I record at 16 bit if I was burning to CD. Or, I record at the bit rate my final destination would be. Less conversions always the better.  Its as simple as that imo.

However, there are excellent bit rate conversion processes like you may use that do the math, and dither it properly . Fabfilter is one that use.

I prefer to use two DAWs to mix down. Example: I track at 24/96 and mix/master down to DAW 2 at whatever sample rate I want. This avoids converting issue and always allows me to hear my "mix down" in real time at the sample rate I choose for the format I want the master to be.

You may find some useful help on conversions here: https://recording.o…

I think at the end of the day, if it sounds good to you, then you are likely on the right track. If you are sending your music to a mastering engineer, they would advise you of what they want.

Hope this helps.

In your case I would mix/master in 24bit, then when it’s time to export/bounce I would do versions in all the formats you expect to use. Like for example 16/44.1 for cd, a high bit rate mp3, and one at the bit and sample rate you used (ie no conversion) because now devices can play back high res stuff, and if you upload the high res stuff to YouTube ect, they will convert it.

the other option would be to bounce it unconverted, then use your standalone program to convert to the various formats.

at this point it’s possible to deliver the final version to the masses without any sample or bit rate conversions, as we are no longer tied to the physical playback medium. This is my reasoning for having a master at full resolution, even if other formats are needed for various reasons like streaming ect. 

i don’t believe higher is always better, rather that we can go thru the entire process at the same resolution, preserve that, and convert to other formats as needed.

Thanks for the advice!  CD Baby wants it as 16-bit 44.1 so I'll have to convert to that as a separate step.  I was trying to combine it in the same step as the mastering, but will bounce and apply instead.

My untrained ears don't hear that much of a difference between 24 and 16 anyway... On the next project I may just track in 16.

If you track 16bit you have to be a lot more careful of your recording levels. You have to record at a much higher level than 24bit. This can run more risk of clipping.

When converting from 24 to 16 bits the conversion throws out the “least significant bit” these are very low level things like reverb trails for instance. 

Bos posted this a few years back that is somewhat relevant

https://recording.o…

We also need to be aware that as we age, what we don't hear is actually heard by others. As we age, we tend to mix things brighter, suck out the mids more and generally just suck at mixing as good as we used to.

damn... ;)

If you record at 24bit, you should mix and master at the same resolution. 

Then when time comes to export the final product, then you can go down to 16bit/44khz for CD or stay at 24bit\48khz for videos. 

I'm not sure about the UV22HR thing.. I never needed a plugin to change the resolution of a project.. 

Does it let you hear what changing the resolutions will sound like, similar to what Ozone does ? 

When I bypass that plugin, and toggle back and forth, it maybe sounds different?  Or maybe I'm just thinking it sounds different.

There’s many reasons why it could sound different.

you can do an easy test. Just bounce the track down with identical settings, except that pluggin, then import the two files into a new session. Now flip the phase on one track. It it’s  dead silent, then there is no difference in the files. If you hear something, that is the difference between the two files.

Yes Null tests is the best way to know if it sounds different. 

I'm guessing that it might add random noise like dithering does.. but what is the plugin we are talking about.? Something you already use or is it hypothetical ?