What is the preferred recording format, WAV or AIFF?

MIDI Life Crisis · Post by **MIDI Life Crisis** » Sat Feb 20, 2016 4:07 pm

And human hearing is not going to improve substantially in the next million years or so. Once you reach "moderately" over that limit, which we surely have, then higher resolution is only a matter of graphic, scientific, or other technical analysis. That means exactly squat (unless there's a sound to go with that...

https://www.youtube.com/watch?v=VPIP9KXdmO0

mhschmieder · Post by **mhschmieder** » Sat Feb 20, 2016 4:46 pm

The more important factor is phase coherence and whatever helps achieve that.

In terms of sample rates, don't forget that humans experience music in the time domain more than in the frequency domain. So, is it possible that more continuity (shorter time span between samples) leads to more relaxing listening experiences? Certainly higher sample rates help with long decays on soloed instruments (e.g. acoustic guitar) or in complex highlighted cymbal work (e.g.).

Super-high rates (e.g. 192 kHz) sometimes are needed for frame-syncing in tightly-edited film soundtracks.

Anyway, my point is that it isn't all about the average human audible frequency range. And of course it has also always been about filter quality (decreasingly an issue) and for synths and modeling stuff can be about anti-aliasing.

The original question is BWV vs. AIFF and that's already settled, as the time stamps are critical but people who work with loops (not me!) might need AIFF for GarageBand compatibility.

It is awfully hard to avoid filter issues at 44.1 kHz so I find it safer to work at 48 kHz, and this also sets me up nicely for having every project film-ready if necessary.

I have a lot of projects at 96 kHz though, as my original digital device was standalone (I was late to computers) and only supported 16-bit but went up to 96 kHz so I needed that in order to capture the delicate details of featured long-decay type acoustic instruments.

I have thought about down-converting all of my 96k projects to 48k, but even though mostly it is only in my legacy projects, I'd still say that 90% of my projects are 96k just because they were almost all done before I bought a computer.

Prime Mover · Post by **Prime Mover** » Sat Feb 20, 2016 4:52 pm

I do quite a bit of folly and sound design type of stuff these days, lots of requirement for slowdowns. Recording at 96kHz preserves a bit more of the top octave if I slow something down to 50%. I know, mics aren't really supposed to get frequencies much higher than 20Hz, but I suspect they do get some, and it makes a difference.

I recently created a really gross sound of a blood splatter. I recorded some water being flicked on a surface at 96kHz. I slowed it down to about 65%, and it just felt really juicy and meaty, without losing the high end.

Aside from that, I stick with 44.1k or 48k, mostly 48.

MIDI Life Crisis · Post by **MIDI Life Crisis** » Sat Feb 20, 2016 4:55 pm

OK, I definitely buy the slow down as well as the synth aliasing, but those are special circumstances. For most use, 48k is just peachy fine. If a special reason to go higher arises, I'll use it, but for general archival and production stuff, it's 48k or bust.

stubbsonic · Post by **stubbsonic** » Sat Feb 20, 2016 7:49 pm

I'm not saying it is likely, but it is conceivable that down the road things like cochlear implants, nano-tech, tactile sensors or other weird brain science could expand our sensory experience. (I.e., like someone invents some sonic equivalent of "night vision glasses" for ears).

MIDI Life Crisis · Post by **MIDI Life Crisis** » Sat Feb 20, 2016 7:57 pm

stubbsonic wrote:I'm not saying it is likely, but it is conceivable that down the road things like cochlear implants, nano-tech, tactile sensors or other weird brain science could expand our sensory experience. (I.e., like someone invents some sonic equivalent of "night vision glasses" for ears).

This may be true, and then again everything above 20k might just sound like a lot of crap. You never know...

mikehalloran · Post by **mikehalloran** » Sun Feb 21, 2016 9:20 am

My 2¢

I can no longer hear above 15kHz but can hear the difference between 48k and 44.1, especially when compared to an analog original. Is in phase distortion caused by aliasing? Perhaps... Is it the actual top end? Maybe...

Does anyone hear a real difference between 48k and 96k in blind listening tests? Consistently? How about recordings done in true stereo?

There is evidence that frequencies in the 22k-25k range are used by the brain to locate images in the stereo field. In an age where most stereo is really processed, panned mono (including most classical recordings if you think about it), this falls into the 'interesting curiosity' category.

I have an early 2-mic stereo record from the late '50s that states you can see frequencies from 16–25k frequincies on the record looking through a microscope. The CD is ok but, listening through headphones, the vinyl kicks butt. Master that one at 48k, 96k and with the best equipment at 44.1 and let's see who can hear the difference.

My half-track mastering deck has a gentle roll-off beginning at 25kHz — or did when I last calibrated it. It's been in my basement since 1999. I have a box of tapes to bake and transfer someday which is why I still have it. Someday, I'll hire someone to help me set it up and do that. Only the tracks I recorded in true stereo will get transferred at 96k; everything else will be 48k.

I was involved in some early experiments with 192k with an ECM artist. The results were impressive but the conditions were so controlled including playback and listening, I figured that 96 or even 48k would have also been quite good. Something about an emporer and new clothes.

Back to the OP. I spent part of my evening yesterday teaching my daughter how to do file conversions in TwistedWave and QuickTime 7. She uses Logic but so many others do not.

