTS), and I've tried both Sonic 4.3 and Arcsoft(they give identical CRC btw)...

I can't reproduce your problem:
Core dtshd
--------------- ------------------ --------------------
RMS power ch0: 3.21% (-29.88 dB) 3.40% (-29.37 dB)
RMS power ch1: 2.82% (-30.99 dB) 2.99% (-30.49 dB)
RMS power ch2: 4.93% (-26.15 dB) 5.22% (-25.64 dB)
RMS power ch3: 1.59% (-35.96 dB) 1.86% (-34.61 dB)
RMS power ch4: 0.88% (-41.11 dB) 0.93% (-40.60 dB)
RMS power ch5: 0.69% (-43.22 dB) 0.73% (-42.72 dB)
Please upload a sample.

[leeperry]

ok, thanks for the test...I was hoping for some measurements...how did you do that? you're comparing the FLAC'ed DTS-HD against the core DTS?

I've already had the problems many times on DTS HD tracks :(

for instance on this stream "DTS Master Audio, 5.1 channels, 24 bits, 48khz"(core: DTS, 5.1 channels, 24 bits, 1509kbps, 48khz), the dialogs were hardly audible, where they were perfectly fine on the core...and -keepDialnorm doesn't work(maybe by design?)

[Revgen]

^If you upload a sample, they may be able to help.

[tebasuna51]

I was hoping for some measurements...how did you do that? you're comparing the FLAC'ed DTS-HD against the core DTS?
I campare a dtshd decoded to wav with ArcSoft (DTS-MA) and decoded with libav (DTS core), the values are statistics from Wavosaur (free audio editor)
...and -keepDialnorm doesn't work(maybe on design?)
Don't use -keepDialnorm never with ac3 or dts (I don't know with TrueHD), let eac3to work.

[leeperry]

righty, I'll do my homework and come back w/ a proper sample...maybe ffdshow's mixer is playing tricks on me after all :confused:

I use my own stereo downmix Logic7 matrix, and this one works fine on all movies...except FLAC'ed DTS-HD :o

[Revgen]

1) TMT2 installs dtsdecoderdll.dll version 1.1.0.0
2) TMT3 installs dtsdecoderdll.dll version 1.1.0.1
3) TMT 3.0.1.133 Patch installs dtsdecoderdll.dll version 1.1.0.5

http://www.arcsoft.com/downloads/digitaltheatre/arcsoft/totalmediatheatre_beta_3.0.1.120_3.0.1.133_update_all.exe

you need to have installation of TMT2 on which eac3to recognize the dtsdecoderdll.dll, i.e. "eac3to -test" returns:



then just replace dtsdecoderdll.dll version 1.1.0.0 with version 1.1.0.1 or version 1.1.0.5 from TMT3 and then when you run "eac3to -test" you should get:



or



4) TMT 3.5 installs dtsdecoderdll.dll version 1.1.0.10, but it doesn't work with eac3to.

BTW, there is also new CinemasterAudio.DLL version 4.3.0.236 from the new version 2.6.3 of SONIC CineVision, which works good with eac3to. if you look here:

http://forum.doom9.org/showthread.php?p=1158014#post1158014

and here:

http://forum.doom9.org/showthread.php?p=1212881#post1212881

you will see that this is the first new version of CinemasterAudio.DLL from very long time that works good with eac3to

I'm having no luck getting arcsoft to work either. Am I screwed if I can't install TMT 2.0?

I bought TMT version 3.0 with the 1.1.0.5.dll and it will not work with EAC3 no matter what steps I take. Only the regular asdecoder.ax works in graphedit. Is there a way I can get a hand on a previous dtsdecoder.dll version?

[xkodi]

I'm having no luck getting arcsoft to work either. Am I screwed if I can't install TMT 2.0?

I bought TMT version 3.0 with the 1.1.0.5.dll and it will not work with EAC3 no matter what steps I take. Only the regular asdecoder.ax works in graphedit. Is there a way I can get a hand on a previous dtsdecoder.dll version?

look here:

http://forum.doom9.org/showthread.php?t=148324

[tebasuna51]

The spanish audio from Blu-Ray movie La Cosa is :
Bitrate 384Kbps
Sample Rate: 48KHz
Channels: 2
299MB, 1h 48min

Without problem when use ArcSoft decoder, but is buggy when use libav.

