Channel swapping?

[`Orum]

I'm shrinking down 6 channel (technically 5.1 I think, as one channel appears to be a LFE channel) TrueHD audio taken from a Blu-ray disk. The audio was initially converted using eac3to from the original TrueHD to FLAC, and then encoded in Ogg Vorbis before being muxed with the video. However, when I play the movie back on my PC with headphones, voice and other mid/high-frequency sound like they are coming predominantly from the right, while low-frequency seems predominantly from the left.

This leads me to wonder if the LFE (or center channel?) and left channel were somehow swapped in the conversion process--or there's some downmixing issues. If I enable ffdshow's mixer, and set it to "2/0/0 - stereo", the problem remains. If I set it to "headphone virtual spatialization", the balance sounds about right, but the audio is much fainter, and even at maximum volume it doesn't sound loud enough. If I disable the mixer and simply use swap channels (swapping "center" and "left"), it sounds much closer to the downmixed 2 channel AC3 track on the Blu-ray disk.

My questions are:
Does Ogg Vorbis have some way of tagging which channel is which, or is it simply by the order of the channels?

What software should I use to listen to each channel individually to decide if it's in the right place, or where it should be? Audacity will open it, but simply assigns the first two channels it sees to left/right and the rest as just individual mono channels, leaving me to guess their placement and if it even has left/right correct.

Is there any way to swap Vorbis channels around without reencoding?

If not, what's the simplest way to reorder them in the FLAC file?

[sehgal.v7]

why don't u let eac3to do the downmixing?? 6channel to stereo? That too with Dolby Pro Logic II matrix..

try) eac3to audio.thd audio.flac -down2

for re-ordering in flac try) eac3to (input file) (output file) (channel order lik- 0,1,2,3,4,5)

[`Orum]

why don't u let eac3to do the downmixing?? 6channel to stereo? That too with Dolby Pro Logic II matrix..
...because I'd like to keep the 6 channel audio. I just think the tracks were incorrectly ordered/tagged somewhere along my conversion process

for re-ordering in flac try) eac3to (input file) (output file) (channel order lik- 0,1,2,3,4,5)
Well, I'd ideally not like to go back to the TrueHD source (as I'm sure it will take longer than reordering the FLAC file I already made). The bigger question is, how do I know what the correct order is supposed to be, and if it's right and this is just a playback issue vs. a transcoding issue?

[tebasuna51]

You can use Foobar2000, BeHappy or eac3to to downmix and encode to ogg.

With eac3to you can use:
eac3to multichannel.flac stdout.wav -down2 -normalize | oggenc2 -o stereo.ogg -

[`Orum]

Since this seems to be confusing people, let me make this very clear. I don't want to downmix the source audio. I'm noticing playback problems when it is downmixed during playback with ffdshow. As such, I think the channels are in the wrong order. Refer to my first post if there's anything else unclear.

[sehgal.v7]

I din told you to get back to TrueHD track.. You can use it with flac too..

>eac3to input.flac output.flac (new channel order lik -0,1,2,3,4,5)

PS- Refer to eac3to first post (-0,1,2,3,4,5 remap the channels to the specified order) & do some homework.

[tebasuna51]

Since this seems to be confusing people, let me make this very clear. I don't want to downmix the source audio. I'm noticing playback problems when it is downmixed during playback with ffdshow. As such, I think the channels are in the wrong order. Refer to my first post if there's anything else unclear.

Sorry, I don't understand first time.
Then you if want encode a 6 channel ogg file use BeHappy (the data is sended raw to avoid problems).

If you want encode using oggenc2 (www.rarewares.org) you must know the bug/problem pending of solution:

If the data is send to oggenc2 with WAVE_FORMAT_EXTENSIBLE header with a MaskChannel 0x3F the ogg output is ok.

If the data is send to oggenc2 with standard PCM header or with a MaskChannel distinct of 0x3F the ogg output is wrong.

Eac3to send the data with WAVE_FORMAT_EXTENSIBLE header with a new style MaskChannel of 0x60F not accepted by oggenc2, then you need a remap like this until a new oggenc2 version:

eac3to multichannel.flac stdout.wav -0,2,1,4,5,3 | oggenc2 -o multichannel.ogg -

[`Orum]

I din told you to get back to TrueHD track.. You can use it with flac too..Yes, but I'd rather not go back to the TrueHD source track if I can avoid it. It's excruciatingly slow, especially because it has to do two passes, and then if I get the track order wrong in my guess, I have to do the conversion all over again instead of just a quicker direct copy from FLAC to FLAC, which I'm sure has to be possible with some audio editing software.
>eac3to input.flac output.flac (new channel order lik -0,1,2,3,4,5)

PS- Refer to eac3to first post (-0,1,2,3,4,5 remap the channels to the specified order) & do some homework.Yes, I can do that but I don't know what the correct order is. After looking around for quite some time, I'm not sure how the various formats tag which channel in the file relates to which speaker, or if it's simply the order they appear in the file. Either way, it looks like some of this information was either lost in transcoding or expected to be in a different order when it was transcoded to Ogg Vorbis. If you know of a site that lists the layout of multichannel Ogg Vorbis (and TrueHD too, if available), I'd be more than happy to read it.

[sehgal.v7]

Thn try wht br0 tebasuna suggested..

[tebasuna51]

You don't need know the internal order in compressed formats

If you ask, you need trust in the answer.

The standard order for uncompressed format (PCM-WAV files) is FrontLeft-FrontRight-FrontCenter-LowFrequencyEffects-BackLeft-BackRight-...

A encoder must accept this uncompressed order and the user don't need know the internal order.
A decoder must supply this uncompressed order not mather what is the internal order.

This order was established in the past century (http://dream.cs.bath.ac.uk/researchdev/wave-ex/mulchaud.rtf) but some encoders (oggenc2) still don't know this.

[`Orum]

The standard order for uncompressed format (PCM-WAV files) is FrontLeft-FrontRight-FrontCenter-LowFrequencyEffects-BackLeft-BackRight-...

A encoder must accept this uncompressed order and the user don't need know the internal order.
A decoder must supply this uncompressed order not mather what is the internal order.
It's interesting that it (the ogg vorbis encoder) should take this order, but doesn't appear to.
This order was established in the past century (http://dream.cs.bath.ac.uk/researchdev/wave-ex/mulchaud.rtf) but some encoders (oggenc2) still don't know this.
That's the document I've been looking for! Unfortunately it's only half the puzzle as to why there is a problem. The other half I found here (http://lists.xiph.org/pipermail/vorbis-dev/2005-November/018300.html). It's strange, because this bug is reported back in Nov 2005 and surely would have been fixed by now. However, they mention "...works on encode and decode to file and reads/writes WAVEFORMATEXTENSIBLE headers as appropriate." I assume FLAC lacks such headers, and this the encoder assumes that the channels are already in the internal order that vorbis uses. Strange behavior, but the workaround appears to work, thankfully.

Edit: Your earlier post makes more sense now--still, new versions of the encoder have been released, so it seems odd that this bug would still be present.