48kHz vs 44.1kHz

LasVegasIRA

LasVegasIRA

I already have a multi-track recorder that records in 16bit/44.1kHz. However, I'm reading on here that I want a minimum of 16bit/48kHz, and that 24bit/96kHz is even better.

So my question is, what is the real difference between them all?

After reading about this on other websites, it doesn't seem as though there's a discernible difference in quality. And since humans can't hear above 20kHz, I can't imagine that the difference is even audible.

And as far as bit rate, I understand that you get more dynamics in the recording with 24bit because you're getting more "headroom," but how much would this really affect your audio takes?

Then there's the whole music aspect. If I want to use music from a commercial cd or an mp3, it's sampled at 16bit/44.1kHz, so does that mean the quality could be compromised with the music?

This is all very confusing... I guess I just want to know if using my 16bit/44.1kHz recorder is going to ruin the audio I get on a shoot. I've used it before on one of my short films, but if I'm going to make a feature length film, I don't want audio that will sound like crap if it's played through a theater sound system...
 
Sort by date Sort by votes
First about bit rates; I'll just keep it very simple and not go into the maths that I have a hard time understanding myself... Recording at 24 bits gives you more headroom than 16 bit; in other words, you can record a little bit louder before peaking. Next stop, true 32 bit...

Now I'll address
since humans can't hear above 20kHz, I can't imagine that the difference is even audible.
Click to expand...

Sound frequencies don't exist in a vacuum. Hmmm, that's literally true as there is no sound where there is no atmosphere since sound moves air molecules. But individual sounds are made up of sound waves by the hundreds of thousands. Stated differently, sounds occupy an entire frequency range, not just a single frequency. For instance, the human voice occupies 80Hz to about 1-3 kHz; however, the sibilant sounds go well above 8kHz. So you can see that it is the interactions of these frequency ranges that give us speech - even when speaking we vary the pitch up and down and use sibilants (C, F, H, J, S, X, Z); and all of these sounds interact, they aren't stagnant. The way that sounds interact with themselves and other sounds create more subtle frequencies (with musical instruments these are called harmonic overtones). For those who are visually oriented think of waves in a pond. Drop one pebble and that is a single frequency. Drop 1,000 pebbles, stones and rocks of varying sizes all over the pond and the waves interact with each other by pushing against each other, washing over each other, crossing at various angles creating new waves in thousands of permutations.

Now, when it comes to sampling rates it's slightly different. You are taking a real-time digital sample of the audio X number of times per second. So at 44.1kHz you are sampling the audio 44,100 times per second, at 48kHz you are sampling the audio 48,000 times per second, at 96kHz - 96,000 times per second, etc. The extra data density provides for more clarity and definition. Yes, you can't hear the upper frequencies, but their interactions with the frequencies you can hear makes the sound richer and more detailed - provided you have a playback system and a listening environment (and your ears aren't damaged) to enjoy the difference. 99.99% of people will never notice the difference. However, where folks in the music and film audio worlds will notice the difference is when processing sounds recorded at the higher sample rates; there will be less degradation of high end clarity and less aliasing and other audio artifacts when doing lots of digital processing like noise reduction and time stretching since there is more data to play with. That includes sounds above human hearing that still interact with the frequencies humans can hear. BTW, HD audio is 24bit/96kHz.

Where 44.1kHz recording can affect your film audio is that it often is imported at the wrong "speed" into NLEs. The audio recorded at 48kHz will match up to one second of visuals (24 frames, for example) but the one second of audio recorded at 44.1kHz will import incorrectly and will play back in 0.91875 seconds. After a minute or so the audio to visual drift will become noticeable.

Just as a little bit of history, 44.1kHz was used (instead of 48kHz) so that the standard vinyl double album could fit on a CD.

I hope that all of this helps.
 
