Sample rates

Does anyone still record at 44.1Khz?
If so, why?
Are you going to release the music on a physical CD?
Does anyone still do this?
I do lots of mixing and production for many clients and when they send me the stems, they are nearly always still at 44.1Khz, unless I remember to tell them to send them with a 48Khz/24bit spec.
So it seems that many people are still stuck in the old CD standard and even have their music creation apps set at 44.1Khz!

Hi,

Do you hear a difference comparing 44.1 vs 48kHz?

1 Like

I use 44.1kHz mostly because I don’t want to have some Projects at one sample rate and others at another.

I’ve been working with audio for 35 years - I can’t really tell the difference. Then again maybe I’m going deaf…

44.1khz for music production
48khz for TV/film

As the previous poster said - I prefer to stick to those rules so I don’t get confused and mix up my sample rates in projects.

Bring back 1/4 inch tape :wink:

1 Like

Streaming services are where people consume the most music today. And almost all of them only accept 44.1 kHz, so there you go.
I mix in 48kHz and export in 44.1.

2 Likes

The difference is when mixing, you have more margin below Nyquist at 48 kHz and it is less prone to aliasing.

For the final export there will be no difference between 44.1 and 48, since aliasing is generated during mixing, it will be printed in the file, regardless of the sample rate in which you export, so it is necessary to set the project and record at higher sample rates in first place to avoid aliasing.

1 Like

Bingo! Was going to say the same thing. Record at the higher sample rates and bit rates then when you are mixing or using processing affects you won’t lose a lot. By nature, processing on the audio file is destructive. This gives you more head room with higher rates.

1 Like

Hi,

Then I recommend to record in the double Sample Rate of the expected result.

In my opinion it’s better to record and mix in 44.1kHz than 48kHz, if the output should be 44.1kHz. The last sample conversion (resampling) by un-whole (sorry, this is probably not English) number is worse than starting at 44.1 from the very beginning.

I use 48 kHz/24-bit by default. This is simply the standard for film/television. (Occasionally, 48 kHz/16-bit is also desired. But then I still choose 24-bit for processing).
For concerts, which are later released on Blu-ray, it is occasionally 96 kHz/24-bit.
Can you hear the difference? Our customers say yes. And some buyers of the Blu-rays claim the same. Since the concerts were already recorded in this resolution, we have no choice anyway. :wink:

This is the case with bit depth, but not with sample rate. Doing so will only remove the high frequencies but it does not degrade the audio, there’s no way people could hear the difference between let’s say 192 → 44.1 and 44.1 → 44.1 . The only way to “see” the difference would be to do a null test with a spectrogram.

1 Like

Hi,

Not really. It resamples the signal. It means at one second, you don’t have 48000 nodes (samples) anymore, but 44100 samples in the second. So the engine has to interpolate/find the new position of every single node.

I know how that works, if you go 96 to 48 there won’t be enough samples to reproduce the high frequencies that were present at 96 kHz, so the shape of the waveform is indeed modified where these high frequencies were supposed to be, but when doing a null test the remaining noise/difference is at -180 dB, even devices that operate at 24 bit cannot reproduce such low levels.
The question is, can you hear the difference ? Personally, I can’t. :grin:

Hi,

I don’t hear 96 vs 48kHz. I don’t hear neither 48 vs 44.1kHz. Therefore it doesn’t make sense to record in 48kHz and then downsample (what definitely affects the signal out of my control). Better to start in 44.1 already for me.

I understand your point, but at 44.1 kHz you are closer to Nyquist, if you put too many saturation/harmonics plugins, you’ll end with aliasing much faster, and this is actually the worst thing, as aliasing is audible, once it’s there it will be rendered into the final file and there’s no way to get rid of it, apart from increasing the working sample rate in first place.

1 Like

There’s also the argument that the filters used are different and potentially audible, meaning that you might like a filter design better when the converter is operating at 48kHz as opposed to 24kHz.

2 Likes