Sample rates

[Jonathan Morrow]

Hi All, following my issue last week with having to convert 48kHz sampled files to 44.1kHz files for podcasting I have come up with a related issue; CD music is always 44.1kHz, so how is music put into a film (if the film is recorded at 48kHz) without there being a noticeable loss of quality in the theatre? If there is less of a problem going from 44.1 to 48 than there is in the other direction, I am wondering if I shouldn't make 44.1kHz my default sampling rate for all recordings. I ask this since I am recording a local musician for inclusion on the sound track of a documentary and I don't want to have to repeat any conversions.

Cheers, Jon

[Marco Leavitt]

There may be a measurable loss in quality in the conversion, but you can't really hear the difference, at least I can't.

[Seth Bloombaum]

Originally Posted by Jonathan Morrow

If there is less of a problem going from 44.1 to 48 than there is in the other direction, I am wondering if I shouldn't make 44.1kHz my default sampling rate for all recordings. I ask this since I am recording a local musician for inclusion on the sound track...

With Marco, (and probably most soundies), I'd recommend recording as much in 48KHz as possible.

First, as Marco points out, in practice, artifacts from such sample rate conversions are not perceptible in the final product. Of course this could cut both ways.

If there is less of a problem going from 44.1 to 48 than there is in the other direction...

There isn't. If we want to go to theory, there's slightly more data/accuracy in the 48KHz recording. Also on theory, if one wished to maximize potential quality, even if imperceivable, one would record on two devices of different sample rates, split the mics, etc.

Third, the consequences of small time base errors from resampling are potentially huge in video/film, and non-existant in music-only playback. For the music market, who cares if tracks are a second longer than the original performance? Some would care if it is a couple cents flat or sharp, but very, very few.

The consequences of that sort of stuff can be both difficult to sort out, and very time-consuming in any sort of picture sync work. Even if you're not shooting picture on your musician, a 44.1 track is going to be handled radically differently by different NLE systems. One NLE might give you no trouble, move it to another facility, (even with the same NLE, but a different sound card!), and your music now ends in a different place relative to picture, or, on a different NLE, might not play at all.

If you're totally an island and don't interact with the world, and your NLE supports multiple sample rates easily, then I guess this wouldn't be as important, but these standards exist for a reason - if all contributors conform to 48KHz that means less grunt-work for everyone, and more time to spend on the creative work. And, you're less likely to get caught out recording some sync sound at 44.1!

[Jonathan Morrow]

Originally Posted by Seth Bloombaum

...the consequences of small time base errors from resampling are potentially huge in video/film, and non-existant in music-only playback. For the music market, who cares if tracks are a second longer than the original performance? Some would care if it is a couple cents flat or sharp, but very, very few.

I hadn't thought of that, nice point.

Originally Posted by Seth Bloombaum

Even if you're not shooting picture on your musician, a 44.1 track is going to be handled radically differently by different NLE systems. One NLE might give you no trouble, move it to another facility, (even with the same NLE, but a different sound card!), and your music now ends in a different place relative to picture, or, on a different NLE, might not play at all.

This maybe what I was seeing when I originally tried to convert my previously recorded 48kHz to 44.1kHz using Tracktion. It didn't do it any good at all. I have to admit though, that I never really checked out the original recording in any great depth, I simpy assumed that the warbles I was hearing on "aah" sounds were artifacts of the conversion process. It is also possible that they were on the original recording.

Thanks for the input, I will go with 48kHz tomorrow, I doubt that a soundtrack will end up on CD anyway, I certainly don't intend to double record. Cheers, Jon