Audio sample (http://rapidshare.com/files/263798729/DTS_384_48Khz_2.0_La_Cosa_2_min.zip)

Using ffmpeg to decode (or Foobar2000 or MeGUI/BeHappy) is also OK.

[odin24]

When you encode with Surcode use the .cpt (compact) format instead the .dts (padded) format. After rename to dts.


I tried searching, but no luck, I even searched through google. I was curious as to what exactly is "DTS zero padding"?

Thanks.

[Octo-puss]

is Madshi still MIA?

[tebasuna51]

I tried searching, but no luck, I even searched through google. I was curious as to what exactly is "DTS zero padding"?.
These 4 dts files (20 sec long) was encoded with Surcode, the padded ones add 0's bytes after the core frame to reach the 2048 boundary (to easy write and search on hardware)

File ........: D:\Test\dts\Jap_padD.dts ... 3840000 bytes
----------------------------------------- First Frame Info
Primary Frame Byte Size .....: 2012 ( 2013 bytes/frame)
Transmission Bit Rate .......: 24 (1536 Kb/s)
...
--------------------------------------------- Revised Info
Total Frames ......: 1875
Duration ..........: 20 seconds. ( 0 h. 0 m. 20 s.)
Pad Bytes min./max.: 35 / 35 (bytes between frames)
------------------------------------------------- End Info

Real frame = Core frame + pad bytes = 2013 + 35 = 2048
Real bitrate = 3840000 x 8 / 20 = 1536 Kb/s
=================================================================

File ........: D:\Test\dts\dJap_cptD.dts ... 3774375 bytes
----------------------------------------- First Frame Info
Primary Frame Byte Size .....: 2012 ( 2013 bytes/frame)
Transmission Bit Rate .......: 24 (1536 Kb/s)
...
--------------------------------------------- Revised Info
Total Frames ......: 1875
Duration ..........: 20 seconds. ( 0 h. 0 m. 20 s.)
------------------------------------------------- End Info

Real Bitrate = 3774375 x 8 / 20 = 1509.75 Kb/s
=================================================================

File ........: D:\Test\dts\Jap_pad7.dts ... 3840000 bytes
----------------------------------------- First Frame Info
Primary Frame Byte Size .....: 1005 ( 1006 bytes/frame)
Transmission Bit Rate .......: 15 (768 Kb/s)
...
--------------------------------------------- Revised Info
Total Frames ......: 1875
Duration ..........: 20 seconds. ( 0 h. 0 m. 20 s.)
Pad Bytes min./max.: 1042 / 1042 (bytes between frames)
------------------------------------------------- End Info

Real frame = Core frame + pad bytes = 1006 + 1042 = 2048
Real bitrate = 3840000 x 8 / 20 = 1536 Kb/s
=================================================================

File ........: D:\Test\dts\Jap_cpt7.dts ... 1886250 bytes
----------------------------------------- First Frame Info
Primary Frame Byte Size .....: 1005 ( 1006 bytes/frame)
Transmission Bit Rate .......: 15 (768 Kb/s)
...
--------------------------------------------- Revised Info
Total Frames ......: 1875
Duration ..........: 20 seconds. ( 0 h. 0 m. 20 s.)
------------------------------------------------- End Info

Real bitrate = 1886250 x 8 / 20 = 754.5 Kb/s

[Pho3n1x]

I have one Question,is there a Decoder for TruHD wich decodes all 7.1 channels with eac3to to mono waves or is the only way 5.1 channel with the Nero or the libav decoder?

[tebasuna51]

I have one Question,is there a Decoder for TruHD wich decodes all 7.1 channels with eac3to to mono waves or is the only way 5.1 channel with the Nero or the libav decoder?
Eac3to decode TrueHD 7.1 to 8 wavs with libav without problems

[SomeJoe]

Don't use -keepDialnorm never with ac3 or dts (I don't know with TrueHD), let eac3to work.

