96k vs 192k Worth It?
- Thread starter cecerre
- Start date
miroslav
Cosmic Cowboy
What's the rest of your rig like...the entire signal chain, from source to final monitoring...will it be able to really make use of whatever 192 kHz gives you?
Also...do you think you will really hear a difference?
If you can say YES without a doubt, then let your budget be the final factor, but you will probably do better to spend that money on some other part of your recording rig, IMO.
Heck....most home rec/project studio people are recording at 44.1 or 48 kHz, never mind 96 or even 192 kHz.
I can record at 96 kHz...and I've been on 48 for quite a few years now, and the rest of my studio rig is pretty decent, and TBH, I don't feel like I'm missing anything.
Also...do you think you will really hear a difference?
If you can say YES without a doubt, then let your budget be the final factor, but you will probably do better to spend that money on some other part of your recording rig, IMO.
Heck....most home rec/project studio people are recording at 44.1 or 48 kHz, never mind 96 or even 192 kHz.
I can record at 96 kHz...and I've been on 48 for quite a few years now, and the rest of my studio rig is pretty decent, and TBH, I don't feel like I'm missing anything.
keith.rogers
Well-known member
This topic has been flogged pretty well and not that long ago here.
I can't imagine the hardware that would be required to actually mix that volume of bits at any significant track count. That would be a ridiculous a lot of money for something I'll never hear!
I can't imagine the hardware that would be required to actually mix that volume of bits at any significant track count. That would be a ridiculous a lot of money for something I'll never hear!
Thank you both for your response. I can say that there is a TON more headroom recording at the 96k rate. I used an Alesis unit at 48k for a long time and when I moved up to 96k, it was very noticeable. The headroom in tracking in my opinion was blatantly obvious and the lows were back to analog days. For me, it was a huge step up. I am hoping someone has made the leap from 96k to 192k and can weigh in on their experiences and opinion of the sonic performance overall compared to the 96k sample rate.
It’s another 300.00 more for a Focusrite capable of 192k sample rate at 24bit depth. While it’s not a deal breaker for me, I’d like to get two units so I can handle 16 tracks at a time, hence the concern.
Thanks!
It’s another 300.00 more for a Focusrite capable of 192k sample rate at 24bit depth. While it’s not a deal breaker for me, I’d like to get two units so I can handle 16 tracks at a time, hence the concern.
Thanks!
Steenamaroo
...
There's no relationship between sample rate and headroom.
Do you mean bit depth?
Do you mean bit depth?
miroslav
Cosmic Cowboy
You're one of of the very few who noticed a "TON" of difference...though not understanding how increasing the sample rate increased your headroom...and that substantially?
Maybe it's just the differences in the analog filtering between the Alesis unit and whatever you moved to, and nothing much to do with the sample rate.
I'm also not understanding why it would affect your lows.
Not just doubting you...but honestly, I think you would have to do a more honest comparison of your entire chain when you used the 48 kHz and moved to 96 kHz.
IOW...people often think they are comparing apples to apples...when in fact there's multiple variables involved.
That said, even if you heard a "TON" of difference between those two...I don't think you will hear the same thing going to 192.
In all my comparisons...the differences were very subtle, at most.
keith.rogers
Well-known member
Steenamaroo
...
That'll be it. Headroom relates to bit depth.
Frequency range relates to sample rate.
I'd be surprised if 1 in 100 people here could blind test 48 against 96 with all else being equal.
Beyond 96? Forget it.
snow lizard
Dedicated Slacker
You'd take a substantial hit in processing power and storage space if you want to run higher sample rates with higher track counts. Some of the prosumer level stuff can sound better at a specific sample rate, but that can have more to do with the design of the converters than the sample rate itself.
192 kHz is more of a marketing idea. A lot of pro level stuff gets done at 48.
192 kHz is more of a marketing idea. A lot of pro level stuff gets done at 48.
If any of the Annyloggy guys look in? I wonder whether any of those run at 30ips? Directly equivalent to going from 40's to 96k and eats resource the same way as well.
Back in the day, a well setup 15ips A77 was flat to 25kHz or so and the only peeps running 30ips were the Dolby A naysayers. Fork, a decent OR machine is easily good enough for most jobs at 7.5ips, certainly FM radio, but editing is more of a pain.
Dave.
Back in the day, a well setup 15ips A77 was flat to 25kHz or so and the only peeps running 30ips were the Dolby A naysayers. Fork, a decent OR machine is easily good enough for most jobs at 7.5ips, certainly FM radio, but editing is more of a pain.
Dave.
snow lizard
Dedicated Slacker
I don't think any comparison to tape is all that useful when talking about sample rates. Apples to buffalo. I think a lot of people used 15 ips for the way the head bump shapes the low end. Digital audio doesn't do that.
There is an argument that going up to around 60 kHz sampling makes it much easier to build the low pass filter required by Nyquist. Beyond that you're getting into the bat range.
