Without getting mathematical, you've already proved it to yourself (and I tend to agree). And you might notice that if you can track down some polls (not just forum polls, but actual polls) of SR usage, you'll find that somewhere between 70 & 80% of full-time professionals record at the target rate (44.1kHz for audio projects).

One thing is for sure -- I'd most definitely NOT record at 48kHz if I knew I was going to SRC it down to 44.1kHz later... It's not like one was made to be "better" -- One is for audio projects, one is for video. Adding 2kHz to the top (of which, 99% of the gear out there can't even come close to in the first place) seems silly when you think about it, no?

You want to beef up your input chain at *any* sample rate? Upgrade your converters.

Cheers for the reply.

The reason I asked was I booked time in a studio and asked what format the files would be recorded in so I can mix at home (Yeah, his face sagged). The owner replied 48khz 24bit. I said I was making CD's not Video and asked if I could record to start with in 44.1... He chomped me... Politely, but still...

He has Apogee AD 16X and DA 16X converters. Which cost slightly more than mine. ;P

If you are going to be doing any pitch processing or time stretching, the higher sample rate may help. However you probably aren't going to notice an improvement unless you get up around 80kHz+

For this reason I use the target media rate.

Thanks for the interesting topic to think about. I'm going to go figure out how to do some testing.

He chomped me... Politely, but still...

I suppose it's not the hill you want to die on - So you might just go with it. If nothing else, he can't use the "Well, you wanted it at a different sample rate and that's why (etc., etc., etc.) happened."

But he should know better also...

And agreed - Some processing may benefit from upsampling (or recording at higher rates) - But not "way up there at 48kHz."

I went into Uni today, and about time too, apparently it's week six ;P

Anyways, I found my old signal processing lecturer.

We spoke for some time about this topic.
It seems there is no direct maths to explain what we are talking about but he gave an explanation which made some sense.

Basically he said that using Shannons theorem would be what you are working with and it's a matter of how far you wish to extrapolate and interpolate the data.

He said off the top of his head up-sampling to something like 480khz then coming back down. I think he was looking for a common denominator to go up to then back... Something both 44.1 and 48 go into somewhat evenly. Who has space to upsample to 480khz+? :P

Basically it was a long winded conversation which ended with him saying there is no "maths" to explain why this happens, only maths to approximate how to do it.

Also we discussed that it may be a case of the shaping the filters give to the sound, as in when recording directly to 44.1 the roll-offs and other sonic characteristics of the way the converters handle the sampling could be different when down sampling.
This is because of the loss of available frequencies at the lower sampling rate.

Yeah, it's a drift.

If you have a good SRC, I think they are very good ones now, I don't really see it as an issue anymore. Bad SRC sounds well... bad.