Is a high sample rate (88.2 or 96) worth it?

[Lugh98]

If I am just going to end up converting a 88.2 to 44.1 when I mix down, is the end result going to sound any better if I had just imported at 44.1 to begin with?

Are there better conversion formats that are better than others?

thanks...

[JohnnyG]

I guess it's all debatable, but I definately have worked in Pro Studios that use 88.2 as their default recording sample rate because the down sampling division is an even 1/2 to get to 44.1 and should be a "cleaner" and easier conversion, rather than using 96K. Also you can definately hear a big difference in 88.2 compared to 44.1, but the difference between 88.2 and 96K isn't so obvious.

Though, I don't know the exact science or "magic" of how or what parts of the higher sample rate get down sampled into the final mix, maybe someone else kind enlighten us.

In the TDM world using higher sample rates would generally make more sense but with LE I guess it all just depends if you have the processing power available and the need for the higher sample rate.

[OddsAre]

Don't bother... especially if you're not using really high end mics/pre's/conversion... its not worth losing the extra power you could be much better off using for additional plugins. I made the mistake of recording my friend at 96k just acoustic and 2 vocals... my iMac G5 had a hard time running just a few plugins... where at 48 or 44.1k I can mix a full band. Plus you'll pay a fortune in hard drive space. Until you step up to a TDM system in a controlled environment, with adequate equipment, keep your sampling rate down... and bit rate up! 24bit!

[comprodman]

At the end of the day it's all down to how much space you want to use against the Power of the Mac and the final clarity. As you cannot hear over 20khz or below 20hz the actual sample rate over 44.1 makes little audio difference, though having twice the number of samples (0's and 1's) does actually make a huge difference to the quality of the audio, in the same way making a brick wall with smaller bricks is much stronger than making a wall with large bricks. The more samples you have the harder the processor will have to work, so at the end of the day it is all down to payoff, a few less samples and a few more tracks with a couple of extra plug-ins? I tend to evaluate a job on a needs basis. If its a paying job and envolves music, I work at 24bit 48K if its just a Voice narration and a few sound effects I work at 44.1 16 bit, I do work for most UK radio stations and even for TV recording Voice Over's and I have never had a complaint about quality.

[nikki-k]

Hi!
Not that it should matter, but I usually work at 24/44.1k...even when I was using my HD4 Accel rig. 24/96k seemed to have a little more sheen, or something, to it. So, I now have some at one rate, and some at another. It really is hard to decide!

To make it easier, and harder at the same time, I would HIGHLY suggest reading both of these:
Dan Lavry paper
Not too rough to understand. He spells it out nicely, and in layman's terms enough so that those with out degrees can grasp things. I happen to agree with the ~60khz theory, which may be why I am torn between which to record at...or, I might just be a little crazy...

Next, if you really want to get a better grasp on digiyal audio, I HIGHLY suggest getting a copy of this:
Digital Audio Explained - by Nika Aldrich
IMO, this is among the top 3 must read books...maybe even the first one to be read! Nika does something interesting: he explains things, but about 99.9% of the time he keeps from pushing his own opinions. This book is fact, and explains theory and proof of various aspects of digital audio, and working with it. If you have good gear, I would not be suprised to see you walk away from this read being happy to work at 24/44.1k. After reading Dan Lavry's paper, you probably wouldn't see a reason for ~192k, but 88.2 or 96k might have a slight appeal. (BTW- yes, that is the Dan Lavry that makes Lavry converters).

Read both of those. Nika's is a HUGE meal to try and eat, and then try to digest. It isn't deep, 4th year physics; but it isn't simple layman either. Might take several reads to fully grasp everything in there, or one, well focused read might do it! But combined with Dan's paper, you should have enough to then sit down and figure out which you want to use. And then simply try both, and use your ears to decide. Your converters might simply sound better to you, and those you are working with, at 24/96k/88.2k...or, 24/44.1 might just do it.

Oh- btw- dont worry about the downsampling thing, or dithering; if you are doing things yourself, then the final product probably will not suffer if you go from 24/96k to 16/44.1. And if you want stellar results, simply send it off to a mastering house with an excellent rep, and let them take care of that conversion!

[Naagzh]

If I am just going to end up converting a 88.2 to 44.1 when I mix down, is the end result going to sound any better if I had just imported at 44.1 to begin with?

It could be, but it really depends on the A/D converter and the wordclock scheme. I've researched this quite a bit on my own, having read Bob Katz' "Mastering Audio: The Art and The Science". The gist of Bob Katz' opinion is that while there may be a slight difference in quality/fidelity between sample rates, there's a bigger difference to be heard between high and low quality converters. A really nice converter at 44.1kHz can easily sound better than a mediocre converter at 96kHz.

The Nyquist Theory mathematically postulates that in order to capture and reproduce the full audio spectrum, a recorder must sample at twice the frequency range (hence 44,100 samples for the range of 20 to 20,000 Hz). However, it has also been postulated that higher-order harmonics inaudible to the human ear (north of 20 kHz) may subtly interact with other frequencies within the audible range, and so those deserve to be captured as well, hence the 48 kHz and higher sampling rates. (In his book, Bob Katz' doesn't side with the Nyquist Theory or the oversampling idea; he just presents them.)

I was interested in upgrading the converters on my 002R, and then I checked out the Black Lion Audio modification. What is particularly interesting about this mod is that while it upgrades the converters (it uses the ones from the MOTU hardware which are a bit better than those in the 002R but aren't pricey), it also upgrades the internal wordclock circuit, and compares the results to a Mytek external converter (connected via ADAT). The results are pretty astounding: the 002R with the mod sounds brighter, fuller, and somehow deeper. Not drastically so, but enough that you might question the worth of a high-end converter. This leads me to believe that sound quality depends way more on the CONNECTION of the wordclock signal than was previously realized. In this case, a less-than-high-end converter connected to a good internal wordclock trumped the boutique converter connected via ADAT. Conclusion: jitter caused by the ADAT cable, probably because the phase detection circuit on the 002R is not so great.


You should run a simple test. Use a direct guitar track and play a song. Then increase the sampling rate and record the same song. Bounce them both down and listen.