eac3to - audio conversion tool

[Thunderbolt8]

whats the best solution again to deal with 7.1 DTS-HD MA tracks which are described as "strange setup" ? I want to convert the track to flac, so what are my options here?

is this information here still correct? http://forum.doom9.org/showthread.ph...93#post1361493 using sonic to convert it to 5.1 makes most sense atm?

[Laurent]

Originally Posted by Laurent

I noticed yesterday a similar problem with many macroblocks in video if I demux video in MKV file rather than a H.264 file. As my process is a 2 steps process using first eac3to and then TsMuxer to produce a M2TS file, I cannot be sure if the problem with video in MKV container is due to eac3to or TsMuxer.

I come back to this problem with more explanations and tests. In fact, the problem is not new, I just verified and I have the same strange behaviour with eac3to v3.21 and eac3to v3.18.

Here is my scenario:
1 - eac3to to demux video in MKV container and convert audio to AC3
2 - TsMuxer to mux audio and video in a M2TS container
3 - playback of the M2TS file with the PS3

Video is H.264. When the file is decoded by the PS3, I have many macroblocks, and the result is the same whatever the version of eac3to (3.18, 3.21 and 3.22).
If at step 1, instead of demuxing video in MKV container, I demux the video in a H.264 file, then the playback with the PS3 is fine without macroblocks.

Note that all these files (M2TS and MKV) are played correctly on the PC.

So the problem is relative to the usage of MKV container, either when produced by eac3to, or when used as input by TsMuxer. And the problem is noticeable only when the final file is played with the PS3.

Now that I discovered this problem, I will of course avoid using MKV container.

[b66pak]

eac3to use haali media splitter for demuxing...try other version (11-1-2009)

[Snowknight26]

No it doesnt.

[Laurent]

Originally Posted by b66pak

eac3to use haali media splitter for demuxing...try other version (11-1-2009)

This is the version I am using.

[asarian]

Originally Posted by Laurent

I come back to this problem with more explanations and tests. In fact, the problem is not new, I just verified and I have the same strange behaviour with eac3to v3.21 and eac3to v3.18.

Here is my scenario:
1 - eac3to to demux video in MKV container and convert audio to AC3
2 - TsMuxer to mux audio and video in a M2TS container
3 - playback of the M2TS file with the PS3

Video is H.264. When the file is decoded by the PS3, I have many macroblocks, and the result is the same whatever the version of eac3to (3.18, 3.21 and 3.22).
If at step 1, instead of demuxing video in MKV container, I demux the video in a H.264 file, then the playback with the PS3 is fine without macroblocks.

Note that all these files (M2TS and MKV) are played correctly on the PC.

So the problem is relative to the usage of MKV container, either when produced by eac3to, or when used as input by TsMuxer. And the problem is noticeable only when the final file is played with the PS3.

Now that I discovered this problem, I will of course avoid using MKV container.

Hmm, you wouldn't be using an older version of CoreAVC, would ya? What you describe is a known issue with it. Just make sure you let ffdshow handle H264, and you should be fine. I stopped using CoreAVC for this precise reason.

[Laurent]

Originally Posted by asarian

Hmm, you wouldn't be using an older version of CoreAVC, would ya? What you describe is a known issue with it. Just make sure you let ffdshow handle H264, and you should be fine. I stopped using CoreAVC for this precise reason.

But the playback problem is with the PlayStation, not the PC. And when demuxing with eac3to, I don't think a decoder is involved.
By the way, CoreAVC is not installed on my PC.

[GoodzMastaJ]

I've extracted the TrueHD track from Evangelion 1.11 blu-ray using eac3to. ffdshow reports it as having channels as shown in the screenshot below.



Note the presence of the side channels and absence of an LFE channel.
In actuality, there are no side channels but there is an LFE and a rear center channel which I suppose would make this a 6.1 track (rather than the 7.0 I'm getting from ffdshow). Using the mixer, I found the LFE channel is playing from what ffdshow thinks is Back Left (all that comes out of BackL is rumbling).

I'm not sure where the bug is, if it's in eac3to or ffdshow. Any ideas on how to pinpoint it? I made a similar post on the ffdshow-tryouts forum to stir up some suggestions there as well.

Here is a short sample that has a pretty good demonstration.
http://www.mediafire.com/?mmzydemji5m

[Killroy™]

Originally Posted by asarian

Using eac3to, any way I can downconvert a ~25GB 192 Khz (!) LPCM track (Akira Blu-Ray) to something more human, like, say, 48 Khz?

OK... I'll be the first to ask... Why would you want to take such a beautiful audio track and throw away 2/3 of the data?

[Atak_Snajpera]

the same reason why 24bit is converted to 16bit. Human hearing has limitations.

[dansrfe]

Currently, are there any speaker systems/receiver's that actually support 192Khz? Now THAT would be the 9th Wonder of the World.

