So, I've been really interested in sampling for a while, and I've been watching tons and tons of tutorials on how to sample and do this and that.

But I'm STILL wondering:

How do I get a sample to be isolated? I mean, lets say I want only the song from a sample (take Skrillex for example, he tends to take a few bits here and there from audio interviews, random phrases from youtube, resampling, transposing it and autotuning and processing it etc etc), how do you actually do this in a proffessional way? I mean, in interviews and stuff ,there's always some crack in the background, Skrillex manage to get a away from this, question is HOW?

I'm also wondering, when sampling, you have no idea what tempo the sample is in, since it's no song (lets say from an interview, just pure talk or scream), lets say a guy is talking reaaally slow, like 90bpm, and I want it up to 128bpm to match my track ,or in even extremer cases like skrillex's dubstep: 149bpm, how do I get it to fit good with the tempo without sounding too "speedy"?

This is another thing Skrillex manages to get good.

I think sampling in general is really interesting, and I would kill to get better at it. Any tips, tricks, tutorials or so on this? I've checked youtube alot for this, but I can't seem to find any good ones that come up with cool tricks and tips and isolations etc to make it sound professional. Most of the guys on youtube just show the basics on how to do this.

I also got FXpansion Geist which I think is awsome for drumbeats, should I use this as a sampler for whole tracks aswell and interviews to pick up parts etc? or should I just stick to the Ableton live 8 method? or is it better to use Logic for this?

So many questions, so much interesting stuff, so little answers.

A helping hand here is very much appreciated.

Thanks alot!

Skrillex? really? how many of these topics do we get?

Okay, I'm going to give a short (helpful) answer.
If it's straight talking, learn syllable phrasing. Just like when someone sings each word or syllables of each word have a "pulse" to them
Since talking would sound weird if you broke up a word in the middle of it, You would need to put the first "strong" letter (words like fIght, the I is the strongest part of the word) on the beat, and if the word has multiple syllables, then try you best to time stretch the word so that syllables land on beats no shorter than a 16th

oh well it doesn't necessarily have to be Skrillex, I just mentioned him since he's one of the electronic artists that sample really good today, using it in a creative form.

I think you misunderstood me, or I simply was unclear. I know how to get a perfect loop going, be creative with the syllables and stuff, it's the actual processing and the sampling I need help with.

I appreciate that you're trying to help, really (don't get me wrong), but try to read again. I'm not a good sampler, but the words and stuff I think I can handle atleast pretty decent good tip though! will keep in mind.

Sounds like to me you were asking how to get a "talking" person into beat.

I should mention that styles like dumbstep are in cut time. meaning they are played at 140BPM (or close to that) but have a 70BPM feel meaning, you can sync the person talking by lowering their speed rather than increasing it.

ah yeah, but I'm trying to recreate this, but in house instead.

And when it's in "housey form", if I slow it down, it's getting reeeeally slow.

yep, dubstep is slow because it's not consisting of a 4/4 kickdrum. Makes the impression it's going slower than it actually is.

Also, how to get samples THAT clean?!

various effects and tools and clean up audio.
EQ's are usually your best friend though. Perhaps use a unique EQ on top of a regular parametric EQ, for example, using a tube-emulating EQ, like Pultronic Tube EQ and use it's "vintage" or "12AX" (or whatever) tube that has a lot of brightness to it, bump the warmth and the brilliance and it will probably restore a lot of the lost upper range that you find on small portable video cameras (I say this because I am familiar Skrillex likes to sample Youtube).
Only problem with doing it that way is that it might boost frequencies that didn't need to be, so you would use the parametric EQ to knock those frequencies back where they should be. Also, the Tube-EQ might not ENTIRELY restore the lost upper range, so you would want to use the parametric to boost the restored frequencies.
Washing things out in a little reverb helps too. Find a nice "smooth" sounding reverb (like what you get from TAL's reverb plugins) with a mid-sized room (you'll have to play around with this) and then play around with the dry/wet. You're looking for something that smooths over any harsh peaks in the audio, not to make it sound like the people are in a live room, so chances are you'll want that wet mix pretty low.
Lastly, just because you "smoothed" a harsh peak in the wave form, doesn't mean it's gone, and you'll probably notice that, so the next thing you'd want to do is normalize the volume with either a limiter or a compressor. I recommend GClip because it's free and gives you a visual aid of the wave form as you "limit" it, not to mention it has a knob to "smooth" out any distortion that is caused by over-limited a some peaks.
Some other effects and tools might help as well, it depends on what you are sampling.

OK so here's an example of what I ACTUALLY mean:

@ 0.25 seconds in this clip they sing "come on baby light my fire":

http://www.youtube.com/watch?v=OdV6Q-WQFeU&feature=fvst

Skrillex made it so it sounds exactly the same as the "it's alright" voice, WITH the same voice, aswell as "LOBBY BITCH" or whatever they say in the end.. same voice.

This is what I'm wondering about, AND, the instruments in the background is gone, you cannot hear them through the mix at all.

