The differences, if any, show up on the top end (I'd say, 12k and up) and depend on converters and plug-in type and design. Cheap converters can sound a bit harsh and "smeared" at 44.1, and distortion and bad compression plug-ins, as well as many synths, can alias and cause treble crud if not designed properly.

Great converters make the difference vanishingly small, only perceptible to me on things such as bright drums and bright vocal/guitar transients. Synths and effects causing aliasing are a real-world issue, but a matter of poor implementation rather than an inherent limitation of a lower sample rate.

If you're hearing a difference in the low end you're doing something wrong, or are hearing treble differences in a way that perceptually suggests a change in the bass. I'd suggest anyone hearing a midrange or bass difference check with high quality converters a signal with no treble (for example a full mix that you have lo-passed above 5k with a sharpish filter) to make sure this isn't the case, because it really shouldn't be.

24/96

Thankfully CDs are dead...Kids wandering around that have never even seen a CD. So is 16/44.1k even relevant in conversation anymore?

The first time I heard 24/96 I could sense the difference. And that is about as honest as I can be. It was not qualifiable by me with catch phrases like “the top end sounded more open” etc... I could feel the difference in the quality of the music.

I work in and stick with 24/96 and do the digital conversion from those files to the highest quality downloadable format I can.

I have grown used to listening to digital downloads in the car and since we got past the MP3 stage to the higher Apple formats, I am really pleased with the results.

I think that by starting off with the highest quality sample rate you can and will make a difference in the digital download version down the line.

Pat

24/96

Thankfully CDs are dead...Kids wandering around that have never even seen a CD. So is 16/44.1k even relevant in conversation anymore?

The first time I heard 24/96 I could sense the difference. And that is about as honest as I can be. It was not qualifiable by me with catch phrases like “the top end sounded more open” etc... I could feel the difference in the quality of the music.

I work in and stick with 24/96 and do the digital conversion from those files to the highest quality downloadable format I can.

I have grown used to listening to digital downloads in the car and since we got past the MP3 stage to the higher Apple formats, I am really pleased with the results.

I think that by starting off with the highest quality sample rate you can and will make a difference in the digital download version down the line.

Pat

I believe many of the streaming services, which are a where lots of people listen, stream in 44.1, so it's still a part of everyday life.

Quite a few records, that have been mastered and released in digital, have not been re-mastered past 44.1, so some records you will only listen to at that rate, no matter where you play them.

If I'm doing something sound-designy like turning a pig snort into a giant alien I'll do 192k. Otherwise, 24/44.1.

44.1
solid and low distortion.

24/192

Audio recorded sound more relaxed, plugins rarely alias, virtual synths often change their sound usually for the better.

I lower resolution on samples or tracks with dedicated plugins for a more hazy sound.

Where possible I like to record at 24/192 because something about it feels more natural and the sound staging is a little bit better and the highs seem more extended, even compared to 96khz. This holds true on most converters I've used.

However, I found that with some audiophile-type DACs I've bought for my home listening system over the years, particularly those that upsample to DSD, sample rate doesn't make as much of a difference as bit depth. There seems to be some impact, but 24/48 or 24/96 seem perfectly adequate for non-PCM playback designs. Even 24/44.1 sounds ok. This tells me that actually the drawbacks of lower sample rates are mostly to do with the interface ADC and especially DAC in question. Granted, every standard chip-based PCM multichannel converter I've heard doesn't sound as good to me at lower rates. But it seems that from the perspective of the actual audio data, there isn't really much difference and with a good enough DAC any sample rate can sound great.

Interesting direction Jon Hopkins takes. I guess it isn't totally different from how some people really cherish 16 or 12 bit samplers in electronic music. On the other hand, I downloaded a 24/96 ambient techno album off Bandcamp the other day it sounds excellent to me. I would prefer that all music be recorded and released in 24 bit.

Used to track 24/88.2Khz but now just do 24/48Khz. I hear the difference enough to warrant 48Khz (plus the whole video thing) but not enough to justify to the higher rates anymore and the additional HD space they take up.

24/48. I do a lot of video stuff.
Tried 96. Listened real close. If there was an "improvement" difference, I couldn't define it. Maybe there was a difference, but bigger is not necessarily better. Just different.

In my so-so setup, i work in 48/24. I render to 44.1/16

but whats weird is i actually like the sound of 44.1/16 better. i am just worried that if i worked in 44/16 all the way through .. it will sound crappier when rendering to 44.1

i am still confused by all the 192 audio interfaces on the market. aren't most bedroom producers going to be working in 48/24 max due to cpu and storage limitations?


it could be a rock and roll thing and 48 is all you need in that realm.

If it just takes code to do it and not fancier hardware (I have no idea), then why not include it? The R&D is paid for at this point.

There are some inexpensive recording interfaces that max out at lower sample rates. And there are some setups that demand lower rates for the most channel count (ADAT connections for example). But yeah, most of the better ones give you the option whether you use it or not. I would be a little suspicious of something that was expensive that couldn't at least do 192. I doubt there are many people doing huge projects at 192 but I'm sure there are a few out there. There was some "big name" producer that swears by it in interviews. Might have been Chuck Ainlay, I don't remember.

<curmudgeon mode on>

1. Nothing is quite as misleading as a fully sighted, non-A-B comparison of two things that actually sound identical.

2. True ABX/DBT of different sample rates and bit depths is monumentally difficult to perform, so everybody's given up and reverts to heavily biased, sighted testing. Thus the "preference" for higher depths and rates.

