16 Bit Session with 24 Bit Input

[rohlfie]

Greetings: Stumped tech support with this one.
I'm now using an RME Fireface UFX+ that only operates at 24 bit. But it could be any interface.
The question I have is what does a 16 bit Pro Tools session do with 24 bit audio when it receives it? I'm just now recalling Pro Tools using a 32 bit floating point bit rate. But since this audio is going to be dithered down to 8 bit eventually, I wanted to get clear on what exactly is going on at each stage in order to avoid unnecessary quantization errors, or the like.
If the audio is truncated on the fly, does it matter if the 16 bit audio is dithered to 8 bit?
Thanks!
I'm running Pro Tools 2018.12.0 on a 2017 iMac (4.2 GHz Intel Core i7 32 GB 2400 MHz DDR4) on macOS Mojave 10.14.4

[64GTOBOY]

IIRC there is a lot of dithering going on under the hood during processing but at 32 bit float it is well below/above the audible noise freq so i wouldn't think that once it's in a session it wouldn't really matter how it started if that means anything

[daeron80]

It does matter. Quantization rounding errors from truncation produce inharmonic partials that are well above the noise floor. They will cause problems with any further processing, and could easily still be an issue after dithering to 8-bit. I wouldn't chance it. Track to 24-bit.

[albee1952]

I'm going to turn this around and ask why you want to record to a 16 bit session? The final product will never be "better" than the source, so I would not start working at the lowest quality format(thank God, nobody has asked to make MP3 a session format). I understand all the arguments about "why bother when the final is going to be 16 bit/44.1 on a CD?", but we still have many users that record at 88.2, 96K and higher...My personal preference is to work at 24 bit/48K, regardless of the final delivery format and here's why(in no particular order):
#1-even my old ears hear the improvement
#2-most video requires 48K audio
#3-many plugins work better, the higher the sample rate
#4-I would go to 96K, but my computer doesn't handle the workload well
#5-recording at 24 bit allows you to record at lower levels for increased headroom.

This might open a can of worms, so I'll just say this is what makes sense for me

[rohlfie]

Thanks for the replies! The reason why I need to record at 16 bit is because the proprietary software used for IVR records at 16 bit, as the target file will be compressed down to 8 bit. I would never record at 16 bit these days unless I had to.
Source Elements Source Nexus plugin makes it easy to forward, uncompressed, audio from a Pro Tools session into the proprietary software in this case. And I suppose I could do a workaround, by using my Sound Devices MixPre 6 and a mic splitter into my Fireface for monitoring/playback.
But I still want to know exactly what happens to the 24 bit signal as it's received by a 16 bit Pro Tools session. For now, I have to assume it is truncated. And quantization errors become much more noticeable at lower bit depths. Although, it may not be an issue after the truncated 16 bit files are dithered to 8 bit.