I ran into an
article yesterday sort of while looking for something else. While it is nothing I didn't know already, it was a good quick reference for file-sizes at the various bit and sampling rates. It goes into more detail, however, about why 24-bit is preferred over 16-bit but why 96 kHz is not necessarily superior to 44.1 kHz.
The first bit is easy: use 24-bit. In fact, I think all audio should be 24-bit. 16-bit might have been great in the 80's when CDs first came out but the dynamic range of 16-bit audio is actually quite small. Maybe in today's louder is better world it doesn't mean much, but it should! Most DVDs and Blu-Rays offer 24-bit audio as well, and the reasoning is so you can clearly hear both the soft and loud parts of a movie. Consider your average horror film. The creepy scenes are partly creepy due to the soft weird music and then a really loud noise. Modern music should not be just off and on. There's beautiful variances in volume that are getting missed due to the lower dynamic range and we need to get past the
loudness war. There's this thing called a volume knob, after all. Even if the resulting medium is going to be 16-bit, having the dynamic range for mastering is still quite important. Long story short, record in 24-bit. Do it.
The sampling rate issue I guess is more blurred. Yes, there is this
Nyquist frequency concept which basically says that you need 2x the sampling rate of your highest frequency to avoid artifacts - or something like that. So 44.1 kHz (the sampling rate of CDs) is good enough since the human ear can only hear up to 22 kHz (half the sampling rate). It's not quite that simple though. The article I ran into points out that some audiophiles consider the greater frequencies as adding depth since the higher frequencies somehow change the lower ones. In other words, while there is no scientific qualification for this, 96 kHz audio simply sounds better.
The practical terms there are quite blurred. However, this whole debate, in my opinion, seems moot for one reason. When you are recording audio, the more samples you have, the more you can massage the audio without introducing artifacts. If I take a drum loop, for instance, and want to both half the pitch and speed, I simply play the samples at a slower rate. 44.1 kHz at half-speed is 22.05 kHz and, at that point, the human ear could hear artifacts. Half of 96khz is 48khz which is still enough to avoid perceptible artifacts. There are other things to keep in mind here. How good is the sampling hardware and what sort of audio distortion might it be contributing for instance. I mean, if you have a noisy sampler that can sample at 96 kHz, it's worse than a very quiet and precise sampler than only does 44.1 kHz.
Nonetheless, hard-drive space is cheap and 24-bit/96 kHz audio is still only 33MB/min (as opposed to about 10MB/min for 16-bit/44 kHz). When you consider that you can buy drives up into the terabytes there's plenty of room for mastering. Even my most complex songs still fit on a DVD and that was with all the raw audio included. That's just not that big of a deal and the ability to do some interesting things with the audio due to the higher sampling rates is well worth it.
So, while this debate will probably rage on, in my opinion, if you can record in 24-bit/96 kHz do it. 192 kHz? Well, that might be getting a bit ridiculous. I would also heavily push for offering your finished products in 24-bit as well (perhaps at 44.1 or 48 kHz). In fact, I really wish places like iTunes would offer at least lossless audio as an option. 24-bits, while just as unlikely as offering lossless audio, would be wonderful too!
(By the way, if you were curious about how my DAW evaluation is going, the winner is likely going to be Ableton Live. That thing owns face!)
Link to
Facebook comments.
Tags:
, loudness war, sampling rate, bit-depth, 24-bit/96 kHz