96k vs 192k Worth It?

Oooo! Matey! You used the 'R' word. No matter the sample rate nor word length the RESOLUTION is still one bit = approx 6dB.



Dave.
 
Lol, easy Dave. I just mean number of units used to measure something, or the smallest amount by which something can be adjusted.

We know we can feel bass that we cannot hear....

......coming out of my $2 earbuds.

See, this works.
We can debate whether there's a measurable difference, then debate whether or not it's perceivable to any given person,
then, perhaps most importantly at HR, debate whether the rest of the equipment is even capable of articulating that difference, or if other weaknesses or failings are going to mask it.
96k and up just seems too much like quality of cable, to me. Get better monitors. ;)
 
Could be, after all, the number of samples doubled and the bit depth went from 16 to 24. Could also have been just a much better quality pre in the Presonus unit : )

The bit depth relates to the headroom, the sample rate relates to the highest frequency that can be captured.

Moving from 16 bit to 24 bit, moves the digital florr from -96dbfs to -144dbfs. Moving from 48k to 96k changes the nyquist frequency from 24khz to 48khz. However, you certainly won't be hearing much over 24khz and will never hear anything above 48khz, so going over 96k sample rate is just a waste.
 
I'll drop in a warning- a guy who gets cited a lot is Oohashi, who did some experiments which supposedly demonstrated that brain wave differences could be measured when bandwidth was extended past the canonical 20kHz. Unfortunately, not only were his controls questionable, several attempts by other groups to replicate his work showed null results. That doesn't stop hucksters with something to sell from citing it anyway...

Reality: if you're a musician, you're unlikely to be able to even hear 20kHz. So anything beyond a 44.1 is for bragging or marketing, not actual sound.
Reality: Very, very few mikes have frequency responses that go into the ultrasonic range being discussed. The two that I've reviewed which do are unsuitable for recording anyway because of their noise floor (engineering = tradeoffs).
Reality: There's no good experiment out there, none, zilch, nada, which supports audibility of increased sampling rates.

One other comment- the reason for using 30 ips for analog tape is extended headroom, lower distortion, and lower noise floor. 44.1 vs 48 vs 96 vs 192 doesn't address that at all. Analog tape is a bad, errrr, analogy.
 
I have one use for my TASCAM DR100Mk3 that absolutely has to use the 192KHz sampling rate .. recording the ultrasonic echolocation calls of bats!

Over the years I've built several 'heterodyne' bat detectors which - by mixing the bat calls with an ultrasonic oscillator - can bring those calls within 'normal ' audio ranges. It works, but it is 'faking' things a bit....and does not capture the subtleties of the original source 'sounds'

Until recently, portable audio recorders with 192KHz sampling rates tended to be really expensive - mostly bespoke machines produced especially for this task. There was the Korg MR1 of course, which had a useful ultrasonic range in DSD mode, but it was still expensive... and by some accounts, not that good?...

Now that machines - like Tascam's DR100Mk3 - have 192KHz sampling and an 80KHz frequency response, it's all becoming a bit cheaper (not a whole lot cheaper as yet, sadly! :( )

So, by adapting cheap Knowles MEMS mics (which have a useful ultrasonic response) to work with phantom power - and a 192KHz recorder - it's now possible to capture the 'real' sound of the bat echolocation calls (although you do still have to 'time stretch' the recordings to actually hear anything of course!).

OK, all a bit specialised of course, but it's not something you can do without 192KHz sampling -- so it does have some validity :)

(Some notes on my home made set up -- and a short sample recording here: http://www.jp137.com/lts/MEMS.bat.experiments.pdf )
 
Am I the only one that looks at that and sees someone needing dental work?

And, Ok, recording bats is probably a valid use case, but FOR HUMANS, it's marketing.
 
I'll drop in a warning- a guy who gets cited a lot is Oohashi, who did some experiments which supposedly demonstrated that brain wave differences could be measured when bandwidth was extended past the canonical 20kHz. Unfortunately, not only were his controls questionable, several attempts by other groups to replicate his work showed null results. That doesn't stop hucksters with something to sell from citing it anyway...

Reality: if you're a musician, you're unlikely to be able to even hear 20kHz. So anything beyond a 44.1 is for bragging or marketing, not actual sound.
Reality: Very, very few mikes have frequency responses that go into the ultrasonic range being discussed. The two that I've reviewed which do are unsuitable for recording anyway because of their noise floor (engineering = tradeoffs).
Reality: There's no good experiment out there, none, zilch, nada, which supports audibility of increased sampling rates.

One other comment- the reason for using 30 ips for analog tape is extended headroom, lower distortion, and lower noise floor. 44.1 vs 48 vs 96 vs 192 doesn't address that at all. Analog tape is a bad, errrr, analogy.

Agree with all of that friend except! I think you will accept that for a given replay gap, upper frequency limit is increased for higher tape speed?

I have already said that "no analogy is perfect".

Dave.
 
I agree with ecc83. With analog tape you lose low end at 30ips but a decent machine will achieve 28kHz, but you’ll be 3db down at 60Hz. A well aligned machine at 15ips will be almost flat from 40Hz to 20kHz, it’ll be about 2db down at 25kHz. I can’t hear this, but test equipment proves it. 30ips was used to reduce the noise floor without noise reduction. It sounded much more open, even to my aged, abused ears, but the head bump was at about 88Hz. The head bump on a 15ips was usually around 40Hz. Head bump is a result of the contour of the head and the way the tape rides past the heads. At least that’s my understanding of it. Studers had wider profile heads than say an MCI. The MCI at 15ips kicked ass in the low freqs.

There is an analogy between analog and digital. The first Sony digital multitracks were 16 bit, most people used 48kHz. These machines ran at approximately 30ips. When Sony went 24 bit they had to speed up the tape to 45ips to get all those extra bits. Thankfully ProTools became the standard and tape cost became a non-issue.

I personally use 48kHz at 24 bit and am very happy with the sound. The only advantage I get from 96kHz is lower latency on virtual instruments. I use a hybrid analog/digital system. I have an analog mixer and many hardware keyboards to eliminate any latency. I use virtual synths very sparingly. Unfortunately I bought a Mackie Onyx 1640 and they stopped writing drivers for it (Bastards!). Having the mixer as an interface was a brilliant idea, too bad they consider it a “legacy” product. So if I want to update my MacOS, I’ll have to build 16 direct channel out cables and buy another 16 channel interface. The stuff works for me now so I’m not updating!
 
Back
Top