Preserving 'Dolby Pro Logic' information when transcoding.

[20/20]

Hi guys.

If a source AC3-track (Dolby 20) has Dolby Pro-Logic information (Dolby Surround)
How can I make sure this Dolby Pro Logic information is preserved after transcoding
to another format like MP3?

*Do I have to use "Stereo mode" with MP3?
*Is there another format you reccomend to do this, and wich settings do I use?

Thanks

[mh2360]

Transcoding a Dolby 2.0 AC3 surround track to any other stereo format should preserve the surround sound, just be sure to choose an adequate bitrate. 192K is what Dolby 2.0 tracks are usually encoded at, so you should use 192k as your bitrate for MP3 for example.

[20/20]

If I had to encode this Dolby 2.0 to 192K MP3, Wouldn't that mean
I'd also could keep the original AC-3 track instead?
Theres no sense in compressing to MP3 at the same 192Kbitrate then, would it?

[Cornucopia]

Automated Dolby ProLogic decoding in an amplifier is actuated by a FLAG in the AC3 metadata/header that accompanies the main data stream. It doesn't necessarily have to be encoded in Dolby Surround to be decoded that way (THOUGH IT SHOULD). For things to be decoded correctly, the flag should be on and the 2 channels should have been pre-matrixed with a Dolby Surround processor.

Now, decoding is NEVER completely transparent, even with uncompressed originals, interim files, and final output. But, ANY kind of lossy compression will further hamper decoding transparency. It has to "guess" about whether is supposed to go to the L or the C or the R or the S channel(s), and the more "muddy" the signal, the harder it is to make that guess.

So when going to another format, you're further compromising the capability of the decoder to do its job--especially if that format is lossily-compressed. Further still if it's a lower bitrate.

Obviously, with MP3 you couldn't maintain 2channel info in the [MONO] setting, but the difference in the other 3 choices aren't as clear-cut.
1. Joint-Stereo (and/or sometimes "Intensity Stereo")
2. Stereo
3. Dual-channel

It is my understanding that an encoder could be set up to encode M+S (Mid + Side) as opposed to L+R, and in #1 that is what is always done, while #2 could do that, but isn't usually set up that way (more likely it just tracks/matches/pairs the bitrate needs evenly to both channels), while #3 is meant for 2 separate single programs, but can be used for stereo (it tracks bitrate needs independently for each channel). Also, with "Intensity Stereo", the M+S actually blends the higher frequencies (for bitrate efficiency). This is quite destructive on surround/phase cues.
Therefore, it makes the most sense to set an MP3 encoder to either [STEREO] or [DUAL-CHANNEL], with program material type deciding which is better.

Barring lossy-compression artifacts, etc., the problem with MP3 is that there is no "surround" [FLAG] that can be put in the MP3 header/metadata. So decoding can't be Automated.

My suggestion is to leave it as the original AC3 if at all possible.

Scott

[20/20]

That sounds very obvious

hydrogenaudio alway stress to use newer lame-builds and that thee -mj is the only way to go.
Idunno.

The lame -ms encodes stereo channels but leaves for the bitrate to fluctuate bewteen l-r if one channel needs more bits.
I think dual channel would be the way to go if I really would like
to transcode to a different format.

Anyway, as fo now I'm more concerned with avisynth and LaserDisc
restauration (and noise, omg what a headache):P

[3dsnar]

All what Cornucopia wrote is true.

Maybe to extend the issue of lossy compression influence on the DPLII information preservation,
it should be mentioned that MP3 (stereo or dual channel) performs fairly well, while AAC or OggVorbis are not recommended.

This is because MP3 preserves phase of the most energy carrying components of the signal spectrum (for all frequencies), while AAC or Ogg use some "phase guessing" techniques (e.g. SBR), which has terrible influence on the DPLII downmix.

Cheers,
3d