eac3to - audio conversion tool

[laserfan]

I recently needed to re-work a BD and keep the TrueHD audio track, and the demuxed result wouldn't re-mux properly. I instead used tsMuxeR to demux and that worked perfectly.

I wonder: does eac3to demux TrueHD tracks (AS TrueHD tracks, not just the AC3 core) properly, and I just had a one-off glitch of some sort?

If demuxing TrueHD *is* a problem for eac3to, are there any other such demuxing gotchas that are known?

[tebasuna51]

I don't know problems demuxing TrueHD tracks with eac3to.

[laserfan]

Originally Posted by tebasuna51

I don't know problems demuxing TrueHD tracks with eac3to.

Thanks.

[Bigmango]

Originally Posted by kevmitch

While waiting to be able to post here for the first time, I posted over at avsvorums about a problem I noticed trying to decode the DTS HD MA sound from the Obilivion blu ray. In short, the center channel becomes distorted in loud bits. Since that posting, I have realised that specifying '-core' option to eac3to fixes the problem but of course its not "lossless", which my decode of the MA track clearly isn't either in this case.

I have uploaded a sample .dtshd file from 8:10-8:20 of the film which exhibits the problem. Also included with the sample are my eac3to decodes of the centre channel with and without "-core" as well as the libav decode, which also exhibits the distiortion. And of course all the logs and eac3to -test results.

I get the same results (bit-for-bit) using eac3to with the Arcsoft decoder versions 1.1.0.0, 1.1.0.7 and 1.1.0.8. I also hear the distortion when decoding the .dtshd directly in mplayer, vlc and audacity (which I think use the same code as libav anyway).

EDIT: On second thought, the distortion is there in core track as well. I don't know how I didn't hear that before.

Same problem here.

I seems the Arcsoft decoder doesn't decode some tracks properly and Oblivion is one of them.

It's amazing that in 2013 we still don't have a reliable way to decode DTS-MA to wav/raw on the PC.

Have you tried the Arcsoft decoder 1.1.0.9 ?

Does anyone finally know what improvements or regressions version 1.1.0.9 brings?

[kukushka]

Originally Posted by Bigmango

Same problem here.

probably it's worth to give it a try with sonic too. but if every possible decoder decode same distorted original, it probably was there initially..

[Bigmango]

Originally Posted by kevmitch

While waiting to be able to post here for the first time, I posted over at avsvorums about a problem I noticed trying to decode the DTS HD MA sound from the Obilivion blu ray. In short, the center channel becomes distorted in loud bits. Since that posting, I have realised that specifying '-core' option to eac3to fixes the problem but of course its not "lossless", which my decode of the MA track clearly isn't either in this case.

I have uploaded a sample .dtshd file from 8:10-8:20 of the film which exhibits the problem. Also included with the sample are my eac3to decodes of the centre channel with and without "-core" as well as the libav decode, which also exhibits the distiortion. And of course all the logs and eac3to -test results.

I get the same results (bit-for-bit) using eac3to with the Arcsoft decoder versions 1.1.0.0, 1.1.0.7 and 1.1.0.8. I also hear the distortion when decoding the .dtshd directly in mplayer, vlc and audacity (which I think use the same code as libav anyway).

EDIT: On second thought, the distortion is there in core track as well. I don't know how I didn't hear that before.

After some more testing with the full movie, this seems to be a problem with the ffmpeg decoder.

BTW, as you are saying the tiny distortion you are getting in your sample is in the original track as well. But, starting at the same time up to 20 seconds later (not in your sample) there often is A LOT of sound clipping (and in other parts of the movie too).

I seem to be getting this with ffmpeg based decoders, and then not every time. PowerDVD plays it fine (also the FLAC).

So, as far as I can see, this is a bug in ffmpeg.

[nevcairiel]

How is it a bug in ffmpeg if the original problem was the ArcSoft decoder causing the issue?

[Bigmango]

Originally Posted by nevcairiel

How is it a bug in ffmpeg if the original problem was the ArcSoft decoder causing the issue?

I first thought Arcsoft was the issue (I played my Oblivion with FLAC backup on XBMC), but after some more testing the ffmpeg based players (tested XBMC, MPC-HC have the problem) seem to be the only ones having the playback problem (the tested XBMC which is playing back the FLAC as PCM on a hardware receiver is on a totaly different system than the tested MPC-HC, totaly different hardware and drivers - both have the problem.).

Furthermore, PowerDVD plays the eac3to+Arcsoft transcoded FLAC fine whitout the clipping that's sometimes produced by the ffmpeg players, and Oblivion is the only movie having this issue (in a media library of ~300 movies).

So, if PowerDVD plays the (Arcsoft decoded) FLAC fine but 2 different ffmpeg players on 2 different systems don't, ffmpeg is the issue.

