DSP Robotics Support

Hi,
I remember few posts about a "single sample based sampler" that can load a single sample and play with the whole range of the keyboard. Something like the one of KG's, but in the most basic form, withput any GUI features. Just the green+midi+blue block so I can add my own knobs, controls, etc. Does someone have any link or ready block in the tool box?

Thanks

There is one in the default toolbox called "wave player". Or it might come from the "audio toolbox" that you can download from the forum "modules" category.

I actually don't need a "song wav. player" player, but an "instrument wav. player"
I mean one of those you can, for instance, load a single pinao sample and play with it the whole keyboard spectrum. Similar to the one that you have shared on the forum a while a go, but just without any additional GUI features.

Here's the toolbox one->

kortezzzz wrote:I remember few posts about a "single sample based sampler" that can load a single sample and play with the whole range of the keyboard.

I was wondering if you meant a sampler that changed pitch over the keyboard range but kept duration constant, so no "Chipmunk" effect.

This topic cropped up on the Gyl Synths FB page a while back and didn't get resolved.

It would be a nice thing to have!

Cheers

Spogg

Spogg wrote:
I was wondering if you meant a sampler that changed pitch over the keyboard range but kept duration constant, so no "Chipmunk" effect.

Yes, spogg
you've spotted that correctlly. I know there is something around because I remember a topic that shared that schematic at the forum a while ago. Just don't remember the name of the publisher nor the files name (which probably sit on my hard drive whithin the rest of the hundreds of downloaded .fsm files )

Nix,

Thanks for that module. It works very reasonably and does the job
If anyone have something better, please share

Spogg wrote:a sampler that changed pitch over the keyboard range but kept duration constant, so no "Chipmunk" effect.

Actually those are different things: a pitch shifter that keeps the duration is not necessarily free of that chipmonk effect. I would say there are three basic repitching algorithms in the time domain:

1. Play at different speed - the simplest you can imagine. Duration and spectral envelope will change (chipmonk)
2. Play at different speed but repeat or skip one cycle every now and then to compensate for the duration change. This one will have a chipmonk effect, although the duration will be corrected for.
3. Play grains of the sound sample at the original speed but at a different rate. The grain size and overlap relate to the original and target fundamental frequencies, respectively, and you pick the grains at constant phase. This will preserve both the duration and the spectral envelope, so no chipmonks here.

I am attaching a demo for the third type , which I ripped off from my MIDI Choir plugin. The demo is monophonic though. The method is preferrable for vocals because we are so much used to the formants being at the right places. It may not be the best choice for other sources.

Have fun!

martinvicanek wrote:Play grains of the sound sample at the original speed but at a different rate. The grain size and overlap relate to the original and target fundamental frequencies, respectively, and you pick the grains at constant phase. This will preserve both the duration and the spectral envelope, so no chipmonks here.

I was always baffled by this sort of algorithm. Specifically, how does it obtain pitch-shifting when the grains are played at original rate? Do the phases of the grains add up in a way that results in a different-pitched signal?

KG_is_back wrote:

martinvicanek wrote:Play grains of the sound sample at the original speed but at a different rate. The grain size and overlap relate to the original and target fundamental frequencies, respectively, and you pick the grains at constant phase. This will preserve both the duration and the spectral envelope, so no chipmonks here.

I was always baffled by this sort of algorithm. Specifically, how does it obtain pitch-shifting when the grains are played at original rate? Do the phases of the grains add up in a way that results in a different-pitched signal?

Would like to know that, too.
I explain it to myself with sine waves. If I split one cycle into 10 grains and then shift them closer together, so that they cover only half the time, it's pretty much the same as playing the sine double as fast. But of course this is just a guess.