Henry Robinett · Post by **Henry Robinett** » Sun Feb 21, 2016 11:54 am

MIDI Life Crisis wrote:
Henry Robinett wrote:
MIDI Life Crisis wrote:BWAV

Everything in 48k/24 bit

THE END
Everyone I work with uses 96k/24. Myself included
The question then is "why?"

Film and video generally require 48/24 unless it's hi-def. 96 doesn't improve anything afaik.

Why consume more disk space? What do you see as the advantage of 96?

I think it sounds much better! Also I'm preserving my own work in case things change. I can hear without a doubt hear the difference. Not even close.

All the best,
http://www.henryrobinett.com
Check out my latest CD on iTunes.
I Have Known Mountains by Henry Robinett
https://itun.es/us/pi6C_

Michael Canavan · Post by **Michael Canavan** » Sun Feb 21, 2016 12:01 pm

Henry Robinett wrote: I can decibel without a doubt gear the difference. Not even close.

That's some funny autocorrect there!
Hey, much better than F-bombs and "bizitch" this and that's by accident.

Henry Robinett · Post by **Henry Robinett** » Sun Feb 21, 2016 12:08 pm

The first time I realized the difference was mixing and mastering a CD of mine by Mark Wilshire, head of audio production for Peter Jackson and all of the LOTR films. We'd been mixing solid for three days at 88.2. When we started mastering I went to the bathroom, came back in the room and EVERYTHING CHANGED. I didn't know much then. Things had sounded three dimensional. Real depth. Amazing. Suddenly it was flat. No dimension. Crestfallen I asked what happened. He apologized and told me we converted it to 44.1. It was terrible.

So FOR ME, that was my blindfold test. I had no idea I was taking a test. Soon afterwards I worked at 96k more or less exclusively. Well it took me a few years to get the gear. I had been an adat (gag) guy. That was mentoring that took me several years of buying and training to acquire the tools needed.

As always YMMV.

All the best,
http://www.henryrobinett.com
Check out my latest CD on iTunes.
I Have Known Mountains by Henry Robinett
https://itun.es/us/pi6C_

EMRR · Post by **EMRR** » Sun Feb 21, 2016 7:31 pm

I moved to BWAV the second it was available in DP.

If you use a 2408mkIII or 24i/o as I did for a long time, higher sample rates are quite clearly superior in sound to lower rates. If you use anything in the new AVB range, it's not as obvious. I hear the same thing about legacy pieces like the Lynx Aurora which were so popular until recently.

I've done everything at 88.2 or 96 for a decade now.

If you work analog at all during the mix process, meaning D/A then A/D again, then why wouldn't you work at the highest quality practical? Phase shift on the top is easily measurable, and it's already pretty bent at 44.1/48 with just A/D, let alone more conversions.

Jesse DeCarlo · Post by **Jesse DeCarlo** » Sun Feb 21, 2016 10:59 pm

Yeah, to be clear, for most people who use high sample rates the goal isn't to preserve ultrasonic frequencies - it's to keep the audible frequency band from getting screwed up.

MIDI Life Crisis · Post by **MIDI Life Crisis** » Mon Feb 22, 2016 7:54 am

And when it gets played back at 44 or 48, aren't those theoretical (and anecdotal) advantages defeated?

EMRR · Post by **EMRR** » Mon Feb 22, 2016 7:57 am

Those advantages exist any time there is processing or conversion. Your mastering guy likely takes it back to analog for some part, and they do prefer an 88.2/32float mix.

FMiguelez · Post by **FMiguelez** » Mon Feb 22, 2016 8:38 am

Until somebody DEMONSTRATES (not asserts) their claims, I will choose to simply take all these Ultra high SR hype claims as what they most likely are> Empty but hyped claims, probably designed by marketing departments, who pray on the gullible.

A double blind listening test would be convincing. Repeatability is a MUST, and the answers must be consistent and statistically meaningful.

There are people who also see things, hear things, and feel things that are most likely not there, and they are as sure about that as they exist. If neuroscience has taught us anything, is that our brains are so easily deceived and confused. Have you ever played one of those optics game tricks???

Can we at least agree on cancelling? If 2 files cancel digitally, (everything else being equal), they are, by definition, IDENTICAL. No gremlins, no magic, no fairy dust. They are identical.
Yes? 'Cause I've heard people claim they can STILL hear differences between 2 identical audio files that cancel each other when one flips one track's phase

Now, there MAY be a few advantages on certain scenarios of using high SRs, but I'm not sure they're worth it (tripling or quadrupling disk space and CPU usage), so that some maracas sound absolutely pristine (which I doubt) in a pimpled-faced 15 year old's iPod and iBuds.

When does this end? Why not go to 500 GHz sampling rate? Why not 128 bit floating point? May Tlaloc forgive we miss some "air" frequencies and definition all they way up there...
And what about the ultra low frequencies? Are we perhaps missing some beautiful elephant frequencies? I think we're wasting our time until we don't have equipment that goes all the way down there, even if we can't hear it... there may be extra information in there that enhances the upper frequencies and our listening experience too!
So don't forget to restock those special crayons which make CDs sound better, or those special stones that make frequencies relate better... Oh, and the carpets! Do not dare letting a carpet touch a cable. You need those cable-elevators if you are really serious about your music

Again, when does all this end?

MOTUnation.com