44.1 or 48khz

[blackcom]

Hi, Andy!
Just wondering wether you prefer to stay in 44.1 or 48khz when mixing and mastering, and why?

best!
Marius Strand

[Hewy80]

Hey mate most pros use 96KHz. My sound card supports that. If you understand anything about sample rates you would realise why the higher sample rate the better quality.
Basically you have an analog signal (guitar, drums, whatever) and when it goes into your pc it needs to be converted into digital. Digital has limitations and one of those limitations is that it only simulates a true analog wave. The sample rate determines when the pc samples the input voltage. The higher sample rate the more often the signal is being measured giving a better image of the analog sound. Each time it samples it makes a digital representation of the input voltage. So basically what you hear from CDs and your pc is lots of little square waves instead of a smooth analog wave. I hope you understand this its sorta hard to explain well you really need a picture to understand the difference.
Also 48KHz means that it samples 48000 times a second where as 96KHz is sampled 96000 times a second. Higher sample rates end up as bigger files as well so u can fill your hdd quickly using the higher rate.
When mixing pros tend to use the highest sample rate and bit depth they can so when they mixdown there is a higher clarity.

[Andy Sneap]

44.1, as I think it sounds better sticking there, rather than sample rate converting later.

[jeronimo]

Originally Posted by Hewy80

Hey mate most pros use 96KHz. My sound card supports that. If you understand anything about sample rates you would realise why the higher sample rate the better quality.
Basically you have an analog signal (guitar, drums, whatever) and when it goes into your pc it needs to be converted into digital. Digital has limitations and one of those limitations is that it only simulates a true analog wave. The sample rate determines when the pc samples the input voltage. The higher sample rate the more often the signal is being measured giving a better image of the analog sound. Each time it samples it makes a digital representation of the input voltage. So basically what you hear from CDs and your pc is lots of little square waves instead of a smooth analog wave. I hope you understand this its sorta hard to explain well you really need a picture to understand the difference.
Also 48KHz means that it samples 48000 times a second where as 96KHz is sampled 96000 times a second. Higher sample rates end up as bigger files as well so u can fill your hdd quickly using the higher rate.
When mixing pros tend to use the highest sample rate and bit depth they can so when they mixdown there is a higher clarity.

Your explanation about SR is right! But your knowledge on how the pros work... is not...

[97reb]

this thread .................................... ten foot pole.

[Hewy80]

Originally Posted by jeronimo

Your explanation about SR is right! But your knowledge on how the pros work... is not...

I think you'll find that most professionals do work in 24/96 thats why all the pro sound cards are 24/96. It seems Andy does not which surprises me however 99% of people wont notice on the final product anyway. Guess its a matter of personal choice and I'd prefer to keep the quality as pristine as possible up until mixdown/mastering.

[egan.]

In my experience most people will use 24 bit but still prefer 44.1 b/c they do not like artifacts of sample rate reductions. Most pro sound cards are 24bit/96k because that is where technology is going however, unless you're recording for DVDA or SACD the sample rate reduction is going to hurt the audio more than it helps b/c the software or hardware (depending on what you are using) has to recalculate the placement of each sample. Going down from 96 to 48 you literally cut the samples in half, but down to 44.1 things get very hairy and the long and short of it is that the information is far less accurate than if you had just recorded 44.1 in the first place. I've read a few articles over the last year that recomend that you run a 96k master analog to a 44.1 machine rather than doing a sample rate reduction.

[Andy Sneap]

egan you are correct, my sentiments exactly

[jeronimo]

Same thought here...
I've never ever seen a pro recording at anything higher than 48K and 99% of the pros record at 44.1K, and I know a lot of "pros" in Brazil, US, UK, Australia, Sweden...

[underground69]

What about bit depth then ? Better to stay at 16 bits or is it better to record at 32 bits and then dither down to 16 bits ?
I understood your explanations about staying at 44 Khz I've heard it before, but then why do most people recommend to work in 32 bits then ? Cause in this case too, you have to get back to 16 bits to mixdown on CD anyway.
thanks for your answers !

