| Author
|
capturing waveforms
|
mikhail
Started Topics :
7
Posts :
18
Posted : Dec 15, 2007 01:05
|
hey there, Im wondering that whats the best way to bounce a waveform generated in some synth in Logic 8?
Im aiming to load the waveform into Kontakt 2.
probably a very simple thing to do but dont know how.
peace |
|
|
Elad
Tsabeat/Sattel Battle
Started Topics :
158
Posts :
5306
Posted : Dec 15, 2007 06:06
|
well..
it should be repetative .. then u need i reken the tinest peace u can of it that still create a whole loop (if kontakt gonna loop it)
for example , a triangle wav will actualy look like triangle , and a sine will look in sine form.. so to capture the wavform u need to detect the form
keep zero cross points to avoid clicks 
 www.sattelbattle.com
http://yoavweinberg.weebly.com/ |
|
|
|
Speakafreaka
IsraTrance Junior Member
Started Topics :
18
Posts :
779
Posted : Dec 15, 2007 17:58
|
|
Trip-
IsraTrance Team
Started Topics :
101
Posts :
3239
Posted : Dec 15, 2007 20:41
|
That's a good idea speakafreaka
Crackling universes dive into their own neverending crackle...
AgalactiA |
|
|
|
Elad
Tsabeat/Sattel Battle
Started Topics :
158
Posts :
5306
Posted : Dec 15, 2007 20:51
|
+1
also lower pitch will defintly help to detect the form easier , tho my guess is u need to zoom 100% to get this tiny piece of gold
i have actualy a folder i found somewhere called "virus waveshape" not sure its realy from virus but anyway fun to mess with 
upload coming.. (hmm i hope its nothing illeagal there.. seems clean to me anyhow) www.sattelbattle.com
http://yoavweinberg.weebly.com/ |
|
|
|
Elad
Tsabeat/Sattel Battle
Started Topics :
158
Posts :
5306
Posted : Dec 15, 2007 21:00
|
|
acidkills
IsraTrance Junior Member
Started Topics :
26
Posts :
431
Posted : Dec 16, 2007 02:43
|
|
mikhail
Started Topics :
7
Posts :
18
Posted : Dec 16, 2007 23:13
|
And you just record different notes to get different pitched waveforms, right?
thanks for the replies |
|
|
|
Speakafreaka
IsraTrance Junior Member
Started Topics :
18
Posts :
779
Posted : Dec 16, 2007 23:59
|
I just use one sample and spread it out from the bottom up. No need to go multisample crazy, unless you are trying to catch something other then the waveform - like a specific envelope sweep of the synths filter.
If you just want the waveform, just get as low a sample as humanly possibly, cross at zero points, and map across the keyboard.
.
http://www.soundcloud.com/speakafreaka |
|
|
|
Elad
Tsabeat/Sattel Battle
Started Topics :
158
Posts :
5306
Posted : Dec 17, 2007 00:48
|
btw if someone hear just clicks instead of sound it meens he need to activate loop mode (just incase)
also note some synths can use it as wavform (fruity's native 3osc and ts404 , fun to have more then 5 shapes) www.sattelbattle.com
http://yoavweinberg.weebly.com/ |
|
|
|
Speakafreaka
IsraTrance Junior Member
Started Topics :
18
Posts :
779
Posted : Dec 17, 2007 01:07
|
|
mikhail
Started Topics :
7
Posts :
18
Posted : Dec 17, 2007 11:06
|
Why do you choose the lowest possible note?
What difference would it make to choose a higher note?
|
|
|
|
Speakafreaka
IsraTrance Junior Member
Started Topics :
18
Posts :
779
Posted : Dec 17, 2007 12:00
|
If you choose a high note it will sound fine when you play it back at the exact same pitch.
However, any wave consists of a number of samples, which could be viewed as a line of points that when played back in order reproduce the given sound.
When you pitch this sound down, you effectively increase the number of points needed to produce the sound.
A low pitched sound is lower in hz, and therefore the waveform is longer.
So if you have a high, short waveform, how is it possible to make it longer?
Well, what the computer does is something called interpolation.
Lets say you want to play something an octave down. This requires exactly twice the number of samples over the original sound. But how can the computer generate this information when it does not hold information for all these samples? (only half of them, as it happens)
Well, one way would be to take two samples, and generate an average value between the two, and then put this in the 'missing' sample area. Or we could get into curve math, but frankly its too early in the day - either way, its a guess. One is a more accurate guess then the other, but basically, you cannot know what is in the missing sample area - you can only say what is the most statistically likely value.
The computer cannot know, and at best you are getting an estimated sound. Of course, it can sound perfectly acceptable for minor shifts, and is a damn sight better then no guess at all, but it can become very noticable over larger shifts, and not at all what the doctor ordered.
If you start with all the information you could possibly need about a waveform (ie a low note) and pitch up, it is perfectly possible to get rid of the uneeded information through the same process of interpolation, with minimal effect on the sound - for example if the sample position does not match up precisely because the waves are playing at different speeds (which *will* happen).
Basically, its much better to get rid of the information you don't need from a complete picture, then guess at information you don't have, and make false presumptions, which is what happens when you sample wavecycles high on the computer.
You are effectively implementing a crude pre recorded low pass filter on your sounds by sampling high.
.
http://www.soundcloud.com/speakafreaka |
|
|
|