3. ITB mixing is never done at 16/44, or even 24/44, it's all 32 or 64 bit floating-point.

4. 24 bit quantization offers no additional resolution, it only lowers the noise floor of the digital system. Distortion is not lowered, detail is not increased. Higher rates do not increase resolution, they increase bandwidth. Nyquist/Shannon says so.

5. True 24 bit dynamic range is not available from any common audio interface, they are all limited to about 20 bits of DR, though they produce 24 bit samples. That means they are all 24dB off of true 24 bit performance.

6. A properly dithered 16 bit system has a similar audible DR to a typical 24 bit ADC which is limited to 20 bits of real DR.

7. Original recordings made at 24 bits and released at 16 bits will sound worse that original 16 bit recordings unless properly dithered before truncation.

8. No acoustic environment can accommodate playback of 24 bit dynamic range without causing permanent damage to hearing.

9. Every acoustic environment (studio) is THE limiting factor to dynamic range, followed by microphone self noise, then preamp and system noise, and finally the ADC. None approach 20 bits of DR.

10. 24/48 is more than adequate for anything except general release, which for music will be 16/44 for a very long time yet. That's "general release" not "boutique".

<curmudgeon mode off>

While higher rates and more bit depths do no good, they also do no harm, other than take up storage space. So if you think you like it better, have at it. You're likely kidding yourself, but that's fine if the end result is happiness. We all kid ourselves into happiness all the time anyway.

Having A/Bx'ed the hell out of 24/48 vs 24/96 vs 16/44 and 16/48 to my ears the sample rate is more important than the bitrate. Turns out I prefer 16/48 better than 24/44...

Going to 96 does nothing for me except for some very rare corner cases (soloed Trombones, Mallets).

Having 24 bits doesn't hurt in this day and age of cheap storage and fast processors. So pragmatically everything is 24/48.

My take is some things get 'better' at higher rates but not necessarily musically usefully better so I guess a bit like Hopkins. I find high rates to make for more clean top end but also a perceived loss of "physicality/meat". And meat is not something I am prepared to give away as to me that makes the music speak more cerebrally than kinestetically and that doesn't work for me. So I land at 48 for still good meat and slightly better tops than 44.1.

This at 24. I have been tempted to try something at 16 though as that also has a similar association in my mind as Hopkins points at, where 24 can go bigger/taller but 16 is compact and straight to the point. I remember when the first records that were started at 24 bit came out and how that seemed to unlock a different place. Different as in win some lose some. Bigger but not necessarily more emotionally powerful.

But they do good though, objectively. For example reducing aliasing in digital mixing, when using dynamics and distortion plugins. And a lower noise floor is a lower noise floor, it just is.

I'm not going to argue about hearing, even though I'd like to. No one will agree, it almost never occurs.

when comparing RME BS Pro..
to the Steinberg UR44C ..
to the Tascam 4x4 ..
is the recommendation that we should not be looking at max sample rate but at at what sample rate the interface can optimally perform in without noise an distortion?

32 bit is just magic numbers.
There is no such thing as "without noise and distortion". Both are measurements relative to something else, ratios. All of the interfaces above are far more than adequate. You won't find any meaningful numbers to compare.

Objective: "not influenced by personal feelings or opinions in considering and representing facts." In other words, measureable.
None of that has anything to do with interface bit depth or sampling rate. DSP within a DAW is aways done at 32 bit or 64 bit floating point, which is effectively bit depth far beyond that of any interface.
...and is measurable. Objectively. Read the specs, sometimes it's in there. 24 bit interfaces almost never provide 24 bit noise floors. "Almost never" because there are a very few that fake it.

George Massenburg prefers to work at 192k as well.

But if you use expensive equipment where some or all devices involved provide around 20bit DR you are fine with 24 bit.

Using 16 bit with this equipment would mean you give up in advance and limit to the 16 bit which could possibly not be the max DR.....

This has been measured. Sample rate matters. Maybe you're not aware of it. Just to clarify, I was talking about recording and mixing at higher sample rates, I was not talking about bit depth.

Thanks for reminding me of words that I already know the definition of, it's a little patronizing.

I am aware that converters have a noise floor. 20 bits is better than 16 bits. So we can agree on that one.

Take a look into “dithering” and “noise shaping”.

I'm aware of what sample rate changes, I cited it exactly in an earlier post. "Matters"...yes, it changes total bandwidth, which is measurable. Audible? Not so much, never been proven beyond the extremes.
Oh, you're talking about recording and mixing! Oh, I thought you were talking about making coffee.

Ok now you're just being a jerk.

24/96. It sounds better to me, after trying other (lower) settings.

Mostly 48k multitrack, 44.1k CD prep. 44.1 discs are upsampled for playback to 384k, makes a HUGE difference. Then they sound more like 24 bit.

Speakers all do 40k+ hz for playback. My cat hears that stuff. If they run out of the room I know something is amiss up there...

I do that too. Sort of, everything that hits the DAC plays back at 192k here, it makes all the various sample rate sources sound more consistent on my DAC.

When the stereo monitor DAC is always switching rates you tend to hear more differences. It's fun and interesting but I like the consistency of the DAC always running at the same settings regardless of the source material.