[blackcom]

Ok, 44.1 is better then 48 then, but how about 88.2khz wich many audio apps support. Going from 88.2 to 44.1 would just be taking half of the samples, or cut in half as like when going from 96 to 48. I guess this would not cause the same amount of artifacts when doing a sample rate conversion.
Some guy also told med that the "calculated result" of effects/plugins comes out better when operating at higher samplerates... he also told me that those higher frequencies will suffer some when running the sound through plugins prosessing the sound so it would be smart to have some to spare.

Nonsense or not? L'rrrrrrrr

[egan.]

Originally Posted by blackcom

Going from 88.2 to 44.1 would just be taking half of the samples, or cut in half as like when going from 96 to 48.

Well I've honestly never heard anyone try it but I can promise you that my "cutting the # of samples in half" line was still a gross over simplification of the process and there we're certainly still be artifacts.

In regards to the bitrate question, the sample rate effects the bitrate, but the bitrate does not effect the sample rate. The bitrate tells you ampltitude accuracy (dB) while the samplerate is measure in time or frames. The amplitude is only measured when the sample is taken. So when you change the sample rate your DAW or whatever makes assumptions about what is going on in between the samples.
When you do a bitrate reduction you lower your accuracy, but most software and hardware adds dither (basically noise) which smooths transitions in between bits. So the long and short of it is that you are defineitly losing accuracy in a bitrate reduction, but not nearly as much. BUT, you need to consider that if all of your tracks are 24 bit and you are 24 bit right up until the two track stage you will gain accuracy over recording 16 bit the whole way b/c the individual tracks will maintain much of their accuracy. This is b/c amplitude in the mix is cumulative so as long as there are other things going on in the mix the accuracy of the individual tracks could stay mostly intact. Once again I've grossly over-simplified.

[Hewy80]

Originally Posted by egan.

Well I've honestly never heard anyone try it but I can promise you that my "cutting the # of samples in half" line was still a gross over simplification of the process and there we're certainly still be artifacts.

In regards to the bitrate question, the sample rate effects the bitrate, but the bitrate does not effect the sample rate. The bitrate tells you ampltitude accuracy (dB) while the samplerate is measure in time or frames. The amplitude is only measured when the sample is taken. So when you change the sample rate your DAW or whatever makes assumptions about what is going on in between the samples.
When you do a bitrate reduction you lower your accuracy, but most software and hardware adds dither (basically noise) which smooths transitions in between bits. So the long and short of it is that you are defineitly losing accuracy in a bitrate reduction, but not nearly as much. BUT, you need to consider that if all of your tracks are 24 bit and you are 24 bit right up until the two track stage you will gain accuracy over recording 16 bit the whole way b/c the individual tracks will maintain much of their accuracy. This is b/c amplitude in the mix is cumulative so as long as there are other things going on in the mix the accuracy of the individual tracks could stay mostly intact. Once again I've grossly over-simplified.

Also when recording at the higher bit depth you must ensure that the signal coming in is as hot as possible otherwise the signal will not be a true 24bit. Cant remember exactly but for every lets say 3dB you loose 1 bit depth ( I know this is not exact I just cant remember the nubers its something like that tho)

And yeah guys after reading a bit more about the sample rate thing you guys are right about artifacts when resampling the files so I stand corrected. Although I do think that these artifacts are very dependant on the software you are using and the algorithms it uses to downsample the files. I guess when I get a bit more experience (no pun intended) recording I'll be able to tell the difference myself between a file that has been downsampled and one that has not.

[Kevin Kitchel]

I hear a difference in 44.1 to 96. I personally think that you should keep everything in the highest/best sampling rate until the end. Most of the pros that I've met or read about use whatever they have. If they have access to 96k, they use it, if not, oh well. I use the highest sampling rate I can, and I think it does help. Also, I know that in the past, mastering engineers would upsample and downsample in the analog domain, but in the last 3-4 years, resampling has come a long way, and artifacts aren't really a factor at the mastering house. And one point, 88.2 => 44.1 isn't just as simple as throwing away every other sample, that would lead to artifacts, rather than smoothing out the waveform like a resample does.

[jeronimo]

I think (I never heard it) every software downsampling will degrade the sound...
Sometimes we just need to measure... does another A/D/A will degrade the sound more or less than a software downsampling?

[Kevin Kitchel]

I'm pretty enclined to go with the software over the hardware.