That's not the point.

 

44.1 kHz and 48 kHz basically share the same sonic properties. The only reason for these two different basic rates (specifically the existence of 44.1) is a historical compatibility with both NTSC and PAL when using video equipment for the conversion.
You should just stick to whatever rate you're using from the beginning.

Interestingly, when you're uploading 48 kHz (or probably higher) to YouTube, it stays at 48 kHz for Opus, but is downsampled to 44.1 kHz for AAC. For example ABBA's new "I Still Have Faith in You": https://youtu.be/pAzEY1MfXrQ

Code:

$ youtube-dl -F https://www.youtube.com/watch?v=pAzEY1MfXrQ
[...]
[info] Available formats for pAzEY1MfXrQ:
format code  extension  resolution note
249          webm       audio only tiny   49k , webm_dash container, opus @ 49k (48000Hz), 1.90MiB
250          webm       audio only tiny   66k , webm_dash container, opus @ 66k (48000Hz), 2.54MiB
140          m4a        audio only tiny  129k , m4a_dash container, mp4a.40.2@129k (44100Hz), 4.93MiB
251          webm       audio only tiny  131k , webm_dash container, opus @131k (48000Hz), 5.02MiB
[...]

 

https://www.soundonsound.com/techniques/secrets-mix-engineers-robert-orton

https://www.soundonsound.com/people/sandy-vee-recording-katy-perrys-firework

 

If your computer has enough CPU power, use 96 kHz. This way your plugins sound good even if they do not have internal oversampling.

But if your CPU can't handle the extra overhead of the high sample rates, then just use 48 kHz.

 

Because they don't know the basics.

 

I record at 96k (for latency reasons), mix at 192k because I use little or no plug-in oversampling, and I render un-dithered at 96k 32bit. Later I convert and dither to other formats as needed.

 

Reasons I use 44.1:
1. The majority of samples are in 44.1 which I don't want to upsample
2. Even though RX8 is good downsampler, it is not a integer multiple, unfortunately nearly everything must still go to 44.1
3. Nearly all my Nebula libs are only given in 44.1 or 96
4. 96 is not great for the resources I have i.e. computational demand when I can oversample nearly everything
5. the difference of aliasing when I think this might occur i.e. for a cutoff ~20kHz that might bleed a small amount of aliasing at 44.1, I will oversample where possible and then any thing else that gets through is negligible

everyone to their own but for me rn 44.1 will never not suffice even if I got a record to chart/prestigious level. it will not be the sample rate that causes a bad song if you utilise oversampling and don't make strange/silly non-linear distortion decisions such as non-oversampled excessive saturation & distortion. of course this will be a whole different matter for film, where the expected delivery is 48kHz.

edit: just wanted to add, if one day I have a super computer that can handle 100+ plugins at a higher sample rate, I will move up to 88.2kHz. but that is just because 'I can'. If it is as the result of one of the previous trade-offs, I simply don't see it being worth it. go 44.1 all the way through and don't overcomplicate it. sonically my mix will sound no different because I will mix properly.

 

Last edited: Sep 6, 2021

Wow, that sounds dirty play all right.

 

Logic Pro

 

logic asks to change the sample rate according to project but it also changes the tone little bit. specially when re pitching compared to 44.

 

seems like it's ok to try out 48Khz but what about samples up-sampling will it cause any issues?

 

Upsample with Voxengo's R8Brain free or Pro, either is excellent and you'll be good. SRC is still quite terrible in some DAWs and plugins. Like Wavelab and old Cubase, old Live etc. Check https://src.infinitewave.ca/ and see. SRC is terribly important for good sounding production.

 

16 is sufficient for most lossy or lossless audio (at reasonable leves) for listening. 16b has a excellent noise floor of c. 120dB. ALthough lossy doesnt have a traditional bitdepth, masking is used so very quiet things won't get stored anyway.

DVDV & BD is specced to 48k+ only, so if you have 44k audio, you must upscale to 48k. Also, 50-90% of all HR music is upscaled (depending on genre) . I think I checked some stuff from the "Acoustic Reality Experience" label in the past, and they upscale 2-6ch 44-48k to 8ch 96k.

see also : https://archimago.blogspot.com/2013/07/list-suspected-44-or-48khz-pcm.html

 

Thanks for the info and the link.

 

we musicians hear the differences, 90% of the non musicians usually do not.

I spent so many years in improving sound quality without understanding how the listeners/buyers function for my genre (EDM).
todays most people like a Song/Track or they don't... simple as is.. and sad as well..

There are Tracks on Beatport, Traxsource etc etc with horrible sound quality but successful..
Many listeners/customers got much lower standards than musicians ..

(I master in 48Khz 24bit wave format for all commercial platforms for myself and all my friends and customers since years)

 

Years and years ago, I was subscribed to Computer Music & Future Music magazines (I'm sorry to say I had to dump my collection when I got divorced). I remember reading that mixing & mastering at 16-bit resolution was likely to create artifacts when applying any kind of FX, since the bit depth resolution made a difference in data processing by any plug-in. I also remember reading that using higher sample rates was only advisable if you were composing, mixing and / or mastering for hi-resolution audio formats (like those used in Blu Ray, in Cinemas, etc) because higher sample resolutions allowed you to get louder bass sounds (like the ones needed for 5.1, 7.1 or Dolby Surround in special sound FXs, like explosions, earthquake rumbling...). Since I normally make & record music, I use 48kHz/24-bits, since it does not put much pressure on my CPU, and the results are almost professional. Most DJs make their mixes from stems at 48kHz/24-bits, from what I've read.

 

""The session was in 16/44.1. I always work like that, because where do people buy their music? On iTunes. Where do they watch videos? On YouTube. If I'm recording an orchestra — and I'm talking about doing a movie score — I may go to higher resolution, but I don't think it's worth recording in 24/96 for pop songs. Of course, 16/44.1 helps to make the session easier to manage, but especially because I'm using the UAD Quad stuff, I'm not in danger of running out of processing power, even though I do everything in the box. I don't use outboard gear at all — these days many plug-ins are as good as the out‑of‑the‑box counterpart. I've compared the outboard Manley Massive Passive with the plug‑in, and I don't hear any difference, it's incredible."

So, the primary reason back in 2009 for using 16/44.1 was mainly hardware limitations which was understandable at the time… but I guess if he was asked today, he would probably respond something els in terms of bit depth and sample rate…

 

Last edited: Sep 5, 2021

This software works better than Voxengo R8Brain in my opinion:
http://www.sonicillusions.co.uk/finalcd.htm

04/04/21 - Added support for upsampling and fix missing last sample bug (v0.28)
07/08/21 - Support gain scaling and dither only (no upsampling) and allow gains from 0.1 to 10.0 (v0.29)

 

Modern music production

 

the higher the frequency, the more points in digital audio that is, the sine will be a true sine and not a saw, the sound is more natural and rounded if resources are available, work at 96khz (especially modulation and processing will work much better) not in vain in many plugins up to 1 MHz there is an upsample uad also uses this
many daw also corrected SRC conversion ableton and wavelab gives a result close to ideal maybe they use Sox technology src.infinitewave.ca
who say the opposite is disingenuous or is a fan of akai s950=)