Do you guys record 24bit 44.1 or 96khz

[Hjlphilp]

I'll tell you what though fellas... concerning certain plugins, A/D converters, the equipment itself, and even mp3 conversion, you will thank yourself for doing this at 96k

I have recently made the switch because of my apollo twin... but what's so cool is the fact that because it's a high resolution interface, I can truly hear and feel the different because of how that machine was made. I got into a huge debate with my buddy and he swore up and down that there was no difference... i'm like no way!

[Bob Bickerton]

I’m reminded of an article, possibly in this magazine, that was published when recording at 96kHz was new. It talked about how several top engineers in London in a well treated studio with optimum monitoring did a blind A/B test and decided after much discussion and debate that a difference was perceivable..............

Go figure........

Bob

[ef37a]

The question has been posed before here I think (and other places) and the result seems much the same. 44.1kHz almost always and 48kHz for video and 24 bits is a no-brainer!

It makes obvious sense to me however that a large, "important" recording event will be captured at the best current technology can achieve without going overboard? Hugh and others have in the past pointed out that there are certain circumstances where the lower sample rates might have a problem. Makes sense therefore to avoid them when it really matters. The "space" argument hardly applies these days!*

96kHz is used in equipment such as monitors mainly I assume because it halves latency and in any case has no downside?

*Just looked up SSD prices. 240G about 20p a gig! Who'ed a thought it just ten years ago?
Dave.

[CS70]

I'll tell you what though fellas... concerning certain plugins, A/D converters, the equipment itself, and even mp3 conversion, you will thank yourself for doing this at 96k

I have recently made the switch because of my apollo twin... but what's so cool is the fact that because it's a high resolution interface, I can truly hear and feel the different because of how that machine was made. I got into a huge debate with my buddy and he swore up and down that there was no difference... i'm like no way!

I havent tried myself but I seem to remember that the consensus here in the last thread on the subject is that it 96KHz can make a perceptible difference in certain circumstances for high frequencies (like cymbals captured by drum overheads) and with most A/D converters.

I'll record some drums next week for the next songs and I'll do 96Khz for the first time.

In any case, if your ears tell you something go for it - but be careful that self-delusion is a common trap in these things. Arrange your friend so that you make a true blind A/B comparison and you'll find out soon enough.

[Hjlphilp]

I’m reminded of an article, possibly in this magazine, that was published when recording at 96kHz was new. It talked about how several top engineers in London in a well treated studio with optimum monitoring did a blind A/B test and decided after much discussion and debate that a difference was perceivable..............

Go figure........

Bob

And it makes sense that it would be knowing that the ear constructs things based off harmonics. So those frequencies outside of our range are still impacting the stuff we can hear.

And 96k might actually cause MORE latency issues. I actually have to make music at 44.1 and bounce it at 96. Which is dope because there's no resampling involved... all I did was switch between to the two rates.

M.2 for the win anyone???

[ef37a]

I know I am just an old bottle swapper but...surely recording at 44.1kHz and converting to 96k is completely pointless? Unless there is a plug in or other thing that needs 96kHz?

And I get the cymbals thing, high level HF but the idea of "out band ear constructs" is an old one, often trotted out but, AFAIK has never been seen to produce data better than that from the blunt end of a pony?

Dave.

[Hugh Robjohns]

And 96k might actually cause MORE latency issues.

Only if your computer is too slow to cope... which isn't very likely.

Doubling the sample rate halves the A-D and D-A converter processing times, which is the bulk of the round-trip latency time.

I actually have to make music at 44.1 and bounce it at 96. Which is dope because there's no resampling involved...

Of course there's re-sampling involved! If you're converting from 44.1 to 96, the signal is resampled to the new, faster, sample rate -- but the audio bandwidth is not extended as that was set by your original sample rate.

...all I did was switch between to the two rates.

It's amazing how easily modern DAWs can perform complex sample-rate conversions in real time (or faster)!

H

[Hugh Robjohns]

...surely recording at 44.1kHz and converting to 96k is completely pointless?

Yes and no... complex non-linear plugins can be computed with far greater accuracy at elevated sample rates -- so there is a point to processing base-rate material at 96kHz or whatever. But there is no significant practical benefit in upsampling a 44.1k recording on its own.