I think "it's alright" and "LOBBY BITCH" (or whatever they say haha, cant even hear what they say!), is form an interview with Jim Morrison:

http://www.youtube.com/watch?v=dAssS1VbEZE here.

HOW to get the same voice from different samples, AND removing the stuff and noise in the background?

THIS stuff right here is some awsome tips! thank you very much man! Really appreciate the brief answers you're giving. This will help me so much with the sampling.

I'm actually thinking of doing just like Skrillex, sampling it straight from youtube, things from spotify, maybe other various clips from various sites.

The question that still remains though is how I can sample from different clips, but get the same voice on it? and removing the instruments etc in the background?

This is a tough one. And unless you have some REALLY high end tools, I am not familiar with any good ways of removing music from vocals.
There was one tool - forgot what it was called - that is like the "photoshop" of music where it displays the sonogram of the audio and you can use a lasso tool to select the sound you want to solo, remove, or clean.
But, if I remember correctly, it cost some couple of grand.

Otherwise, you'll have to deal with stems. Google searching "[song name] remix stems" or "[song name] a Capella" (or of the likes) might turn up some results, but it's kind of touch and go, really.
Mostly you'll only find stems from modern popular songs - you know, the ones you hear on the radio. Though a couple of times I have ran into stems for Queen, Nirvana, Katrina and the Waves, and surprisingly Royksopp (which is hardly a big name band)

There are some cheaper tools to remove the a capella but none of them - from what I have seen - have given good results. Usually A LOT of the background music still remains and still remains at a fairly high volume (so you can't exactly mask it with your music) and they usually remove important frequencies which is good to remove the bass and some of the drums, but it also removes some of the body of the vocals.

that's the really tricky part... I mean lets say I want to sample from youtube or from some other kind of media, there are no stems for that..

and honestly I really doubt that Skrillex was using, or even is using any tool that cleans this up, I think it's all about the processing inside of the DAW, but I can't figure out how.. kinda frustrating, because this is what you really need when working with samples. A good way to isolate stuff and sections.


Also, there seems to be no explanation for WHY he uses so much random stuff, he seems to just take a bit from and here and there and just put it in and repitching it etc. So bascially it doesn't matter what you sample at all? aslong as it's creative and unique and fits into the mix, then it's game?

There are some audio editors that might be available to help you get close to what you want and then from there you could use effects to get the final bit.
I am unfamiliar with what all Melodyne does, though I think it's mostly for Vocals, I think you'd want an Audio editor similar to that.

And again, to go back to what I said before, for videos from Youtube, you'll need to just use an EQ or two. They'll be your best friends.
Pop an EQ in plus a Spectrum Analyzer and listen for the sound sounds you don't want.
Wind, for example. Sound like an upper-mid sound when it comes through those tiny tinny microphones on handy cams, so look in the analyzer and see if you can see the wind in the frequencies, if you can, work delicately with a small bandwith ("q") and try to remove it's frequencies without reducing the body of the voice. You'll have to play around A LOT with this until you can find the sweet spot.
This of course will leave holes in the frequencies where you did the cuts to remove the wind which is where that reverb will come in handy as well. it will smooth over those holes if you add a dash of a few different effects as well. depending on the song style, you might be able to get away with using a little bit of distortion on the voice, it will be noticeable, but might work with the song, and if it does, just leave only a little distortion, try to use it to reduce the size of the hole from the EQ.
One thing you ought to try, and I'm not sure if this will work, so play around with this, is layer the sample twice.
One layer will be the main one everyone hears while the second will act as a way to fill in the frequency holes.
Lower it's volume A LOT when compared to the main sample and try washing it out in reverb. We're talking like 100% wet, 10% dry, then use a gate to cut off it's resonance so it doesn't linger around too long.
Try that see if that helps. It's just an idea, not sure how well it will sound.

AWSOME stuff and techniques here! I will try to apply them and see if it works out, a million thanks!!!

If anyone else, or if you hear about some good audio editing program where I can fine tune and remove stuff in the background, please let me know! thanks alot man!!! really appreciate all this to the max.

This might be the software ntom was referring to Izotope RX2. When i went to the site to get the link this Skrillex interview happened to be on the front page also so i thought i'd post it for you Skrillex Interview. As for giving you tips, well i haven't got a clue but it's probably a lot of work. The best story i've read about such a thing is when Junkie XL did the Elvis remix, all he had to work with was a mono recording (he did also have a different stereo mix but the mono had more vibe apparently) and he explains how he did it in a fair amount of depth and it was indeed a lot of work. He had to cut up the audio into lots of parts crossfade/eq and compress each part, i'll see if i can find it as i read it in a magazine but it should be online somewere.

whoa griffin, this helped alot! I'm getting RX2 straight away and have a mess about! seems like an amazing plugin, abit TOO amazing?! thanks alot for the info! read the whole interview aswell, interesting stuff. After that reading I think I'm getting Trash too, lol. And if you find that interview, please link! very interesting stuff.

EQing
Mashtactic
RubberFilter
and very very careful 'tops and tails' on samples.

Just that will get you very very far.