eac3to - audio conversion tool

[tebasuna51]

Even eac3to have problems with eac3 in mkv container (not in m2ts).
And default Aften encoder is deprecated.

The best way I know to recode is:

FFMPEG -drc_scale 0 -i "input.mkv" -vn -acodec ac3 -ac 6 -ab 640k -center_mixlev 0.707 "output.ac3"

or the desired bitrate
or -map 0:2 instead -vn if is the second audio track in mkv, or nothing if input.eac3

[Ripman]

Is there an option that can be passed on the command line to tell whether a particular constant 24bit input audio file contains only 16bits of audio data - that there are "superfluous zero bytes"?

eac3to will remove these "superfluous zero bytes" using a second pass - a great feature. And I am wondering if there is an option to tell whether there are "superfluous zero bytes", but without taking the action of actually removing the "superfluous zero bytes".

Here is an example log:

eac3to v3.31
command line: eac3to  "02_Schumann-Violin-Sonatas-Op-105-and-Op-121.flac" "..\wav\02_Schumann-Violin-Sonatas-Op-105-and-Op-121.wav"
------------------------------------------------------------------------------
FLAC, 2.0 channels, 0:03:26, 24 bits, 2640kbps, 352.8kHz
Decoding FLAC...
Writing WAV...
Creating file "..\wav\02_Schumann-Violin-Sonatas-Op-105-and-Op-121.wav"...
The original audio track has a constant bit depth of 16 bits.
Superfluous zero bytes detected, will be stripped in 2nd pass.
Starting 2nd pass...
Reading WAV...
Stripping zero bytes...
Writing WAV...
Creating file "..\wav\02_Schumann-Violin-Sonatas-Op-105-and-Op-121.wav"...
eac3to processing took 23 seconds.
Done.

[Q-the-STORM]

-24 should encode with 24 bitdepth, but I don't know if the "superfluous zero bytes" message comes up, you'd need to try it yourself...

[Ripman]

Originally Posted by Q-the-STORM

-24 should encode with 24 bitdepth, but I don't know if the "superfluous zero bytes" message comes up, you'd need to try it yourself...

It'd be more of a "stat" type of function ... the implication being that you need to go through the file anyway to determine that it doesn't have anything over 16its.

I can process something like the log I posted, and then use an re to match the string with the message. -- I was looking for something more precise.

Btw, here is an example of such a file - 24bit/352.8/2ch flac. I did a post maybe a year ago where I had a problem with 352khz files.
https://www.dropbox.com/s/uuhv97hzo2...hDXD.flac?dl=1

[Snowknight26]

Originally Posted by Ripman

Is there an option that can be passed on the command line to tell whether a particular constant 24bit input audio file contains only 16bits of audio data - that there are "superfluous zero bytes"?

You have to decode the whole audio track, just like you did.

Originally Posted by Ripman

eac3to will remove these "superfluous zero bytes" using a second pass - a great feature. And I am wondering if there is an option to tell whether there are "superfluous zero bytes", but without taking the action of actually removing the "superfluous zero bytes".

-no2ndpass, stops the 2nd pass from occurring, but what's your goal behind all this anyway?

Originally Posted by Q-the-STORM

-24 should encode with 24 bitdepth, but I don't know if the "superfluous zero bytes" message comes up, you'd need to try it yourself...

-24 defines a PCM input track as 24-bit in case eac3to can't automatically detect it. You're thinking of -down24, which reduces the output bit-depth to 24.

[Q-the-STORM]

Originally Posted by Snowknight26

-24 defines a PCM input track as 24-bit in case eac3to can't automatically detect it. You're thinking of -down24, which reduces the output bit-depth to 24.

yeah, I meant -down24

[Music Fan]

@ MrVideo : try to keep your eac3 track as is before to recode it, it will perhaps be compliant with your decoder, even if you wanna put it on a dvd-video.
I know one example of commercial dvd with ac3 @ 640 kbps : Pink Floyd, Pulse (at least the european edition). It contains 2 ac3 tracks, one @ 640 kbps, the other @ 448 kbps. And it's decoded by my equipment.
But this is probably a different kind of encoding, a "simple" ac3. Yours was maybe encoded with an eac3 encoder, unless there's a header misinterpretation of your file wich is maybe also a "simple" ac3.

[MrVideo]

Originally Posted by tebasuna51

The best way I know to recode is:

FFMPEG -drc_scale 0 -i "input.mkv" -vn -acodec ac3 -ac 6 -ab 640k - 0.707 "output.ac3"

Thanks. I will put that into my scripts.