Most of UAD's modelled plugins run internally at 192kHz, for example. However, their plugin itself performs the necessary upsampling and downsampling internally if the source is a lower sample rate. So there is no benefit at all in upsampling a 44.1k source separately, in that context. The same applies to oversampling D-A converters: there is a benefit to the reconstruction filter accuracy (and jitter reduction sometimes) by upsampling, but it happens automatically in the converter. It makes no difference if you feed it 44.1 or separately upsampled 96k audio!

...the idea of "out band ear constructs" is an old one, often trotted out but, AFAIK has never been seen to produce data better than that from the blunt end of a pony?

Much of the workings of the ear are still not properly understood. It appears there are separate processes and mechanisms involved in discerning spectral content (which really doesn't go beyond 20kHz even on a really good day), and temporal resolution used in identifying source locations which could be as precise as 10us resolution (which implies an equivalent frequency response over 100kHz -- which is where the confusion comes in).

H

[James Perrett]

...temporal resolution used in identifying source locations which could be as precise as 10us resolution (which implies an equivalent frequency response over 100kHz -- which is where the confusion comes in).

There is a common fallacy that the temporal resolution is determined by the sample rate. The relationship is actually far more complex (as I understand it) and 44.1kHz at 16 bits is sufficient to resolve these timing differences (given that the filtering performs appropriately).

[Hjlphilp]

Of course there's re-sampling involved! If you're converting from 44.1 to 96, the signal is resampled to the new, faster, sample rate -- but the audio bandwidth is not extended as that was set by your original sample rate.

H

Right I should've made myself clear, I was describing what the guy after you was talking about which was upsampling source material. But since I never bounced it I can get that illustrious 96k and still flow without my comp slowing me down

Now you guys are schooling me on this temporal resolution. I took psychoacoustics which is why I figured the ear wanting things to make sense was coming into play, and I still think that it is... uad's plugins gave me a clue as to why it sounds better.

Now keep in mind... ive had these monitors for over 5 years. And when I tell you I heard the difference i'm like WHOOOOOA THINGS WILL NEVER BE THE SAME!

I was debating recording at 192 for the sake of uad... but was like "latency doesn't matter in the mastering phase so..."

(Btw, did you know there are machines with audio cards producing 32 bit 384k audio... joint is dumb right? And that's for gamers not even audiophiles)

[Bob Bickerton]

Didn’t I hear the first law of psychoacoustics states that the perceived sonic improvement attributed to a new piece of gear is directly proportional to the cost of that new piece of gear, provided the listener is aware of the cost of the gear and paid for it him/herself?

Bob

[CS70]

Didn’t I hear the first law of psychoacoustics states that the perceived sonic improvement attributed to a new piece of gear is directly proportional to the cost of that new piece of gear, provided the listener is aware of the cost of the gear and paid for it him/herself?

Bob

My research indicates that the law applies also if the listener didn't pay directly for it, maybe we should join forces to write the paper!

[CS70]

(Btw, did you know there are machines with audio cards producing 32 bit 384k audio... joint is dumb right? And that's for gamers not even audiophiles)

32 bits allow to represent over 4.29 billion numbers, that translates to a dynamic range of about 192dB.

Handy if you are a dog, I suppose.

[MarkOne]

(Btw, did you know there are machines with audio cards producing 32 bit 384k audio... joint is dumb right? And that's for gamers not even audiophiles)

32 bits allow to represent over 4.29 billion numbers, that translates to a dynamic range of about 192dB.

Handy if you are a dog, I suppose.

More handy for that soundscape that includes the sounds of a pile driver, demolition hammer at close range together with an accurate sonic capture of a group of fluttering buterflies! - I only mention it because the aforementioned pile driver, demolition hammer combo are this afternoon's sonic backdrop here

[TheMiller]

The only additional information captured by 96k is in the bandwidth above 20kHz.

Personally, I can only hear to 18kHz - but maybe some of you who are younger than my 47 years can still converse with bats, and also that you find this significant enough to use 96kHz on fancy recordings.

There are however two reasons to use a higher sampling rate:

1) more data means lower latency - I know it sounds crazy, but it turns out to be true.

2) any processing that adds distortion will add non-harmonic aliasing which translates as a harsh high end - and 96kHz reduces this problem considerably. This is NOT an issue if the plug-in upsamples internally.

3) You like big files and you cannot lie.