96kHz or 192kHz interface?

[thiswayup2010]

Hey.

What's your opinion on it - is it essential to record in 192k? If the final CD or Mp3 is gonna be 44.1k anyway....surely 96k is more than adequate?

Assuming always 24-bit of course.

[hollowsun]

Many would argue (myself included) that 96kHz is overkill, let alone 192!

And depends on the music you're recording...

A string quartet with great players and great instruments in a beautiful acoustic carefully recorded with well placed and well selected mics through quality A-Ds MIGHT benefit from 96kHz and you MIGHT hear a difference if you're listening in a well treated control room on top of the range monitors (and you're assuming that your target listeners are audiophiles with a quality hi-fi who MIGHT also hear a difference ... perhaps).

But otherwise.... nah!

essential to record in 192k? If the final CD or Mp3 is gonna be 44.1k anyway

Instead of 'essential', I think the word you were maybe looking for was 'insane'!

[humandrums]

you MIGHT hear a difference if you're listening in a well treated control room on top of the range monitors

id defo have to disagree on this having upgraded from behringer ada8000's that go upto 48khz to motu 8 pre's to go with my motu 828 mk3 that go upto 96khz same mics/monitors/instruments, and i defo defo could notice one hell of a difference, but that might just be me

[ConcertinaChap]

Hugh has written a few times on the forum about the maths on this, which AIUI shows that the sampling frequency can represent waveforms at up to half its own frequency. Thus 44.1kHz is capable of frequencies over 20kHz. 192kHz would probably be useful in a studio run by bats but not otherwise.

CC

[Jack Ruston]

Basically there are advantages of 88.2 or 96 over 44.1 or 48. These are mainly that the reconstruction filter is moved further away from the audible band, and that some processes like compression seem to behave in a rather more pleasing way at higher sample rates.

The amount of 'difference' depends on your converters. Extremely high end conversion tends to sound closer between the rates, because of the quality of the filtering stage. More affordable converters can sound way better at higher rates because the difficult bit is moved away from what we can hear.

Higher rates carry a processing and storage hit.

There is no credible argument for the general use of 192.

Personally I find that there's something nice about 88.2 on gentler acoustic type stuff. Heavier rock...44.1 usually.

J

[jaminem]

you MIGHT hear a difference if you're listening in a well treated control room on top of the range monitors

id defo have to disagree on this having upgraded from behringer ada8000's that go upto 48khz to motu 8 pre's to go with my motu 828 mk3 that go upto 96khz same mics/monitors/instruments, and i defo defo could notice one hell of a difference, but that might just be me

or maybe you're hearing the difference between a £150 interface and a £650 one?

[chris...]

id defo have to disagree on this having upgraded from behringer ada8000's that go upto 48khz to motu 8 pre's to go with my motu 828 mk3 that go upto 96khz same mics/monitors/instruments, and i defo defo could notice one hell of a difference

That doesn't necessarily prove anything about sample rates.

"Better-sounding kit than behringer exists shocker. Film at 11".

[narcoman]

ha!

96 does sound better than 48 - but I'd defy anyone to hear it on anything but optimal monitoring in a high end room. The rest is confirmation bias.... Jack outlines the real advantages.

[Steve Hill]

id defo have to disagree on this having upgraded from behringer ada8000's that go upto 48khz to motu 8 pre's to go with my motu 828 mk3 that go upto 96khz same mics/monitors/instruments, and i defo defo could notice one hell of a difference, but that might just be me

You're confusing different topics: I noticed an improivement when I upgraded from cheapo Motus such as the 24i/o or the 2048 to Motu 192s. I noticed again when I upgraded to Prism converters. I didn't change the sample rates I was using. I was just experiencing the benefits of better converters.

I am unpersuaded that there is any need at all to record beyond 96k, ever, for any purpose.

[RegressiveRock]

"Film at 11".

Gear porn at 12! (is probably the key motivation for 192)!

[Edd M]

I was under the understanding that although we don't necessarily hear above 20kHz, we can sense it.

So information up there could help add extra information psychologically, whether it be direction, detail or space recognition.

The opposition does have a point in that if you can't prove it and it doesn't make a difference to you, then why use it? Of course that's true, when has production ever been about doing something because you should?

I tend to agree with those who aim at 48kHz, due to it allowing the data collected in the audible range to be more precise, and perhaps up at 192kHz it'll be the most precise representation of the analogue source.

