time comp 5.1 Stems for conv 24 to 25fps

[yazoo]

Hi all!
I have done a 5.1 Mix that is recorded at 24b48khz 24fps.

For our first screening we had to timecompress it to show it in 25fps PAL.

It seems like there is some phase issue between the channels.

Is there a better alternative to Digidesigns timecomp for tasks like this?

[quadraphonics]

There are several. You can create a new session and import the audio doing a sample rate conversion, which I have found to be the best.

You could use a program like Pitch 'n Time Pro to do this as well.

Randall

Originally Posted by yazoo

Hi all!
I have done a 5.1 Mix that is recorded at 24b48khz 24fps.

For our first screening we had to timecompress it to show it in 25fps PAL.

It seems like there is some phase issue between the channels.

Is there a better alternative to Digidesigns timecomp for tasks like this?

[stangroom]

Originally Posted by quadraphonics

You could use a program like Pitch 'n Time Pro to do this as well.

Pitch and Time is WAY better than Time Shift. If that's not n option play with the different monophonic/polyphonic choices. Poly should be the set though.

Good luck!

[cjguitar]

There is also X-Form from Avid, which is supposed to be better than TCE, but I tend to go the sample rate conversion route, although that does noticeably change the pitch.

[danijel]

Originally Posted by yazoo

It seems like there is some phase issue between the channels.

Whatever you use, make sure that the program is fit for polyphonic time-stretch. Prosoniq Timefactory has the mode where you import 6 monophonic files (for 5.1) and he keeps them aligned after the process. Also, Nuendo 3 works great (MPEX2 algorithm, licensed by Prosoniq), if you set the mode to polyphonic-best, and stretch it as one 5.1 clip/region.
In the thread title, you mention "stems". I don't know of any software that can stretch several stems and keep them phase-aligned, but that is usually not an issue, because they carry non-correlated material.....

[Steven1145]

After lots of testing, Waves Soundshifter Offline works very well. But only from v6 onwards.

[TiagoJSilva]

If you're doing a pure varispeed 24->25fps, it shouldn't make a difference if you use AudioSuite plug-ins in Mono mode, Multi mode, or importing the files to a new session with SRC. This apart from Digi's TC/E, which I would avoid for anything apart from the TC/E Trim tool in DX tracks. As long as the precision of the plug-in is enough, the same number of samples should be converted for every channel.

If you're pitch-shifting to compensate for the varispeed, then you'll need an AudioSuite plug-in with Multichannel mode (I recommend PnT or X-Form). Or an external processor, like the Genesys Harmo or the Dolby 585 Time Scaling Processor. All channels of the stem will have to be run at the same time, other wise you'll get phasing when folding to mono, LoRo or LtRt.

Since no "harmonizer" is perfect, doing the harmonizing stem by stem generally leads to the best results, since you can use the appropriate settings for each kind of sound. Some algorithms are very good for voice but not that much with music, and vice-versa.

[petessound@hotm]

plus 1 for import session data.

set the source sample rate to convert from film to PAL. no need to make it any more complicated than it needs to be with plugins

[RScott]

Also, if you're trying to do pitch corrected time stretching and you're finding that it is making too many artifacts, but don't want to do a Varispeed sample rate conversion because of too big a pitch shift, you can try this:

Find out (with a sample test piece) what the greatest amount of pitch shift you could accept is before it sounds too pitched. Test with music and dialogue. Then use a Varispeed or other sample-rate-variable stretcher to stretch this amount and make up the remainder with a pitch-corrected stretching method.

[TiagoJSilva]

Originally Posted by petessound@hotm

plus 1 for import session data.

set the source sample rate to convert from film to PAL. no need to make it any more complicated than it needs to be with plugins

Well, what about Gérard Depardieu speaking like an adolescent, then?

To pitch-shift or not to pitch-shift, that is the question...

[neilwilkes]

Originally Posted by yazoo

Hi all!
I have done a 5.1 Mix that is recorded at 24b48khz 24fps.

For our first screening we had to timecompress it to show it in 25fps PAL.

It seems like there is some phase issue between the channels.

Is there a better alternative to Digidesigns timecomp for tasks like this?

Can you timestretch the 5.1 interleaved file - timestretching 6 separate files will cause phase issues, but the interleave will be fine & phase coherent.

[Kuba_Pietrzak]

As for now (as I know), there is no a perfect tool, which would give you almost unheard results with such time compression or expansion (+- 4%).

In the old days, I used "traditional" conversion (with pitch change) and compensated it with TC M6000 with its harmonizer program. This worked rather good, however there were artifacts in the low end.

Digi's TCEX has lots of time issues, but works pretty well with unprocessed mono dialogs, but there is a need to do it region after region and check sync. Serato Pitch'n'Time has almost perfect timing, but the sound of dialogs after such processing is, I would say, far from perfect.

The best way is to use multiple methods, especially if separate stems are available.

regards,
Kuba

[quadraphonics]

Correct me if I am wrong, but the pitch difference we are talking about here is a musical 1/2 step, right?
Unless your show is a musical, I am not 100% sure I would worry about it. Every single pitch correction software will induce some level of artefacting into the process. That is why on DVD releases, many of the big studios stopped pitch correcting the elements which were going to PAL. The 5.1's sounded phasey and the folded down poorly. Even with Pitch n Time and the TC Box.

Randall

Originally Posted by Kuba_Pietrzak

As for now (as I know), there is no a perfect tool, which would give you almost unheard results with such time compression or expansion (+- 4%).

In the old days, I used "traditional" conversion (with pitch change) and compensated it with TC M6000 with its harmonizer program. This worked rather good, however there were artifacts in the low end.

Digi's TCEX has lots of time issues, but works pretty well with unprocessed mono dialogs, but there is a need to do it region after region and check sync. Serato Pitch'n'Time has almost perfect timing, but the sound of dialogs after such processing is, I would say, far from perfect.

The best way is to use multiple methods, especially if separate stems are available.

regards,
Kuba

[Kuba_Pietrzak]

Yes, the difference is close to half step (half tone)..., so the main problem is, that sometimes such a pitch change, changes actors personality which is even more important, if this is well known actor and his "natural" voice is also well known. The second thing is, that there are composers, who may not let you do such a change, because their music "changes so dramatically", which is even more problematic, when they hear absolute pitch. From sound quality point of view, yes the safest and the best solution is the "classic" slow down (or speed up) with pitch change.

There is a way to achieve really good result and preserve original pitch:
- reedit the music to match new "speed" (perhaps Serato)
- take the source dialog session and line by line change using Digi's TCEX (which speeds unprocessed mono voices really well) and MIX DIALOGS AGAIN
- there should not be a problem, if BGs and Sfx (unless they contain tonal content) are pitch changed
- perhaps correct tonal content from Sfx and sound design, if there is a conflict with music

This takes a lot of time, but give much better results.

regards,
Kuba