[nevcairiel]

So you're saying ffmpegs FLAC decoder is the issue, right? Not actually related to DTS directly at all?
I tried listening to the FLAC center channel samples posted above, and i couldn't really hear any obvious distortions, but maybe i just dont know what to listen for (or my player is fixed)

[Sparktank]

Regarding distortion: I've been meaning to start a new thread with heavy investigation into this.
Intense testing with libav/arcsoft/sonic/etc.

I didn't observe any (for Oblivion) using Foobar2000 with the center-FLAC.
I can't remember if it was apparent in MPC-HC (FFDshow), but I don't remember there being any.
The distortion wasn't immensely apparent for me as it was for you.

Even opening in Audacity it didn't show any clipping (marked in red). It does show that it's cut off at .8 ("The vertical scale displays amplitude when showing the waveform" from Audacity Help)


Using PaulStretch (to slow down the audio by repeating samples numerous times) shows clipping when slowed down to 10x the normal rate.


Which is where you hear the perceived distortion ("perceived" as in it's not apparent to everyone).

----

This could be a result of various setups of software/hardware. (why some can hear it and some cannot)
I've disabled all "effects" that my soundcard has to offer (ie, "loudness equalization", etc).
I've even disabled any normalization or regain effects in FFDShow when playing back the FLAC in MPC-HC.
Also MPC-HC has all its internal effects disabled to play back the audio with the lest altering/manipulation.

I haven't done any testing with the elementary stream (dts-hd ma) though.
I have Arcsoft 1.0.0.9 and can do a test to compare but I don't have any detailed knowledge or software to really examine the differences between the other versions.
But Foobar2000 has a "compare two files" to detect any offset in volume, etc.

----

I observed the same thing with Tron Legacy. Although, with TL, the distortion is much more apparent than that of Oblivion.
After looking at the Oblivion samples (which played fine with barely any distortion in Foobar), I went to rip my BD of TL to test for distortion.

It seems that depending on the player it (Tron Legacy) did have severe bass distortion.
I observed distortion on different instances. Notably when playing back on MPC-HC with FFShow.

Playing back a converted WAV (5.1 (down6) or 7.1) on Foobar doesn't have any distortion but playing the WAVs on MPC-HC did however (FFDShow again).
I have not tried using LAV filters though or other.
Downmixing (either 7.1 or converted 5.1) to 2.0 (-downStereo; -7.xxdb automatically applied with 2nd pass) did playback on all media players without any distortion.
---
I'm still playing around with different setups to test with the most variables.
I'm even running through SoX for "stat" and "stats" (though that info is quite useless to me until I learn it).

But the distortion seemed very conditional.
I just need to remember to document everything I did to see where distortion was obvious (and not, in case of Foobar??)

Though I'll be creating a new thread for this with mega details.
I'm really curious about some movies having this weird distortion anomaly (fact: some hear it; some don't).

X-Men: First Class has it at times as well.
I can hear it when watching the movie on my system (but since I upgraded my BD player, it might be a different experience now).

Sorry for the off-topic details, will create a new thread.
This is highly intriguing.

[madshi]

If the problem is in the ffmpeg decoder then using madFlac should work without issues?

[nevcairiel]

The ffmpeg FLAC decoder is perfectly lossless, so its more likely some other combination of the playback chain.

[Bigmango]

Originally Posted by nevcairiel

So you're saying ffmpegs FLAC decoder is the issue, right? Not actually related to DTS directly at all?

Yes, or something else in ffmpeg.

Originally Posted by nevcairiel

I tried listening to the FLAC center channel samples posted above, and i couldn't really hear any obvious distortions, but maybe i just dont know what to listen for (or my player is fixed)

Yes, these samples don't have the distortion. There is some kind of "stuttering" at one point , but it is also present in the original DTS track.

What I am however experiencing, is sound "clipping" (some very high and loud fast "pops" or "ticks") which happens several times starting at around 08:05 to 08:30 in the movie (and at some other places in the movie too when there's some high pitched sound).

edit: I think this is what Sparktank explains above.

[Bigmango]

Originally Posted by nevcairiel

The ffmpeg FLAC decoder is perfectly lossless, so its more likely some other combination of the playback chain.

If this is really the case (?) then it should be something else in the ffmpeg playback chain as Openelec (XBMC) playing back as PCM on a hardware receiver and MPC-HC are both having the issue, on totally different hardware (+ Openelec is Linux and MPC-HC is on Windows).

[Bigmango]

Originally Posted by Sparktank

Which is where you hear the perceived distortion ("perceived" as in it's not apparent to everyone).

It comes out very lound and clear on my 7.1 home theater system (pay back with Openelec XBMC mediacenter).

On windows MPC-HC with headphones it is less noticeable, but still there. However, the strange thing is that sometimes the clipping is there, and sometimes it isn't. This needs more testing...

