PlanetZ forums - scopeusers.com

96k is wasteful of resources, huge files that will lose all that extra edge anyway when downsampled. so what if it's "better". you won't be any more likely to sell a cd or win a grammy....

this reminds me of why i sold my 2" 24 track tape machine....

right guys, using the available quality stuff plus making the right choices considering eq divorcing of the parts and good eq-ing makes 98% of the result.
I give away those last 2%, if it means recording at 96kHz

I still say the same:

If you are making up a REALLY special job for film production or a very important release in DVD, ok, it is suitable, otherwise it is a waste.

I too agree with Astro, that there is a lot of psychological impression going on with this.

Let me tell you about a little experience we did with an uncle of mine. He liked very much red wine, and was always telling how "good" his perception of red-wine taist was... He knew about every single good and expensive wine in the world. He had also a collection of bottles at home.

My father once said he liked a cheap red-wine and he assured my uncle that this cheap wine was as good or better than a very expensive wine he bought for a party. Well...

There was a silly discusion about which wine was better, and so my father would deffend his idea and my uncle his. My father went in secret to the kitchen, and swapped wines... he put the cheap wine into the expensive bottle, and the expensive wine into the cheap bottle.

My uncle continued with his silly request, and my father said: "Let's see what people think". He gave halp a glass of wine to everybody there, and everybody said the "expensive" wine was far better... of course, this "expensive" wine, was in fact the "cheap" one my father swapped.

When my father told them this, everybody was amazed and confused. My uncle was as RED as his "expensive" wine... That's what I think too, about audio, even if I agree with those who can afford recording in such high settings without problems

Now i know,I like this stories.
cheers everyOne.

What ???
Higher sample rates are not about voodoo, placebo, hype, marketing ... whatever.
Just record any synth at 96 Khz, do a sonogram and unless on lowpassed presets, territories above hearing range are usually not a barren land and all those sonic landscapes are just fold back into hearing range when using 44.1 ... just in case you doubt about your ears / monitors !
Now using 96 K is roughly 2 times more demanding for something different and not necessarily 2 times " better " ... BUT I prefer to go that path and many painful " unexpected DSP overload " than keeping 44.1 ANY day and I regret I hadn't go into a more " wasteful attitude " just for more pure egoist sonic pleasures.
Also downsampling with decent antialias filters do little harm to original files, I did many tests and it was hard to notice 96 K downsampled to 44.1 properly and original 96 K.
I also view 96 K samples as " dry " if you want to make them " wet " by downsampling without antialiaising filters well you can do ... at 44.1 you have fully aliased wet samples and well ... you're are screwed ...
For good measure I'am more than ready to embrace 192 K for that extra leap even if it sucks my DSP, CPU, HD, RAM ... a huge grin is priceless !!

so you think 96k sounds better because it's more precise, don't you ?
What if... you're fooled and that's just the wrong conclusion - it's perceived better because it's constantly distorted at a very low degree...

you know that jitter is inevitable and results in slight distortions of the waveform.
A high jitter rate is perceived as very unpleasant and with decreasing values the sound gets better and better - but when it's almost completely jitter free, the perceiption suddenly changes and the sound is described as cold, thin and sterile.

Think of it as similiar to a good tube amp.

I consider the story above trustworthy because the dude who reported it was experimenting with extremely precise measurement gear and fooled himself during a blind test.
He used a system that could deliver a defined rate of jitter (which he forgot to shut off) and when he finally reached the sweet spot with that WOW that's the prefect sound, he suddenly got aware of his mistake. The 'cold' sound he experienced before was in fact the almost jitter free signal.
Of course this was a stunning and funny (non-linear) experience

So what does it have to do with high sample rates ?
At 44.1 k the clock deviation allowed before the waveform gets noticeably (dunno to which degree) distorted is... 6 picoseconds, roughly the 150th part of a nanosecond.

For obvious reason this is not too easy to be kept constant, but at least it's no fairy tale - people here have reported that a 'studio clock stabilized' digital signal is noticably better, on the same gear.