[Killroy™]

Originally Posted by dansrfe

Currently, are there any speaker systems/receiver's that actually support 192Khz? Now THAT would be the 9th Wonder of the World.

Most modern AVR's support 192Khz with no problems.

[Midzuki]

Originally Posted by dansrfe

Currently, are there any speaker systems/receiver's that actually support 192Khz?

Digital: In the world of today, 192kHz audio should already be something quite trivial.

Analog: In the world of today, speakers and audio amplifiers that support up to 96kHz still are very rare, and (IMHO) somewhat pointless...

Originally Posted by GoodzMastaJ

I'm not sure where the bug is, if it's in eac3to or ffdshow.

Both eac3to and ffdshow use libavcodec for TrueHD decoding,
however I daresay eac3to "is less unreliable than" ffdshow.
I'd recommend converting the original 6.1 audio to a 6.1 .WAV
(assuming disk space is not an issue...)

Evangelion 1.11

I didn't see it, and I didn't like it.

(Yes man, quite often the remakes are quite inferior to the original versions.)

[dansrfe]

You know this might be a n00b question but I still don't understand exactly what a digital speaker is and what connectors are used with it. In a normal home theater setup where, for example, a htpc is outputting digital S/PDIF audio via HDMI (coming from the sound card or mobo sound module built-in to the graphics card), then is transmitted to the receiver all through a single HDMI cable, then then the receiver processes and decodes the audio if bitstreaming. But after that doesn't the receiver send a analog signal to the actual speakers? I'm totally confused about this. What exactly is a digital speaker and what is an analog speaker? What sort of connectors does a digital speaker have?

[Midzuki]

Thanks to some "oddities" of the English language, and also thanks to the marketing people , the word "speakers" became somewhat... undefined. From what I've read so far, it's never very clear when the word loudspeaker refers to the (so-called) "driver" itself, or to the "box" (enclosure) which contains one or more "drivers"... Anyway, usually the expression "digital speakers" applies to a set of "powered" drivers which includes a Digital2Analog converter. ACTUAL digital speakers are an entirely-different beast, and there is an article on Wikipedia about them. HTH.

[Killroy™]

Originally Posted by dansrfe

You know this might be a n00b question but I still don't understand exactly what a digital speaker is and what connectors are used with it. In a normal home theater setup where, for example, a htpc is outputting digital S/PDIF audio via HDMI (coming from the sound card or mobo sound module built-in to the graphics card), then is transmitted to the receiver all through a single HDMI cable, then then the receiver processes and decodes the audio if bitstreaming. But after that doesn't the receiver send a analog signal to the actual speakers? I'm totally confused about this. What exactly is a digital speaker and what is an analog speaker? What sort of connectors does a digital speaker have?

By "digital" we usually mean HDMI connectors. LPCM or bitstream of the audio source. "Analog" usually refers to the old multi-channel RCA-plugs used in older AVR's. Audio from the HTPC or STB comes from the analog output (HTPC) or the multi-channel analog outputs of the STB's. Since only very expensive HTPC analog cards could output 192Khz (properly) it is unsure if you will get a un-molested 192Khz output.

No such issues with the HDMI outputs.

The signal from the AVR to the speakers is always the same (analog) but can be degraded with lower quality wire. But even modest 24GA wires should have no problems with 192Khz at very short throws. Always use 12GA wire for optimum quality.

[dansrfe]

ah. I had this idea that digital speakers were some obscure, advanced, audio professional type stuff.

[Xorp]

Decoding this 6.1 DTS-MA track with eac3to 3.22 and Arcsoft 1.1.0.7 results in distortion: http://stfcc.org/misc/7channels.dtshd

[tebasuna51]

Originally Posted by Xorp

Decoding this 6.1 DTS-MA track with eac3to 3.22 and Arcsoft 1.1.0.7 results in distortion: http://stfcc.org/misc/7channels.dtshd

Read this post and next (the channelmapping is solved with eac3to v3.21)

Test also my sample in the post using Arcsoft 1.1.0.0.

[tebasuna51]

Originally Posted by GoodzMastaJ

I've extracted the TrueHD track from Evangelion 1.11 blu-ray using eac3to. ffdshow reports it as having channels as shown in the screenshot below.

Note the presence of the side channels and absence of an LFE channel.
In actuality, there are no side channels but there is an LFE and a rear center channel which I suppose would make this a 6.1 track (rather than the 7.0 I'm getting from ffdshow). Using the mixer, I found the LFE channel is playing from what ffdshow thinks is Back Left (all that comes out of BackL is rumbling).

I'm not sure where the bug is, if it's in eac3to or ffdshow. Any ideas on how to pinpoint it? I made a similar post on the ffdshow-tryouts forum to stir up some suggestions there as well.

Seems a ffdshow problem because eac3to decode the TrueHD fine to a wav 6.1 with MaskChannel 0x070F (FL FR FC LFE BC SL SR).
Try play the wav file.