KVR Audio

pough · Post by **pough** » Wed Sep 15, 2004 2:45 am

Robert Randolph wrote:Anyways, never trust your ears!?!

What the f**k, this doesnt have ANYTHING to do with music anymore then does it? Why are we testing samplers? They are musical instruments... that we hear with our ears...

Cmon now, I love technical babble as much as the next guy, but never trust your ears? That's baloney.

None of your senses should really be trusted much beyond simple preference. When you can hear differences that don't exist (and can't hear differences that do exist) maybe you should consider keeping preference separate from science and focus on how the two impact each other and where they tend to really meet. Which might not be the same place for different people.

And isn't that what this is all about? Some of these conclusions may end up being pure geek and others pure art, but it's worth the effort to find out, IMO.

Trust your ears to lie to you regularly and trust yourself to enjoy it.

Robert Randolph · Post by **Robert Randolph** » Wed Sep 15, 2004 2:49 am

I was pretty sure we're talking about preference in the case george quoted...

if we're talking science, sure. Agreed.

But this has turned into a preference discussion by WilliamK.

So hard to enter the conversation, because I respect all the parties except Urs (uh, evil mac guy) greatly... Particularly when all the devs have amazing products all vested in this discussion, and they all sound fabulous.

george · Post by **george** » Wed Sep 15, 2004 2:52 am

The never trust your ears statement is because a singular person auditive range won't be the same as others. Maybe you are feeling like getting the perfect mix, but unfortunately can't hear anything beyond 15khz when others will.

pough · Post by **pough** » Wed Sep 15, 2004 2:54 am

Robert Randolph wrote:I was pretty sure we're talking about preference in the case george quoted...

if we're talking science, sure. Agreed.

But this has turned into a preference discussion by WilliamK.

Okay. Nevermind, then.

Robert Randolph wrote:So hard to enter the conversation, because I respect all the parties except Urs (uh, evil mac guy) greatly...

Um, More Feedback Machine?

Robert Randolph wrote:Particularly when all the devs have amazing products all vested in this discussion, and they all sound fabulous.

Yup. I think I own quite a few of them. I tried importing that 15kHz WAV into WusikStation. Some interesting results, from a scientific point of view. From a more musical point of view, every single note on the keyboard sounded like pure shit!

ascdi · Post by **ascdi** » Wed Sep 15, 2004 2:54 am

urs sed:

However, the spectrums on that site do not say anything about aliasing. An aliased sine always remains a sine and will always do so.

someone care to explain this? i can load up a sine and put it through a s/h and it sounds gronky. if that's not aliasing, what is it?

i mean, i understand that a sine at nyquist is a sine because of the filtering of the d/a whatever thingy that urs talked about on the last page of this thread. but say you take a "sine" stored as WAV data @ 44.1 and load it into a DAC that is operating at 192khz... then you'd get small nonharmonic compenents, right?

edit: furthermore, that is exactly what a sampler is doing: if you put a 22050hz sine into a sampler and play c4 it will "sound" fine, but if you play c3, you'll be effectively playing a 11o25hz sine @ 22050hz "resolution" under the *same* DAC bandlimiting thingamabob, and any nonharmonics generated by the aliasing that are not above the 44.1k DAC filter will be audible.

which is why testing wusikstation at 192k misses the point. of this supposedly pointless thread, but hey, let's not split hairs. personally, i like to know exactly what my gear is doing, even if people who are smarter than me tell me that it's impossible to notice it.

WilliamK · Post by **WilliamK** » Wed Sep 15, 2004 2:57 am

george · Post by **george** » Wed Sep 15, 2004 3:02 am

ascdi wrote:i can load up a sine and put it through a s/h and it sounds gronky. if that's not aliasing, what is it?

Try again tomorrow. Mr. Urs just got back from the pub

but say you take a "sine" stored as WAV data @ 44.1 and load it into a DAC that is operating at 192khz... then you'd see small nonharmonic compenents, right? that is what george's test is all about, right?

When you have a pure sinewave tone, it is very easy to check, since you will only trigger the fundamental frequency and the rest will result into aliasing (junk).

Majken · Post by **Majken** » Wed Sep 15, 2004 3:08 am

However, the spectrums on that site do not say anything about aliasing. An aliased sine always remains a sine and will always do so.

This I didn't follow either. As I see it aliasing is when the sinewave goes from being a pure sinewave into being something different (high frequency harmonic content). Where the product would result in a slightly jagged sinewave if you were to look at it using an oscilloscope.

/Majken

george · Post by **george** » Wed Sep 15, 2004 3:14 am

WilliamK wrote:That's what I'm trying to get here. From a musical point of view, who will use such high frequencies?

It's not the use, it's what they bring. Frequencies whose aliasing get reflected above nyquist will add extra junk to the mix.

I know, this was used to test the aliasing problems. BUT, I did a simple test using a more simple file, and the same MIDI file you guys used for the test. And aliasing, even with linear-interpolation, is very low, inaudible.

Audible concept is quite personal. It's not the same using a SB live and el-cheapo speakers than latest EMU card and decent studio monitors.

Majken · Post by **Majken** » Wed Sep 15, 2004 3:22 am

I'll be happy to spot aliasing for you once you've got some samples. I'm running an E-MU 1820m and some Sennheiser HD-570 headphones. I'm also oversensitive to high frequencies. So if they're there I'm likely gonna cry like a baby

/Majken

george · Post by **george** » Wed Sep 15, 2004 3:26 am

Majken wrote:I'll be happy to spot aliasing for you once you've got some samples. I'm running an E-MU 1820m and some Sennheiser HD-570 headphones. I'm also oversensitive to high frequencies. So if they're there I'm likely gonna cry like a baby

Beware with headphones and high pitched sinewaves! You should at least consider setting sample volume about -12dB first.

Majken · Post by **Majken** » Wed Sep 15, 2004 3:30 am

Guess why I was crying when listening to the rar'ed wavefiles on your site

/Majken

george · Post by **george** » Wed Sep 15, 2004 3:34 am

Yeah, I should put some listening warning... just thought about cases of 15khz sine tests at 0dB peak on 100W monitors

autloc · Post by **autloc** » Wed Sep 15, 2004 3:58 am

Why not include some additive resynthesizers in the test? Perhaps it's slightly beyond the scope of your investigation, but given that you can't (generally) be certain of the methods being used internally by the samplers, it might be interested to have Cameleon or WNAdditive as points of comparison. (Of course I would expect them to do much better, but it could be an interesting extension)

btw, one of the links is dead: http://www-2.cs.cmu.edu/afs/andrew/scs/ ... iasing.pdf

why not add the wikipedia article?:
http://en.wikipedia.org/wiki/Aliasing

george · Post by **george** » Wed Sep 15, 2004 11:15 am

autloc wrote:Why not include some additive resynthesizers in the test?

The test is aimed to samplers and their resampling algorythms. If you synthesize a sine you shouldn't get any alias at all. I have put a example using Vertigo, and it should be about the same no matter what additive synth you use (if the waveforms are sinewaves, of course).

btw, one of the links is dead: http://www-2.cs.cmu.edu/afs/andrew/scs/ ... iasing.pdf

Thanks, I have just removed it.

why not add the wikipedia article?:
http://en.wikipedia.org/wiki/Aliasing

Good suggestion, it will be added asap.

The aliasing thread