Now try that trick with a higher sample rate - afaik the required precision is almost impossible to be achieved with affordable constructions.
And since those cards (for high sample rates) do not have any special circuitry for exactly that purpose, you may assume that the jitter amount (in very small units) is present up to a degree where it starts to resemble a dither algorithm.

Which would explain (to a degree), why it sounds so good
But why apply all that effort and waste of resources, if it's only a side-effect that causes this kind of sound improvement ?


The industry uses 96/192k for nothing but cheaper filter design AND to tell customers the sound will be better because the number is bigger.
They don't develope for the one audiophile out of ten thousand folks, who can both afford the gear and has the hearing cababilities to notice the improvement

cheers, Tom

<font size=-1>[ This Message was edited by: astroman on 2005-05-10 19:53 ]</font>

@ Decimator, give us a short 44.1 audio file containing two identical 44.1 and a 96-->44.1 recorded music fragments.
I'd like to be convinced

Does your average "person on the street" really notice the difference between a 44.1 or 96Khz recording? I know that, when I do get round to actually finishing anything, I will play it to my mates who won't notice how wonderfully glorious the bass sounds and how well it integrates with the kick drum (which I spent the last two days trying to perfect!) let alone the superior quality of 96Khz . But, then again, 15 years ago I didn't think I would ever need a mixer and could just use some plugs and cables from Tandys - "EQ" I don't need that!

<font size=-1>[ This Message was edited by: hesnotthemessiah on 2005-05-11 00:47 ]</font>

beerbr, you're absolutely right... A pitty this kills any option of developing for 96kHz, SDK comes only w the large decks, and most DP's are on that too afaik.

On 2005-05-10 17:53, decimator wrote:
...and all those sonic landscapes are just fold back into hearing range when using 44.1 ...

oops, overlooked the foldback - which cannot be denied and is one of the reasons that filters are easier to design at 96k and above.
But doesn't that imply that a 'perfect' anti-alias filter (which exists only in theory of course) would make both sampling rates identical to our (limited) perceiption ?

cheers, Tom

I run a lot of sonograms because I love the beautiful patterns I see at times but still don't see always a clear relation between what I hear and what I see ...
Anyway, at 44.1 Khz you have no headroom : any spectral content above hearing range generated by those digital synths will be fold back, at 96 Khz you have a decent headroom : up to 48 Khz and even 20+ Khz above since a frequency of 70 Khz sampled at 96 Khz will fold back at 26 Khz ( 48 - ( 70-48 )) so out of hearing range.
However I've seen sonograms so crowded at 96 Khz that I really want to be on the next 192 Khz wagon on some cases and I expect some excellent surpises !!!!
In the meantime at 96 Khz all my synths sound more expensive : cleaner, punchier, sharper and I'am maybe more sensitive since I began to be not very happy with what I heard on 44.1 Khz
I don't have time to spend on A/B, anybody can do it : take a synth ( Lightwave is a good canditate, at least on some presets I made ) and if there's a noticeable difference between 44.1 and 96 it's because there's some spectral content above hearing range, simple as that ...
As for resampling with antialias filters, I'am happy with Sound Forge but R8brain free and R8brainpro from Voxengo are more professionnal.

it's not the event itself that some doubt that 96khz improves, the evidence is in the hearing. it's the final product vs. resources used....

hi

when creamware will solve 96khz problem
with our current cards?
I'd like to works in some project at this sample rate like is wrote on the cards
when i have bought them.
18 masterverb on 44khz/vs 1 masterverb on 96khz
maybe they could optimize some things
in the SFP system software,
or the development on our current cards is finished????

no, it's delayed, as you know they're after new resources

I think only new hardware can solve 96Khz problem. The main reason is insufficient internal dsp memory and maybe more important interchange dsp bandwidth on STDM bus in current DSP hardware. Also serious problem is lack of fast external memory access in current CW hardware due it architecture using shared host memory.
Will the new hardware ever be or not - I don't know. The main reason of course incompatibility old staff with the new (if it will be)

the developement on ANY computer component is NEVER finished as the platform is never stable and finished. as to compatability, the jury is out. the new sharc chips do many times the work and they are code compatible with the old ones, so anything's possible at this point....