For my AV receiver (Pioneer VSX-1018AH-K), TrueHD will not play properly unless it is extracted with eac3to using -keepDialnorm.

Have never had a problem letting eac3to remove it with ac3 or DTS, however.

[Snowknight26]

You sure you're using the latest version? Even when its removing dialog normalization, its setting the value to -31dB. In some earlier versions, it set it to 0dB, which might be causing your issue.

[Revgen]

is Madshi still MIA?

Madshi is currently fighting Krell on Altair IV. He'll be back next month.

[M-Blaster]

Hi, by downcoding Renaissance HD-DVD comes the following error.

eac3to l:\mainevo.evo e:\ren
EVO, 1 video track, 2 audio tracks, 1 subtitle track, 1:45:09
1: VC-1, 1080p24 /1.001 (16:9) with pulldown flags
2: DTS Master Audio, German, 5.1 channels, 16 bits, 48khz
(core: DTS, 5.1 channels, 16 bits, 1509kbps, 48khz)
3: DTS Master Audio, English, 5.1 channels, 16 bits, 48khz
(core: DTS, 5.1 channels, 16 bits, 1509kbps, 48khz)
4: Subtitle (DVD), German
Track 2 is used for destination file "Ren".
This audio conversion is not supported.

even the option -core does not help.

DOes somebody have an idea how to strip this movie evo?

[yesgrey]

I don't know if you can simply use e:\ren to specify the directory for the demuxed files.
Try switching to e:\ren directory and then run from there eac3to l:\mainevo.evo -demux

[tebasuna51]

Try with:
eac3to l:\mainevo.evo e:\ren.*

[M-Blaster]

how stupid I was :(

Now it runs with
"eac3to l:\mainevo.evo e:\ren.mkv" :D

[SomeJoe]

You sure you're using the latest version? Even when its removing dialog normalization, its setting the value to -31dB. In some earlier versions, it set it to 0dB, which might be causing your issue.

I normally use v3.14 because latest version has some type of problem where the last line of output to stdout is not present, so my front end cannot parse certain data.

But I have also tried v3.16 and get the same issue.

Apparently, my AV receiver is highly sensitive to a TrueHD stream. Unless -keepDialnorm is used, TrueHD output is very distorted.

It is no big deal, I have dialogue normalization turned off on the AV receiver anyway.

[pascalwil]

Hi

Anyone knows if/where eac3to source code is available? Thanks.

Cheers.

[kypec]

AFAIK the source code to eac3to has never been released by madshi. Considering how long he's not been around this thread it seems rather disturbing. Development of such a great tool should continue because even revision 3.16 is still not "perfect" I think and many things can be improved on it.

[leeperry]

I don't see why he should go GPL, he's very busy IRL...be glad he made it public in the first place!

[jj666]

It would be easier to do the same route as Neuron2 and make it payware.

-jj-

[M-Blaster]

hello I want to downmix a song from the credits of a mkv movie file.
Soundstream is 1,5 mBit DTS 5.1 with 24 bits.
I splitted the credits with mkvtoolnix.

eac3to e:\moviecredits.mkv e:\credits.flac -down2 -mixlfe
eac3to e:\moviecredits.mkv e:\credits.flac -down2
eac3to e:\moviecredits.mkv e:\credits.flac

Problem is that the classical song of the credits is ok, but when the song from the artist begins the sound is totally centered on the left.
I tried with the 3 parameters.
I don not habe arcsoft dts so that "libav" was used.

[Guest]

It would be easier to do the same route as neuron2 and make it payware. The donation I require is modest but it does give me an ongoing incentive to continue development and support. It doesn't earn me a fortune but it pays for my web hosting and some hardware/software purchases.

[jmvh]

Back to getting the best decoders for use with eac3to under Windows 7 x64...I can confirm the earlier posts that eac3to can work with the the TMT 2.0 codecs on Windows 7 x64 RTM. My PC originally had eac3to working with Nero 7 and TMT under winXP. I copied the required files from my winXP installation into the eac3to directory and used the Regserv32 command under Windows 7 x64.