Upvote 0 Downvote
Excellent info, Alcove. Thanks. I have a question though:

Alcove Audio said:
Where 44.1kHz recording can affect your film audio is that it often is imported at the wrong "speed" into NLEs. The audio recorded at 48kHz will match up to one second of visuals (24 frames, for example) but the one second of audio recorded at 44.1kHz will import incorrectly and will play back in 0.91875 seconds. After a minute or so the audio to visual drift will become noticeable.
Click to expand...

What if I convert the audio sample from 44.1 to 48kHz before I add it to the clip? Will that "cure" the drifting issue?

Just as a little bit of history, 44.1kHz was used (instead of 48kHz) so that the standard vinyl double album could fit on a CD.
Click to expand...

I read that on another site. I found it very interesting until a couple of people called BS on it... The guy originally stated that someone at Sony, when developing the cd, thought it was good rule of thumb to be able to put Beethoven's 9th on a single disc and therefore lowered the sampling rate to accomplish this. But a few other people stated that this was a fallacy and that the sampling rate was due to the technological limitations of the time. They went on to state that if it were a time consideration, they would have just made the discs a little larger... Still a good story though...

I look forward to your answer. And thanks again!
 
Upvote 0 Downvote
LasVegasIRA said:
What if I convert the audio sample from 44.1 to 48kHz before I add it to the clip? Will that "cure" the drifting issue?
Click to expand...

Yes.

LasVegasIRA said:
I read that on another site. I found it very interesting until a couple of people called BS on it... The guy originally stated that someone at Sony, when developing the cd, thought it was good rule of thumb to be able to put Beethoven's 9th on a single disc and therefore lowered the sampling rate to accomplish this. But a few other people stated that this was a fallacy and that the sampling rate was due to the technological limitations of the time. They went on to state that if it were a time consideration, they would have just made the discs a little larger... Still a good story though...

I look forward to your answer. And thanks again!
Click to expand...

All of the answers are completely correct and 100% wrong. They are all sort of interlocked with cost/data ratios, not to mention marketing, becoming involved as well. Yes, they could have made a CD bigger, but a 48kHz CD with enough data space for a double album (or Beethoven's 9th :D) would have been closer to the size of a vinyl 45rpm single and the marketing folks did not want CDs associated with vinyl in any way; they wanted to so "See! You get two albums on a disk half the size of a 45." That was the hook; it's small and it doesn't scratch or skip (:lol:). So the size of the disk was determined more by marketing than by technological requirements. Being unable to change the size of the disk to accommodate the data requirements they changed the data requirements down to 44.1kHz. Another consideration was the cost/ratio involved in making 48kHz disks and playback devices; they would have been significantly more expensive. There will be many who will argue both sides, and frankly, I don't really care. The stories are fun, and, as they said at the end of "The Man Who Shot Liberty Valance" - "When the facts conflict with the legend, print the legend."
 
Upvote 0 Downvote
Not sure of the technicalities behind it, or the validity tbh, but have heard from a couple of sound designers I was workng with that they prefer 48KHz over 96 as there is some theory going around that the DACs operate slightly unnaturally at 96. I think it may be linked to most recorder's hardware records natively at 48 but the 96 is often a software mechanism.
24/48 tends to be what everyone asks me to record so that is my default.
I can't pretend to understand but the sound designer gets what the sound designer wants.
 
Upvote 0 Downvote
There is a difference between recording at 16 bit and having a final mastered product at 16 bit.
The dynamic range of 24-bit recording is about 144dB. With 16-bit it is only 96dB. That is a much larger range before noise can become an issue. Also 16 bit recordings can tend to sound more brittle than 24 bit recordings. If possible record at 24 bit, even if the final master is going to be dithered down and wind up at 16 bit.
The difference in sound quality between 44.1kHz and 48k is neglible. You are not likely to hear a difference in quality between these too sample rates, but you are likely to hear a difference between a whole lot of tracks recorded and mixed at 16 bit v 24 bit.
 
Upvote 0 Downvote
You must log in or register to reply here.
Back
Top