The aliasing thread

[valley]

I think that those graphics only scares people from good software. I'm really surprised that no one has sued those owner-sites for "false-information".

Because, as you yourself admit in your very next sentence, the tests are correct.

But in a "musical" way, they aren't. Get it?

No, because that isn't what they are setting out to determine. George stated way back at the start of this thread that a number of things, of which sound quality is only one, come into play when choosing an instrument.

I own a Korg Prophecy. Frankly, by most standards it sounds crap, but I wouldn't part with it 'cos it sounds crap in a way I love, and it is a great control synth.

I also own Vanguard, which aliases like hell, and sounds a little cold to boot. I love Vanguard too. It has a sound I can't get anywhere else, and the UI is laid out such that I can get that sound without any effort at all.

Sytrus sounds great, but I hate the UI, so I left it.

[WilliamK]

..

[René]

Urs is correct in his statement. You are not.
Truth is not achieved thru democracy.

-René

[donkey tugger]

Truth is not achieved thru democracy.

-René

Blimey, quite the philosopher!

[WilliamK]

..

[René]

I will shut-up now. Maybe I'm an ass, maybe I'm stupid, but again, maybe I'm a genious and the problem is not me. Hummm, Einstein had the same problem trying to prove his theories...

Now we're understanding each other. Einstein liked the sound of sfz tho...

-René

[Majken]

Now we're understanding each other. Einstein liked the sound of sfz tho...

-René

Now that's just bollocks, Einstein was a big Synth1 lover. He was always bashing Rasputin for favoring Crystal btw. Ah yes, those were the days!

/Majken

[WilliamK]

..

[george]

Before anything else, I just would like to show that I'm not alone on all this.

You only need a 15khz sinewave sample as root (C-4) and proceed to play from C-4 to B-4 and B-3 to C-3 (backwards) at 80 bpm.

As you use a sine wave and the spectrums posted on that site show multiple frequencies, the phenomenon proved is not aliasing. Period.

Regards,

Urs

Check Urs quote later on:

Now - and here I admit that some of my yesterday's observations have to be reconsidered - these additional spectrums (or are they called "spectra"?) do of course alias, if crossing the Nyquist boundary. I think I can say that to some extent this is what we see on the pictures.

[bmanic]

It would be quite nice to know what monitoring system people have as I guarantee you that a lot of the "aliasing is not heard in a mix", "you can't hear it with chords", "aliasing is a myth" etc threads would die quite quickly if people would have access to a professional monitoring chain. A soundblaster live + hi-fi speakers is NOT the answer. KvR sometimes goes into the mode of " I don't own a formula 1 race car but I sure know how it feels to drive it as I have a sporty Honda ". Most folks would be surprised how easy it is to do an ABX double blind test and get 100% accurate answers when you have a good quality monitoring chain.

However, aliasing does not equal bad sound. Not at all. Alisaing can be very nice but it can also (and is more often) be a problem.

My point? DO trust your ears for scientific tests but only in a controlled environment where all the 'mud' (read: consumer equipment, bad accoustics etc) is removed. Of course it helps if you're not deaf.

Does this text belong in this thread? Well, in a way it does. I for one hear the sinc 256 vs linear interpolation in a FULL MIX if I pitch up a closed hihat or crash sample by as little as one semitone. Now add on top of that ALL the samples in a mix that I tweak a little to get the timing/length/timbre correct by tweaking a couple of cents. It does add up, quickly, and makes quite a difference.

My 2 non scientific but ear VERIFIED (abx) cents.
bManic

[WilliamK]

..

[WilliamK]

..

[stefancrs]

OMG. Some people seriously needs some perspective.

1) Everyone knows that a spectogram does not show how good a sampler is, it only shows the result of a certain input.

2) For the ones wanting hifi resampling, under "all" circumstances, this is a really big deal.

3) For the ones wanting a cool gui / nice workflow, (1) might not at all be a big deal.

4) This goes to WilliamK actually. Dude, calm down. If you feel that _everyone_ tells you that you are wrong, you actually might be. You do have some points, but on the whole, you are quite wrong. For some reason you have totally ignored my posts about spectral content (repitching high amplitude stuff near nyquist IS tough, and it will probably happen even in wusikstation).

No hard feelings though, I just think people actually should _read_ the points risen by others.

Btw, anti-aliasing as opposed to aliasing (when not going to extremes), can give you the _illusion_ that you have lowpassed the sound, when on the contrary you have distorted the sound in the latter case.

Now, would someone please show me zero-latency lowpass filters :) Analog or digital.

[arguru]

I would to comment a few last words on this subject.

As developer, and from my position as DSP learner (despite i'm developer audio software for some years ago, I have basic knowledge of DSP culture, specially from a theorical perspective), but i'm with george and rene in this area, and with probably along a vastly amount of developers and dsp enthusiasts, in fact, aliasing, interpolation and resampling, along with filters, are the heart and the soul of digital signal processing.

In the other side, if we're talking about creating a 'overall' good product (synthesizer, sampler, whatever), the subject is far from represent a critical piece of the final thing. Probably there're 300 things more important in the task of create a -good- product.

Maybe I sounded a bit arrogant sometimes, but Willian, dont take too seriously, and no one said in this thread that your product(s) are crap (nothing far than that). Ironically, most famous (and by definition, best) gear ever (DX7, Akai Samplers, etc) aliased as pigs being raped.

Argu

[WilliamK]

..