KVR Audio

quantize · Post by **quantize** » Fri Aug 05, 2005 7:32 am

yeh but if you follow that logic to its conclusion, 200GB wont ever be enough either..

Failed Muso · Post by **Failed Muso** » Fri Aug 05, 2005 4:17 pm

quantize wrote:yeh but if you follow that logic to its conclusion, 200GB wont ever be enough either..

Exactly

Which is what drives technology on & on, and makes us even lazier and unimaginative

It's a shame really, because I have sample sets here dating back to Akai S1000 and some of those sets are just great, yet they fit on a floppy disk ! All of this achieved by very clever use of loops and other tricks. Even now, I have sample sets that are designed for my S6000 or Z8 that, at 20 or 30MB, are considered tiny by today's standards but are as good as sets 10 times their size because they employ clever programming.

Take a look at Hollow Sun's sample CD's. Perfect examples of high quality, low footprint sample sets that professional musicians and producers are currently raving over.

korgsp200 · Post by **korgsp200** » Sat Aug 06, 2005 3:00 pm

Closed

Failed Muso · Post by **Failed Muso** » Sat Aug 06, 2005 3:05 pm

korgsp200 wrote:thats there. but look at NDB samples.. check out their loops.. www.ndb.hu they are long.. pretty long and of high quality... samples quality , resolution and length do matter and loops too do matter.. where loops are made without crossfading.

Am I the only one who doesn't quite understand that reply ?

korgsp200 · Post by **korgsp200** » Sat Aug 06, 2005 3:07 pm

Closed

Failed Muso · Post by **Failed Muso** » Sat Aug 06, 2005 3:20 pm

korgsp200 wrote:lol.. i am talking bout crossfading loops... some use this to make loops faster but this is risking the quality of samples...

It all really depends on the sample itself. If it is a long evolving pad sound, then a longer loop is probably necessary, but for many sounds, a short loop is more than adequate. You can use the internal filters and synthesis methods of your chosen sampler to further tweak and shape the sound.

Cross-fading is best used to clean the loop points up, no matter how long the loop is. The loop points should be determined by the character of the sound and to facilitate a good long sustain but at the same time, minimising the memory footprint.

korgsp200 · Post by **korgsp200** » Sat Aug 06, 2005 3:23 pm

Closed

Sascha Franck · Post by **Sascha Franck** » Sun Aug 07, 2005 1:36 pm

Btw, for anything sort of artificial, I'm usually getting the best results when setting up a loop x-fade ratio of 50% in WaveLab. I'm allways trying to find good start and end points (read: the sound should be similar), then apply an x-fade with the mentioned setting. I do experiment with the x-fade's "power" parameter though.
Of course things are becoming more difficult on a variety of situations. Should there be some more or less clear rhythmic information in the sample, you need to be more careful and find out about the best loop range suiting the rhythmic elements first. Yet, the 50% x-fade setting is usually working fine.
And then there's the decay portion of the sample. Should I want to keep it, the x-fade method won't work anymore, but WL offers some "post-loop fade" functionality, that you can usually get away with.
Obviously, this method won't work as good for natural sounding samples. For those I try to find a good loop point without x-fading and then try to keep the x-fade value very small (if required). But I have to admit that there's quite some samples that won't loop too well.

Anyways, IMO you can get away really fine with the 50% method for most synthetic sounds, assuming you sort out anything rhythmically happening first.

zircon · Post by **zircon** » Mon Aug 08, 2005 5:06 pm

Well, I do work with mainly synthetic saples. I'll have to give that a shot. The problem with crossfades is that there is a very slight 'jump' in the natural cycling of the sound due to phasing, flanging, unison, detuning, chorusing, or various other factors. That's what gives away the loop, and you can't really avoid it without sampling the sound for 3 minutes.

Sascha Franck · Post by **Sascha Franck** » Mon Aug 08, 2005 5:19 pm

zircon wrote:Well, I do work with mainly synthetic saples. I'll have to give that a shot. The problem with crossfades is that there is a very slight 'jump' in the natural cycling of the sound due to phasing, flanging, unison, detuning, chorusing, or various other factors. That's what gives away the loop, and you can't really avoid it without sampling the sound for 3 minutes.

If you'd supply some sample, I could demonstrate things.
Seriously, the 50% setting is working on a LOT of material, assuming you stay away from involving any attack stages (which often sound pretty much different).
Of course things are getting a LOT more difficult as soon as there's a lot of internal movements involved.
Also, finding proper loop points on sounds that fade out quickly is another nightmare...
But it's all fine with most synth sounds. You gotta make the right decisions before you even start to sample things.
The key question being: Are my LFOs/filters/envelopes/FX and whatever doing something that can't be reproduced later on (when loaded in my sampler of choice)? If yes, I keep them, if not, I switch them off.

Sascha Franck · Post by **Sascha Franck** » Mon Aug 08, 2005 5:37 pm

I did a quick example of the 50% setting I'm using:
http://home.arcor.de/s.franck/temp/Testpad.exe
(selfextracting archive, no worries!)
I included both the original and the looped wave file.
Searching for proper loop points inside the unaltered file would be close to impossible, as there's quite some crossmodulating and FX stuff going on. Using approximate loop points and then applying short crossfades would be pretty much noticeable too.
With the 50% setting you almost don't notice the looping.
I'm sure a similar technique is found in things used for automated sampling, such as Sample Robot or Disco DSPs Highlife.

Zero-X Seamless Looper..