There is an argument that going up to around 60 kHz sampling makes it much easier to build the low pass filter required by Nyquist. Beyond that you're getting into the bat range.
miroslav
Cosmic Cowboy
Well! I don't like to disagree with my youngers and betters but! I think tape speed IS comparable to sample rate!
For any given replay headgap, doubling speed doubles the HF cut off point but in the same way as sample rates above 48kHz 'waste' resources with modern technology, tape heads got so bloody good that 7.5ips was easily good enough.
The bass "woodless" as they were dubbed here by our sadly missed Angus MacKenzie are an artefact caused by the head profile and it can only be optimized for one speed AFAIK.
There are other benefits to 15ips over 7.5 of course, bit better HF squash and 3dB noise improvement and of course going to 30ips gets these better again but at great cost.
My point being that LIKE >15ips >44/48kHz uses more memory and CPU grunt for almost no demonstrable benefit.
Analogies are never perfect but HANDY sometimes?
Dave.
For any given replay headgap, doubling speed doubles the HF cut off point but in the same way as sample rates above 48kHz 'waste' resources with modern technology, tape heads got so bloody good that 7.5ips was easily good enough.
The bass "woodless" as they were dubbed here by our sadly missed Angus MacKenzie are an artefact caused by the head profile and it can only be optimized for one speed AFAIK.
There are other benefits to 15ips over 7.5 of course, bit better HF squash and 3dB noise improvement and of course going to 30ips gets these better again but at great cost.
My point being that LIKE >15ips >44/48kHz uses more memory and CPU grunt for almost no demonstrable benefit.
Analogies are never perfect but HANDY sometimes?
Dave.
gecko zzed
Grumpy Mod
I am very fond of analogies, and there may be a case to compare digital to tape, but the factors influencing differences between 7.5, 15 and 30ips are themselves quite complex, and probably don't shed much light on the differences between 48, 96 and 192.
A more revealing analogy is one implied by Miro: will my Toyota Corolla go faster if I swap its 4 cylinder engine for an F1 engine? The answer to that is 'no'. And the reason is that the whole of the car -- transmission, suspension, tyres and wheels and everything else -- needs also to be changed to be capable of dealing with the power.
And even were you to do that, you would not be able to go much faster because the roads travelled won't allow it.
A more revealing analogy is one implied by Miro: will my Toyota Corolla go faster if I swap its 4 cylinder engine for an F1 engine? The answer to that is 'no'. And the reason is that the whole of the car -- transmission, suspension, tyres and wheels and everything else -- needs also to be changed to be capable of dealing with the power.
And even were you to do that, you would not be able to go much faster because the roads travelled won't allow it.
snow lizard
Dedicated Slacker
ecc83 said:
Fair enough. My point being that 30 IPS is more subjective. 192 kHz sampling is more surreal. There aren't many sound transducers around handling 80 kHz.
miroslav
Cosmic Cowboy
The other thing about tape speed...specifically 15 ips and 30 ips...very often speed selection is not done for the similar kind of reasons as you might choose 48 kHz over 96 kHz.
IOW...there's often a production decision to use one tape speed over the other, as the impact on the final sound of each speed choice can be used creatively, and not purely analytically, as you would with digital sampling choices.
That said...I've heard of some cases where people favor older, 16 bit processors to more modern 24 bit ones with higher sampling rates, because the grainy quality they add may be desirable, in some cases, to some people.
It's not unusual that even to this day, people try to find ways to compare analog tape to digital recording. Not to mention, that digital as advanced as it has become, can't shed it's desire to BE compared to analog.
I'm looking at the back cover of my current Tape Op, and there's a full page ad for UA Apollo interfaces...and the caption reads:
Superior conversion meets the tone, feel and realtime workflow of classic analog recording.
IOW...there's often a production decision to use one tape speed over the other, as the impact on the final sound of each speed choice can be used creatively, and not purely analytically, as you would with digital sampling choices.
That said...I've heard of some cases where people favor older, 16 bit processors to more modern 24 bit ones with higher sampling rates, because the grainy quality they add may be desirable, in some cases, to some people.
It's not unusual that even to this day, people try to find ways to compare analog tape to digital recording. Not to mention, that digital as advanced as it has become, can't shed it's desire to BE compared to analog.
I'm looking at the back cover of my current Tape Op, and there's a full page ad for UA Apollo interfaces...and the caption reads:
Superior conversion meets the tone, feel and realtime workflow of classic analog recording.
The other thing about tape speed...specifically 15 ips and 30 ips...very often speed selection is not done for the similar kind of reasons as you might choose 48 kHz over 96 kHz.
IOW...there's often a production decision to use one tape speed over the other, as the impact on the final sound of each speed choice can be used creatively, and not purely analytically, as you would with digital sampling choices.
That said...I've heard of some cases where people favor older, 16 bit processors to more modern 24 bit ones with higher sampling rates, because the grainy quality they add may be desirable, in some cases, to some people.
It's not unusual that even to this day, people try to find ways to compare analog tape to digital recording. Not to mention, that digital as advanced as it has become, can't shed it's desire to BE compared to analog.