All is good for DTS-MA using TMT. However, for the life of me, I can not get Nero 7 decoders to install. Despite successfully installing in my winXP, and activating with by BD/HD-DVD plugin, I can not repeat that under Win7. I've installed Nero as admin and normally, I've activiated the plug-in. But Showtime doesn't work (errors saying can't build graph on any content - DVD blu HD - and doesn't seem to recognize that BD is active since it doesn't have the BD enabled skin).

Anyone have success with Nero 7 win 7 x64 and care to share how you accomplished that feat?

[Jeff Flowerday]

Back to getting the best decoders for use with eac3to under Windows 7 x64...I can confirm the earlier posts that eac3to can work with the the TMT 2.0 codecs on Windows 7 x64 RTM. My PC originally had eac3to working with Nero 7 and TMT under winXP. I copied the required files from my winXP installation into the eac3to directory and used the Regserv32 command under Windows 7 x64.

All is good for DTS-MA using TMT. However, for the life of me, I can not get Nero 7 decoders to install. Despite successfully installing in my winXP, and activating with by BD/HD-DVD plugin, I can not repeat that under Win7. I've installed Nero as admin and normally, I've activiated the plug-in. But Showtime doesn't work (errors saying can't build graph on any content - DVD blu HD - and doesn't seem to recognize that BD is active since it doesn't have the BD enabled skin).

Anyone have success with Nero 7 win 7 x64 and care to share how you accomplished that feat?

How important is nero just for DD+ decoding? libavcodec will also decode DD+ by default with eac3to.

I've already ripped my HD-DVDs when my machine was vista now that it's 7 and I'll probably never rip any quantity of HD-DVDs again I'll live with libavcodec.

[nurbs]

libavcodecs dd+ support is not complete, but nobody ever provided sample files for the features that aren't supported so it's probably safe to assume there are no HD-DVDs that use them.

[Revgen]

^DD+ 7.1 audio is virtually non-existent. HD-DVD's never used it and Blu-Ray never will have it since the discs can hold up to 50gb. 7.1 audio typically uses True-HD or DTS-HD.

[Fiffy]

The libav E-AC3 codec does support 7.1, so that is not the problem. The concern is whether the decoder (which is still quite new) works correctly. Apparently it produces different output than the Nero codec (which is Dolby certified). Also take a look at this thread:

http://forum.doom9.org/showthread.php?t=145636

[xxx666yyy777]

I normally use v3.14 because latest version has some type of problem where the last line of output to stdout is not present, so my front end cannot parse certain data.

But I have also tried v3.16 and get the same issue.

Apparently, my AV receiver is highly sensitive to a TrueHD stream. Unless -keepDialnorm is used, TrueHD output is very distorted.

It is no big deal, I have dialogue normalization turned off on the AV receiver anyway.

SomeJoe - how do you turn if off? which receiver?

Thx

[<><]

I have retail version of Arcsoft TMT 2.0 installed on Win7 (with latest patch/update), added the bin path the path Environment Variable, but it still says that it is not installed when running the -test command....any ideas?

[Revgen]

The libav E-AC3 codec does support 7.1, so that is not the problem. The concern is whether the decoder (which is still quite new) works correctly. Apparently it produces different output than the Nero codec (which is Dolby certified). Also take a look at this thread:

http://forum.doom9.org/showthread.php?t=145636

libav does not support EAC3 (Dolby Digital Plus) 7.1. Only 5.1. It supports 7.1 for Dolby True-HD aka MLP.

http://forum.doom9.org/showthread.php?p=1313704#post1313704

AFAIK, there are no 7.1 EAC3 Blu-Rays or HD-DVD's. Discs that use 7.1 are usually Dolby True-HD or DTS-HD.

[jruggle]

The libav E-AC3 codec does support 7.1, so that is not the problem. The concern is whether the decoder (which is still quite new) works correctly. Apparently it produces different output than the Nero codec (which is Dolby certified). Also take a look at this thread:

http://forum.doom9.org/showthread.php?t=145636
Both AC3 and E-AC3 use random noise as part of the decoding process. From the tests I have done so far, the difference between the output of libavcodec and Nero is approx. the same as the difference between using 2 different types of random number generators in the libavcodec decoder. The type of RNG used is not defined in the specification. It even says that the level of the noise, which should be 0.707, can be approximated to 0.5 and still be compliant.

