need advice importing 88k audio into a 96k session

[zapple]

Howdy,

I am running Pro Tools 9.0.2 with an Mbox Pro 3 on a Macbook Pro with OS 10.6.7.

I'm working on a new CD and have been tracking at a sample rate of 96k / 24 bits. Even though this project will probably only end up on CD, in theory I want to record at the highest resolution available to me at this time.

I am collaborating with a remote musician who has an Mbox2 and is running Pro Tools 8. As I understand, he can only record at 88k using that hardware / software combo.

He could not open my 96k session, so I sent him an mp3 which he is recording over.

When I import his audio (via the import session data dialog) back into my session, his audio is way out on the right of the timeline at like the 25 minute mark.

Why is this?

In the past, I have been easily able to sync up imported audio tracks by visually aligning a unique wave form at the beginning of a track. This first test session I imported seems to "drift" a little, so that I'm getting some flanging between the reference mp3 he recorded over (which I also imported back to my session) and the audio in my session.

Is this because of all the conversions I'm doing (bouncing to mp3, importing that mp3 to PT8 at 88k, reimporting to PT9 at 96k)?

Is there a smarter, better way to deal with this?

Any thoughts / problems with recording audio at these two different sample rates (88k and 96k)?


Thanks!

[Top Jimmy]

The Mbox2 cannot do anything over 48KHz.

Mp3 encoding/decoding is not sample accurate so you can expect phase shifting.

I would send him a 48K wav file rather than an mp3. Have him record his parts in a 48K session and send consolidated wav files back to you starting from the same place in the session as the file you sent him. All you would need to do import with sample rate conversion and line up his parts with what you sent him.

If you are trying to spot his files and they go to the 25 minute mark, that just means that is where he did his work at in his session.

[zappity]

Thanks for your note Top Jimmy! Good to know about the mp3 encoding.

Any thoughts about combining/importing audio recorded at different sample rates into a single Pro Tools session? Can I trust the PT "import audio" dialog to do all of my crunching cleanly?

Also, not knowing much about the technicalities of dithering down to 44100 at 16 bits, are people recording at 96k / 24 bits for projects that ultimately end up on CD?


Thanks again!

[Top Jimmy]

Originally Posted by zappity

Any thoughts about combining/importing audio recorded at different sample rates into a single Pro Tools session? Can I trust the PT "import audio" dialog to do all of my crunching cleanly?

Yes. If you want the best PT can do, you can set the SRC quality to "Tweakhead". One thing to beware is that there is currently a known bug when trying to import audio of varying sample rates all at the same time via the Import>Audio window. Currently, mixing the sample rates on import will result is some of them importing correctly and some not. The workaround is to only import files of one sample rate at a time.

Also, not knowing much about the technicalities of dithering down to 44100 at 16 bits, are people recording at 96k / 24 bits for projects that ultimately end up on CD?

Maybe yes, maybe no. While I have recorded some bands/music at 96KHz, they were worth doing it. Also, this was on an HD setup that could handle the double rate throughput without hiccup. Even with a modern CPU, if you want a decent track count with a Digi firewire interface, you had best stay at standard def rates.

It used to be that sample-rate converting was atrocious so the consensus was to avoid it at all costs. Today the SRC algorithms are vastly improved and we have a fair number of plugins that use SRC to process at higher rates. I have yet to hear someone complain about the SRC used by any of these plugins.

Personally, even though SRC is much better than it used to be, I still fall back on the idea that when the final product will be at 44.1KHz, that's the rate at which I record.

[jntracks]

i think computers are good enough at math to handle all the sample conversion you need. 44.1k can represent up to 22.05 khz, that's higher than you can hear but the advantage of keeping your session at the highest resolution is the processing will happen at higher resolution. so the inaudibly short bits of transient audio that need to overlap and process and sum and phase with each other can do so at high resolution.

but you're really splitting hairs.

and of course dither down to 16 bits should be the very final process that happens to the audio before the laser hits the lacquer. maintain the maximum dynamic range of each of those samples no matter what resolution.

[zapple]

Thanks fellas. Very helpful info.

Thanks for the reminder about that bug Top Jimmy...I did read about that.

Z!

[Park Seward]

Originally Posted by Top Jimmy

Personally, even though SRC is much better than it used to be, I still fall back on the idea that when the final product will be at 44.1KHz, that's the rate at which I record.

Agreed. There is no improvement if you record at 96 and downsample to 44.1 vs. recording directly to 44.1