Why use -drc_scale if it is set to 0? Isn't that the default?

If the audio being recoded is from EAC3, should the -center_mixlev not be used, as it should already be that from the time it was initially encoded?

[tebasuna51]

Originally Posted by Ripman

eac3to will remove these "superfluous zero bytes" using a second pass - a great feature. And I am wondering if there is an option to tell whether there are "superfluous zero bytes", but without taking the action of actually removing the "superfluous zero bytes".

That is the correct answer:

Originally Posted by Snowknight26

-no2ndpass, stops the 2nd pass from occurring, but what's your goal behind all this anyway?

And the correct question also.

[tebasuna51]

Originally Posted by MrVideo

Why use -drc_scale if it is set to 0? Isn't that the default?

The default decoding is apply the Dynamic Range Compression (DRC) found in the stream (attenuate volume), set to 0 cancel that action.

If the audio being recoded is from EAC3, should the -center_mixlev not be used, as it should already be that from the time it was initially encoded?

The center-mixlevel is a metadata present in AC3 5.1 header to be used when the 5.1 is downmixed to 2.0.

The ffmpeg default put -4.5 dB instead -3 dB and I recommend use the big value to avoid low dialog volume problem. Other encoders use -3 dB by default, I don't know for what ffmpeg put -4.5.

[Ripman]

Originally Posted by Snowknight26

-no2ndpass, stops the 2nd pass from occurring, but what's your goal behind all this anyway?

I didn't know about the -no2ndpass -- it doesn't show on the options screen (or even page 1 of this post). Anyway, using that opt prevents the 2nd pass.

Goal? --- Get my money back from the site that sold the file as 24bit. (The idea would be to use eac3to to confirm bit-depth before processing any files I purchase online. In the case of the 16bit-352.8khz flac I posted, I imagine that the producing firm had an encoder option set incorrectly.)

[Music Fan]

Originally Posted by MrVideo

Thanks. I will put that into my scripts.

Did you see my post ?

[MrVideo]

Originally Posted by tebasuna51

(snip)

Thanks for the info. Mucho appreciated.

[MrVideo]

Originally Posted by Music Fan

@ MrVideo : try to keep your eac3 track as is before to recode it, it will perhaps be compliant with your decoder, even if you wanna put it on a dvd-video.

Sorry, missed this post at first. I don't do DVD-videos. I do Blu-ray. My amp has no issue with the variant EAC3 stream.

The program I use is BDRebuilder and jdobbs is a stickler for files being Blu-ray compliant. His latest release is supposed to now handle EAC3, but he wants the compliant version, so he tried to make it so and it fails. I've asked to use the stream as is, with an option to recode if the user so desires.

Yours was maybe encoded with an eac3 encoder, unless there's a header misinterpretation of your file wich is maybe also a "simple" ac3.

It is EAC3, but not the Blu-ray compliant kind.

[Music Fan]

Ok. Your player could maybe play it anyway, even on a Blu-ray. You can try with a short extract that you export with TSMuxer in Blu-ray mode (and burn it with ImgBurn for instance), it will keep sound as is.

[MrVideo]

Originally Posted by Music Fan

Ok. Your player could maybe play it anyway, even on a Blu-ray. You can try with a short extract that you export with TSMuxer in Blu-ray mode (and burn it with ImgBurn for instance), it will keep sound as is.

Yes, it will, but it won't be an authored disc with a menu and multiple files in the menu to select for playback, i.e., BDRebuilder.

[Q-the-STORM]

If I remember correctly, you should be able to simply replace the m2ts file... so use your regular software to create the BD, but then remux the m2ts with tsmuxer....

[iSeries]

Hi, I'm extracting a DTS-HD MA 7.1 audio track to 7.1 WAVs, but eac3to is getting about 10% of the way in and then saying that clipping is detected and that a second pass will be necessary. Then at the end it does the second pass and applies a gain of -0.31db. I've never seen this before, is this normal?

[tebasuna51]

Is not normal with a DTS-MA, but I see some DTS-MA bad encoded with this problem.

You can let the -0.31 dB (little atenuation) or use the parameter -no2ndpass to obtain the full volume with a little clip (unnoticeable most the times).

[iSeries]

Thanks, actually I got it wrong, it's a DTS-X soundtrack (Jason Bourne). Are there any known issues decoding this format?