[tebasuna51]

I make some test over the dtshd sample included. And I want post my opinion about this.

1) Decoded to wavs directly (ArcSoft 1.1.0.0 and libav) result the same output than decode the Flacs included in the sample.
For me don't exist problems related to Flac encoder/decoders.

2) Decoding the "core" with ArcSoft produce the same (more or less) output than decoding with ffmpeg/libav (or libdts with NicAudio).
Then I can't suspect problems with ffmpeg/libav decoder.

3) There are clip (see attached image) in the wave form.
This can be listened like distort sound, more or less clear depending on audio system chain.
But the clip isn't at 0dB, that suggest the problem was in the original source and isn't a decoder problem.

4) But there are a problem that I never see. The lossless output from DTS-MA decoded with ArcSoft is 3dB loud than the decoded core (ArcSoft or libav).
Not only on clip parts but over the full sample.

I checked all my DTS-MA samples and always the volume is the same decoding lossless or core.

The attached image is from my decoded wavs but is the same than flacs included in the sample.

[Anakunda]

Hello, may I get some help how to properly setup eac3to for conversion 7.1 channels audio, whenever I try to convert 8-channel track (DtsHD master audio) to AAC it always produces useless 8channel digital noise. All 6-channel tracks convert nicely so the problem must have to do something with the channels layout. Maybe wrong channels mapping? Or is that encoder bug? Is Arcsoft decoder decoding the audio improperly?
Using pipe, the only way I know:

C:\Users\Anakunda\AppData\Local\Temp>benice "I:\video\(workset)\MARVELS_IRON_MAN_3_BD_BLUEBIRD\english71.dtshd" -d -55 --down16
Executing eac3to "I:\video\(workset)\MARVELS_IRON_MAN_3_BD_BLUEBIRD\english71.dtshd" stdout.wav -55ms -down16 -normalize | D:\media\AAC\qaac.exe --ignorelength --threading --verbose --no-delay --rate keep --quality 2 --tvbr 100 -o "I:\video\(workset)\MARVELS_IRON_MAN_3_BD_BLUEBIRD\english71.m4a" -

eac3to v3.27
command line: D:\media\eac3to\eac3to.exe "I:\video\(workset)\MARVELS_IRON_MAN_3_BD_BLUEBIRD\english71.dtshd" stdout.wav -55ms -normalize
------------------------------------------------------------------------------
DTS Master Audio, 7.1 channels, 24 bits, 48kHz
(core: DTS, 5.1 channels, 24 bits, 1509kbps, 48kHz)
Decoding with ArcSoft DTS Decoder...
Applying RAW/PCM delay...
Writing WAV...
Creating file "stdout.pass1.wav"...
The original audio track has a constant bit depth of 24 bits.
Starting 2nd pass...
Reading WAV...
Reducing depth from 64 to 24 bits...
Writing WAV...
Applying -0.09dB gain...
Creating file "stdout.wav"...
The processed audio track has a constant bit depth of 24 bits.
eac3to processing took 49 minutes, 59 seconds.
Done.

qaac 2.19, CoreAudioToolbox 7.9.8.3
english71.m4a
Prepend 960 samples (0.02 seconds)
Using default channel layout.
Output layout: 7.1 (C Lc Rc L R Ls Rs LFE)
Enable threading
AAC-LC Encoder, TVBR q100, Quality 96
2:10:32.839 (4.6x)
375976304/-1 samples processed in 28:09.990
Overall bitrate: 665.123kbps
7812/7812 chunks written (optimizing)
Encoding done.

[tebasuna51]

Originally Posted by Anakunda

Hello, may I get some help how to properly setup eac3to for conversion 7.1 channels audio, whenever I try to convert 8-channel track (DtsHD master audio) to AAC it always produces useless 8channel digital noise.

Without problems here with a simple channels test DTS-MA 7.1.

The ArcSoft (1.1.0.0) output is correct (FL FR FC LF BL BR SL SR) the qaac (2.19) Output layout: 7.1 (C Lc Rc L R Ls Rs LFE) is the default like you can see here, but is played correctly in my 7.1 system with mpc-hc, and decoded correctly with NeroAacDec.

Try without --threading to see if there are problems in your system with pipe and multithread with big files.
Try also decode the DTS-MA to wav and encode the wav file.

[Anakunda]

Originally Posted by tebasuna51

Try without --threading to see if there are problems in your system with pipe and multithread with big files.

Thanks dude, don't know what's --threading exactly for but it seems without it the output now is perfect
For some reason with 6channel audio qaac with threading support worked correctly.

[carlmart]

It's been a long time since I could use eac3to. I wonder what might be happening.

I have already installed both eac3to and put the GUI on the same folder. But when I load a file as I remember I did, something seems to go wrong because it doesn't accept any command any longer.

Sorry if this question is so primary.