Timestretch and pitch shift?

Whats up everyone. I got a problem its happened a few times. Ill have a sampel about 10 seconds long and I speed it up, so I timestretch it a bit. Then I want to pitch it up a bit.. and after I do that the quality goes way down and it sounds like its distorting. Has anyone else had this problem?

This will always happen since the machine is manipulating the sound directly. It's like fat women trying to squeeze their flabby feet into a pair of stilletos. Looks like they're baking bread or something. Anyway, their feet look all f*cked up. Same thing happnes to your samples.

Maybe you should try cutting the toes off your sample, or cutting your feet up into equal portions. That way, when you time-stretch and pitch shift, the effect may not be so obvious. I've never tried it to be honest with you. My footwear selections are always well-considered.

Good luck!

Dolo wrote:This will always happen since the machine is manipulating the sound directly. It's like fat women trying to squeeze their flabby feet into a pair of stilletos. Looks like they're baking bread or something. Anyway, their feet look all f*cked up. Same thing happnes to your samples.

Maybe you should try cutting the toes off your sample, or cutting your feet up into equal portions. That way, when you time-stretch and pitch shift, the effect may not be so obvious. I've never tried it to be honest with you. My footwear selections are always well-considered.

Good luck!

lolo. so its better to chop my samples before i timestretch if thats what i understood

Dolo... you're analogy wonders and amazes me... but possibly not in a good way.

To the OP... I think what you probably need to know is that time stretching (playing the sample slower) is a process of making something out of nothing, and so is pitch shifting (down). The MPC or any other sampler has to take a best guess at what should be filled in to all of those extra samples that you are asking it to create in order for the output to ultimately still be 44.1khz at 16bit.

Going the other way, time compression (still called time stretching, but making the sampler play faster) is a process of throwing away, and so is pitch shifting (up). The MPC or any other sampler has to take a best guess at what information is useless and can get tossed from all of the samples that will be played in order for the output to ultimately still be 44.1khz at 16bit.

Technically it is quite a bit more complicated than that, but to give you a good explanation... you will tend to want to play a sample faster through either time stretching or pitch shifting up. This will sound much more natural than time stretching to make a sound slower or pitch shifting down. I'm not saying you shouldn't do any time stretching or pitch shift down as it can produce some amazing artifacts, but you need to understand the concepts of what is happening before you can really take proper advantage of it in a piece of music.

You can go either way, but under most circumstances you don't want to pitch shift/time stretch up or down much past a few semitones anyways. The more you go up or down from the base sample the more data the MPC (or other sampler) will either have to make up on it's own or throw away.

Hope that helps...

EDIT: One last comment... given all of this info... This is typically why you'll see people record off of records at a slower RPM, it allows them to play the sample at a much slower speed if they chose, and they can still play it at the "correct" speed and it will still sound natural. It takes a reasonable understanding of math and octaves to really get what is going on, but if you poke around the forums or google I'm pretty confident you'll find some very useful info.

Thanks guys. I don't know what got into me today.....

Anyways, yes, stupid analogies aside, perhaps chopping up your 10 second sample into smaller slices may remedy the problem somewhat.

Also, why are you time-stretching and pitch-shifting the same sample? Try to get the sample you're using closest to the bechmark that you're matching it to.....

i.e. sample the record playing either pretty close to the tempo you need it at (and then pitch shift), OR sample the record so it's as close as possible to the pitch you want it (and then time-stretch).

I don't know if either of these will work, but just a thought! My guess is the less you manipulate the sample in a digital sense, the better it will sound during playback. Unless you want to go all Glitchy and sign to Warp Records???

Just to put this out there--

If you really need to timestretch something and keep the quality nice, I find that most computer programs will do a much better job than my MPC1000. I've been rocking an old version of Sony Soundforge for timestretch and pitchshift for years, which I definitely endorse if you have a computer.