Using Mid-Side decoder (recomendasions)

[funkytwig]

Read a bit about this and for stereo mixes, it seems the Mid-Side decoder is recommended if you can try listening to the DCP but I have a couple of questions.

First can work out exactly how to do this, is there an easy-to-follow how-to somewhere?

Secondly, I am curious exactly what this should improve, how would L/R only sound bed /wrong without it?

Also, how might it sound worse, what may it do to the sound?

I've done quite a few simple L/R DCPs and never had any complaints but I'm keen to investigate this.

Lastly, can this be checked with a more 5.1 system, I do have one but guessing the answer is kind of but it should be tested in a cinema.

Regards,
Ben

[Carsten]

Compared to upmixer A and B, which also create surround signals, the Mid/Side Decoder ist fairly safe to use. You do not necessarily need to check the result in a cinema. You can also use DCP-o-matic player, and play with the playback matrix to listen to the 3 channels created to get a good idea.

Now, sometimes we have to deal with bad audio, and this MAY create problems with the decoder. For instance, I rarely come across stereo audio from dubious sources where audio is phase changed between left and right. In that case, the Mid/Side decoder will extinguish audio or dialog. But you can easily find that out by listening to the resulting DCP audio on your ordinary stereo system and switching playback channels in the audio matrix.

To use it, there is not much to do - if you have stereo audio, enable Mid/Side decoder under DCP->Audio, and choose a DCP audio channel number of at least 4 or 6.

- Carsten

[Ryan Gallagher]

I've recently started playing with MSD 3.0 versions too.

Noticed this quirk, not sure if it is a bug or I am doing something wrong:

If I create a 2.0 DCP initially, then without starting a new project, set it to MSD and also export a 3.0 version...
(On windows 10, 2.16.88 DOM)

Then if I use the player to verify both DCPs. The 3.0 one will verify without errors, but the 2.0 one will complain about an asset ID mismatch in the ASSETMAP. Before I caught this by verifying locally I would run into failed ingests with the 2.0 version.

If I delete the DCP output folder for the 2.0 one and re-export it after switching back to "none" in the audio processing. This second export of the 2.0 DCP verifies without error.

I did not snag a screenshot or log during verify of the failed one, but could probably reproduce it by repeating my steps with another file.

EDIT to add this is exporting using the Batch processor, not waiting for the first to complete before putting the 2nd in the cue.

[IoannisSyrogiannis]

That seems to me like the assetmap file being overwritten when the second version is created.
I don't think it has to do with 2.0 or 3.0 audio.
It could be a bug, but not a bug relevant to the specific thread.