|
24-bit question
|
View this Thread in Original format
| Microlab |
So I've been producing for years without really thinking about all that bits and sample rate thing when today suddenly it occurred to me "Why not try 24/96, just what my interface allows me?". A couple of articles and Youtube vids on the topic have certainly shed some light on the whole bit/rate thing however still the question is still there - How will I benefit should I transfer to 24/96 and will this really improve the quality of my mixes? Do all professional producers out there record at higher than 16/44.1? And increasing your bits is only possible when recording analogue audio and rendering your audio for the sake of mastering, and it is useless when producing a regular EDM track?
Any comments are welcomed! |
|
|
| evo8 |
imho monitoring at 96khz wont make any difference, a lot of plugins these days upsample to 96 anyway im guessing
bit depth: i usually export from Live in 24 bit then down to 16 bit for mp3 - once again im not sure it matters
you can convince yourself of the benefits but blind testing is the only way youll know for sure |
|
|
| DJ RANN |
Ahhh, I miss these questions. This is the Production forum of old.
Here's my take and I'll split them up in to two different arguments.
First sample rate....
Just talking strictly from a technical audio quality standpoint, yes, 96k is better than 44.1k.
Firstly, there's sevreal instruments and sounds that actually go above the inherent frequency limit of 44.1k (22k) so that isn't being properly captured or reproduced on a 44.1 system.
Then, there's a lot of evidence that shows that things like upper harmonics of certain sounds such strung, high wind and even perc instruments are only properly captured (and therefore represented) at a higher sample rate. In these cases, it's a little less about hearing the actual sound (mainly because it's outside the normal range on non-mutant hearing) but the harmonic artifacts being present. It sounds bizarre, but I've witnessed (heard?) it in action and it's real.
Then, you have to realize that sample rate is resolution on a horizontal plane. Think of this way: with images, once we get past 16 frames per second, our brain stops seeing individual frames or pictures, and starts to recognize a fluid image. Hearing is the same way however the threshold and way we hear is completely different so it's no where near as much of a finite thing like our sight is.
However and of course obviously, 44,100 "slices" is already so high that we have no idea this is a cut up waveform as such but to the trained ear, there is a difference in perceived quality when upgrading to higher sample rates - it results in more detail of represented waveforms.
There's also another argument that on sounds with a very low frequency and therefore a long waveform (in the double digit hz) the higher sample rate gives a "smoother" representation of the sound because it's such a multiple of the available frequency range.
The biggest argument for it in a real world setting is the extra quality, specifically detail in both tracking, and monitoring.
For recording, if you have a great mic/beautiful front end, there's a strong argument to capture it at as high a sample rate as possible. We can always go lower if needed as technology allows us to bounce down to a lower freq without any real loss so why not?*
For monitoring, it's the extra detail you get and therefore, the more you hear, the better you can do your job of engineering and mixing it.
One score engineer I've worked with, Shawn Murphy, only works in 96k - no 96k, no mix, and if you hear his score mixes (Force awakens, Bourne, Saving Private Ryan, Titanic etc) they kinda speak for themselves. Take a second to think about the kit you need to do 300+ channels of 96k.
So *technically* speaking, yes, 96k is better.
Why not in real world use?
1, Lower end sound cards, even though they can do 96k will actually introduce distortion and ultrasonic artifacts at 96k so you're not actually getting the full benefit and it may be better to stay at 44.1.
2, Disk space. Not too much of a concern these days, but understand the 96k files are a lot larger.
3, A lot of plugins are coded for 44.1 and cheat their way to 96 by oversampling.
4, Some argue that higher sample rates actually have lower fidelity because of the way PCM works are our ears hear sounds on a biological level, although I've only ever really seen this argument truly demonstrated at really high sample rates like 192k which I personally don't see merit in.
In other words, if you have a good pro soundcard, plenty of disk space, a pristine signal chain and are recording or monitoring in the right environment, then go for 96k.
Otherwise, you may get no benefit or even have worse results.
Now Bit Depth:
Firstly, 16bit does not actually technically cover the human rage of hearing so even though it's the standard for commercially produced audio, it's actually a compromise. yes, we're talking extremes here but nonetheless it technically doesn't do the whole task at hand in terms of covering the volumetric range of human hearing.
Bit depth isn't really about resolution, like sample rate is; it's about headroom, noise floor or subjectively said, breathing room.
Think about this: 16b calibrated at 0db, usually has mixes done at as much as -6dbfs below that. That means losing an entire bit of depth, which is now not just in the extreme poles of hearing range but in the normal audible hearing range.
24bit gives a range that is so far beyond our range of hearing, that we have so much mroe to play around with. Why is that important? Dynamics. You can make your sounds much more dynamic as there is a far greater range of volume to play with and the inherent noise floor related to the audio signal. It results in 48db of extra dynamic range.
That is it. There's no other true advantage of extra bit depth.
Should we use it? yes. For the relatively little extra space a higher bit depth uses, the dynamic range increase and lower relative noise floor is well worth it.
Now your real problem is that 24/96k takes up 3 times more disk space.
Finally and just to clarify, 32bit floating point has nothing to do with "bit depth" in this context. That's purely a computer based calculation for processing numbers in a system and is NOT bit depth used in recording or sound reproduction.
My advice is to use 24 bit wherever possible, and use 96k as well if you've got the setup. |
|
|
| Microlab |
| Whoa thanks for the detailed answer. But still the question - can I still record high-quality label friendly material at 16/44.1, right? |
|
|
| tehlord |
| RANN for president. |
|
|
| evo8 |
| quote: | Originally posted by Microlab
Whoa thanks for the detailed answer. But still the question - can I still record high-quality label friendly material at 16/44.1, right? |
of course! 44.1khz is a standard for a reason, guarantee you most of the professional dance music producers are doing just that |
|
|
| DJ RANN |
| quote: | Originally posted by Microlab
Whoa thanks for the detailed answer. But still the question - can I still record high-quality label friendly material at 16/44.1, right? |
yes, 16b/44.1k is fine - that's CD quality and the average listener will be none the wiser. However, if you are actually recording anything via a mic, at least go to 24bit in your preferences. You can always bounce down to 16b but you at least want the source recording in 24bit. |
|
|
| Microlab |
| Thank you all mates |
|
|
| Magnus |
| quote: | Originally posted by Microlab
Whoa thanks for the detailed answer. But still the question - can I still record high-quality label friendly material at 16/44.1, right? |
Absolutely. And RANN, amazing response. Thanks for that! :) |
|
|
|
|