I'm looking at the back cover of my current Tape Op, and there's a full page ad for UA Apollo interfaces...and the caption reads:
Superior conversion meets the tone, feel and realtime workflow of classic analog recording.
Agreed, but I was limiting my analogy (as indeed all analogies are!) to the 'resource/cost/perceived benefit' and 30ips ON THAT BASIS does compare. For sure very few mics get TO 20kHz leave alone beyond it but that bald fact still does not stop pre amp mnfct making products that go into the low MW band! (and PLEASE anybody, DON'T start on about bloody 'phase'!).
Bolting a more powerful engine into my Proton WOULD make it go faster if the rev limit was the same since final speed is determined by HP/drag and the latter goes up as the cube (iirc) of velocity? So, you would need a lot more power to get a small increase in final speed. The Law of Diminishing Returns AGAIN! Of course the transmission would blow very soon, clutch probably (stick shifter) .
"Production reasons"? aka 'Poncey producers' who want "Total Transparency' and know FA about the technicals.
Our view of the whole universe is an analogy filtered through our pitifully limited senses and capability to understand.
Dave.
Steenamaroo
...
I don't think any analogy is really going to work unless the limits of human perception are in the equation.
If I make a guitar tuner that does a thousandth of a cent per turn, or someone makes it that there are several thousand steps available in midi velocity, that compares.
Someone's raised the resolution of something to the point where the ability to perceive the difference is debatable.
There's bound to be a direct comparison in photography - A certain resolution after which people argue there's just no point?
Is it better? I don't know. Can you hear/see the difference?
One guy says yes, another guy says no.
One guy can measure it.
The other guy can't hear it.
No one really 'wins', I don't think.
I like the car analogy in that, at some point, other parts of the chain render the main change pointless.
No matter what someone thinks on the sample rate debate (or converter debate), it's always worth pointing out that the environment, monitors, and ears,
are usually weak links that require attention before worrying about the real subtleties.
If I make a guitar tuner that does a thousandth of a cent per turn, or someone makes it that there are several thousand steps available in midi velocity, that compares.
Someone's raised the resolution of something to the point where the ability to perceive the difference is debatable.
There's bound to be a direct comparison in photography - A certain resolution after which people argue there's just no point?
Is it better? I don't know. Can you hear/see the difference?
One guy says yes, another guy says no.
One guy can measure it.
The other guy can't hear it.
No one really 'wins', I don't think.
I like the car analogy in that, at some point, other parts of the chain render the main change pointless.
No matter what someone thinks on the sample rate debate (or converter debate), it's always worth pointing out that the environment, monitors, and ears,
are usually weak links that require attention before worrying about the real subtleties.
keith.rogers
Well-known member
The other aspect of any analog-y with tape is that higher tape speeds (and wider tape!) delayed/reduced the inevitable, human-audible [negative] impacts of those mixdowns/bounces and copies, and even the initial record/playback degration to master-cutting. Not an issue at all with digital.
But, I honestly don't know what those ultra-high frequencies might do to the human body in some intangible way in a real world, e.g., symphony orchestra in a great hall. We know we can feel bass that we cannot hear, so just because we can't hear it doesn't mean it's not wiggling something somewhere I suppose. But, here, we're talking about audio that is captured imperfectly to begin with using devices that measurably cannot get all of it, and then played back on other devices that don't do much better. Bass (or a dog whistle) I cannot hear is not even coming out of my $2 earbuds...
TBH, I'm actually a little more bothered by the thought that 192kHz sampling is introducing bits into the data that we know nothing in a studio is actually capable of capturing, so whatever is there is either pure noise or just a bunch of zeros taking up disk space, which then gets carried though the process, ultimately to be introduced to some D/A algorithm, where it might actually change things, but not in any way that could be reflective of reality.
Really, I think this is largely a "we've got this and it's 2x better than studio X so you definitely want to give us your business" thing. There's not one bit of value in it for home recording. My $.02. Now, back to my popcorn..
But, I honestly don't know what those ultra-high frequencies might do to the human body in some intangible way in a real world, e.g., symphony orchestra in a great hall. We know we can feel bass that we cannot hear, so just because we can't hear it doesn't mean it's not wiggling something somewhere I suppose. But, here, we're talking about audio that is captured imperfectly to begin with using devices that measurably cannot get all of it, and then played back on other devices that don't do much better. Bass (or a dog whistle) I cannot hear is not even coming out of my $2 earbuds...
TBH, I'm actually a little more bothered by the thought that 192kHz sampling is introducing bits into the data that we know nothing in a studio is actually capable of capturing, so whatever is there is either pure noise or just a bunch of zeros taking up disk space, which then gets carried though the process, ultimately to be introduced to some D/A algorithm, where it might actually change things, but not in any way that could be reflective of reality.
Really, I think this is largely a "we've got this and it's 2x better than studio X so you definitely want to give us your business" thing. There's not one bit of value in it for home recording. My $.02. Now, back to my popcorn..