[RonaldoSan]

I always extract the best audio track to wav (better source material, better encoding result), unfortunately some Scandinavian Blu-rays only offer the standard AC-3 track (The Departed for instance). I noticed I was getting some large wav files from those tracks (compared to TrueHD tracks).

I'm not very confident in the Scandinavian publisher, besides lacking quality audio tracks on Blu-rays, The Dark Knight has a separate AC-3 track that is exactly the same as the embedded AC-3 track.

The Dark Knight

3: AC3, English, 5.1 channels, 640kbps, 48khz
4: TrueHD/AC3, English, 5.1 channels, 48khz
(embedded: AC3, 5.1 channels, 640kbps, 48khz)

Anyway, I extract the TrueHD and AC-3 tracks to wav and the output looks like this:

[a03] Extracting audio track number 3...
[a03] Decoding with DirectShow (Nero Audio Decoder 2)...
[a03] DirectShow reports 5.1 channels, 24 bits, 48khz
[a03] Writing WAV...
--------------------------------------------------------------
[a04] Extracting audio track number 4...
[a04] Extracting TrueHD stream...
[a04] Decoding with libav/ffmpeg...
[a04] Writing WAV...
--------------------------------------------------------------
AC3.wav - 6912 kbps 7,34 GB
TrueHD.wav - 4608 kbps 4,89 GB

TrueHD is lossless and AC-3 is compressed. When you extract to wav I assume you don't loose any information, so how come the AC-3 extraction is much bigger?

[tebasuna51]

...
AC3.wav - 6912 kbps 7,34 GB
TrueHD.wav - 4608 kbps 4,89 GB[/CODE]

TrueHD is lossless and AC-3 is compressed. When you extract to wav I assume you don't loose any information, so how come the AC-3 extraction is much bigger?

You decode to wav (not extract).
This TrueHD track seems have 16 bitdepth (audio samples with 16 bits), when decode from compressed audio you obtain 24 bitdepth then the size is 24/16 = 1,5 greater.

Of course the TrueHD.wav is the exact source audio.

[RonaldoSan]

You decode to wav (not extract).
This TrueHD track seems have 16 bitdepth (audio samples with 16 bits), when decode from compressed audio you obtain 24 bitdepth then the size is 24/16 = 1,5 greater.

Of course the TrueHD.wav is the exact source audio.
I see. Thanks for clarification.

[LeXXuz]

I have demuxed my Harry Potter 3 HD DVD with eac3to and got a TrueHD track with a delay of 2002ms.
I got a file named: "feature_1 - 4 - TrueHD, 5.1 channels, 48khz DELAY 2002ms.thd".
As a next step I wanted to transcode this file with eac3to into dts (using SurCode).
From what I understand eac3to first decodes this TrueHD track into several WAVE tracks and feeds them to the Surcode Encoder afterwards.
Is there a way to get rid of this delay during this process?

[nurbs]

Look at the first post. I assume adding +2002ms to the command line would do the trick.

[LeXXuz]

Look at the first post. I assume adding +2002ms to the command line would do the trick.

Ah thank you. I must have missed this as I looked into the command line options last night. :o

Why was the delay not removed during demuxing like for the E-AC3 tracks?

[TinTime]

Why was the delay not removed during demuxing like for the E-AC3 tracks?

I think eac3to can't apply delays to TrueHD tracks - that's why the delay is written to the file name.

However, as you're converting to DTS anyway why not go straight from the source EVOs to DTS, instead of demuxing the TrueHD track first? If you transcode then eac3to will apply the delay as needed.

[LeXXuz]

However, as you're converting to DTS anyway why not go straight from the source EVOs to DTS, instead of demuxing the TrueHD track first? If you transcode then eac3to will apply the delay as needed.

Because Nero AC3 Decoder does not allow to process more than one track at a time. That's why I first demux all the tracks I need (usually 2 E-AC3 tracks at least + TruHD) and process them afterwards.