Realistically, surely having the most accurate representation of the original sound is the most important thing, so we should embrace higher sample rates!

[humandrums]

cut me some slack guys im a drummer!!! i just remember when i was asking on here a while back when i first got the behringers that using them digitally at 48khz i wouldnt really hear an improvement by getting better pre's but that better mics was the way to go (i couldnt afford new pres and new mics at the same time), the setup i have now when i recorded my kit using the same mics in the same room at 96khz again digitally i dunno fair enough if its the better unit than the behringer honest injun i never changed nuffin else in the setup guv, it just sounded so much richer and a much fuller sound with really clean crisp top end in it

[narcoman]

I was under the understanding that although we don't necessarily hear above 20kHz, we can sense it.

So information up there could help add extra information psychologically, whether it be direction, detail or space recognition.

The opposition does have a point in that if you can't prove it and it doesn't make a difference to you, then why use it? Of course that's true, when has production ever been about doing something because you should?

I tend to agree with those who aim at 48kHz, due to it allowing the data collected in the audible range to be more precise, and perhaps up at 192kHz it'll be the most precise representation of the analogue source.

Realistically, surely having the most accurate representation of the original sound is the most important thing, so we should embrace higher sample rates!

Depends on what your view of super harmonic information is...... Most analogue electronics generate a lot of noise up there (look on a scope) so all you're doing is capturing stuff you don't want anyway ...

[thiswayup2010]

You wonderful people, thanks for the input. I've always thought 192kHz was a bit extreme....pointlessly extreme, that is.

I think the moral of the interface story is.....have decent preamps!

[Andi]

Depends on what your view of super harmonic information is...... Most analogue electronics generate a lot of noise up there (look on a scope) so all you're doing is capturing stuff you don't want anyway ...

Easy - just stick a 4th order LPF on it at around 17K!

[Wonks]

cut me some slack guys im a drummer!!!

OK, I'll type slowly.

If you still have the behringer, then record the same sound using the same mic through both A/Ds at the same sample rate. That will tell you whether the better sound is due to the pre-amps or the sampling rate.

[chris...]

or the A/D conversion

[MarkOne]

or the A/D conversion

Or indeed the analogue stuff before/after the converters

Or the power supplies

[ef37a]

This will rumble on till the end of time I guess but I would like to make two observations.

Very few mics actually GET to 20kHz. A few are 3-4dB down and falling at about 23khz and a VERY few go up to 30-40kHz and I cannot think of any other source that goes past 18kHzish? Certainly not MY area of interest, guitar amp/speakers!

Secondly I have seen specs' of converters that whilst running at 96kHz and above still limit the top end to 25kHz or so!

Dave. (Oh! and a MOTU sounds better than a Berry? VERY Cilla Black that!)

[Wonks]

or the A/D conversion

Or indeed the analogue stuff before/after the converters

Or the power supplies

I should have just said (and I meant) interface.

[jmedigital]

I was under the understanding that although we don't necessarily hear above 20kHz, we can sense it.

So information up there could help add extra information psychologically, whether it be direction, detail or space recognition.

The opposition does have a point in that if you can't prove it and it doesn't make a difference to you, then why use it? Of course that's true, when has production ever been about doing something because you should?

I tend to agree with those who aim at 48kHz, due to it allowing the data collected in the audible range to be more precise, and perhaps up at 192kHz it'll be the most precise representation of the analogue source.

Realistically, surely having the most accurate representation of the original sound is the most important thing, so we should embrace higher sample rates!

Hold me back!

This is not how it works. Higher sampling rates simply allow you to record higher frequencies.

An 18kHz signal is not sampled any more accurately at 192kHz as it is at 44.1kHz.

[ef37a]

Edd.M
Even IF we could "sense" above 20k, I repeat, where are you going to get the frequencies from?

You would have to restrict your microphone choice VERY severly and for sure you could use NO dynamics or ribbons and even for that handful of 30kHz + mics, cable capacitance would start to come into play (find the "mic splitter" thread and read Bill W's paper).

Audio is the only engineering discipline (I use the term loosely in this context!) where some designers go for crazy bandwidths. You don't design a ships' anti-roll servos to work at 1000Hz!

Dave.

[humandrums]

so should i record everything at 22khz to save disk space?? or doesnt it work that way??? why is there the option to be able to record at 96khz if its pointless??

[RegressiveRock]

so should i record everything at 22khz to save disk space?? or doesnt it work that way??? why is there the option to be able to record at 96khz if its pointless??

No