First post here, so I hope I don’t make any newbie mistake :-p
I have recently bought my first Elektron device (Octatrack) and must say that I’m very enthusiastic about it! I hesitated a long time because of the ‘complexity-rumours’. It took some time to read the full manual, but it’s such a great document that it makes the Octatrack very accessible for new users.
Anyhow, to the point: I am trying to sample for the first time, and am facing a dilemma: my sample source is an old drum machine, which produces a lot of background noise. So I’m wondering if i should sample in the Octatrack with the volume of the external device set to maximum (which also increases the noise level) Or should I decrease the volume a bit on the drum machine, so that I get less noise in the incoming signal. And then of course add some additional gain with the Octatrack afterwards to make up for the more quit input?
I’m afraid that it won’t make much difference which way i choose, because if I add more gain afterwards in the Octatrack, the noise will also increase again. So basically the signal to noise ratio will always remain no matter how i sample?
Thanks for your fast replies!
I tested the noise gate and that’s a nice tip! I set it to about -70dB and it cuts all the noise. But i guess it doesn’t impact my signal to noise ratio. It just cuts the noise whenever no sample is playing.
I will try to do both and compare. My main fear is that i don’t trust my ears :-p
Does ‘no clipping’ in the OT means that the amplitude doesn’t reach the top/bottom of the AED screen?
It’s just being careful not to overload the inputs with too loud a signal coming in. There’s a green/yellow/red LED on the OT that you want to not be red, but 3 flashing colours isn’t the most sensitive input level monitor.
Really, it’s just something to be aware of if you want to make clean recordings.
And yes, have a look at the waveform after to see if it looks good, give it a listen and I’m sure you’ll do some great things with that old drum box
And yes - try all sorts - even record it one take where you tweak the volume and slice it & normalise the slices individually… Some of my favourite slice files are the same sound repeated from the blofeld with the patch decreasing volume until it’s almost overtaken but digital noise. Then slice and normalise each slice so that I can graduate noise ratio while playing chromatic. I do this with volca keys too - the noise on that has different character with each semitone
Slight diversion but this is the kind of stuff that can be a bit irresistible on OT
Thanks for all the tips! I will continue experimenting and see what sounds best.
Just wondering one more thing: doesn’t normalising change the overall ‘sound feel’?
In case you didn’t notice yet: I’m also new in the world of sampling
if u don’t trust ur ears, trust ur eyes in this case…u can always see if something sampled came in too hot, if u check the sample in ot’s audio editor…clipped audio has obvious hard flatted/cutted top and end peaks in it’s curves…
but in general, the mk1 does likes green flashing on it’s input indicator led’s only, while the mk2 is fine when these indicator led’s sometimes flash a LITTLE orange and even rare red…if u end in red only u’ve gone too far…
many gear, old stuff in special, adds more noise on the last 10% of output gain…
proper gainstaging is key to everythig in the end and it always starts at the signal sources…
so full drilled output might be fine, but the danger of additional noise comes into play…while a little less than all too the max is always fine…
ot does not add serious noise…not on the way in, not on the way out…
even if u read the manual, and ur one of the few mentioning in it in such a positive way ;-),
keep always in mind that the main volume output knob is only adressing ur actual headphone levels…i kept on wondering, even knowing, many times in the beginning…
gainstaging is complex inside of the ot…amp page, atributes page, main volume, track volume…
and another little thing, if u don’t trust ur ears…even if u would, ur brain will always tell u in natural born reflex, the louder signal sounds better at first glance…
so normalizing is really just lifting the signal to overall max peak…which won’t do anything if ur signal has already reached max peak/digital clip due to a too hot sampling process…
never hesitate to experiment and test things by trial and error…for now, there is no right or wrong…only fooling around with lucky dips and happy accidents…u got a great sonic safari waiting for u…enjoy the funride…
Thanks for the detailed answer! Very helpful!
I don’t see any clipping so I guess I’m good Should I always aim for the level ‘as loud as possible without clipping’ when I’m sampling?
On the last 10%: I didn’t notice relevant increase in noise, but didn’t notice any increase output neither. So I didn’t take the risk and just went for about 80% - 90% of the max volume.
It’s good practice to leave a little bit of headroom. I leave 1dB but I’m thinking I might push that up to 6. Obviously, everybody wants their samples to be at the same level and for that to be exactly loud enough, but if you make a practice of normalising everything all the way to 0dB, you can expect to get inter-sample peaks.
Without taking a deep dive, it can help to understand a little bit about how digital audio works. Audio files are basically lists of coordinates for drawing curves. It’s a big game of join the dots. When you play an audio file, you don’t hear steppy, bitty dots, because playback devices take a look at all of the discrete points and draw long, wavey, wiggly, continuous lines that make it all fit. These wiggles are how your speakers vibrate. Inter-sample peaks are where the curves go over the edge of the paper, so the peaks get chopped off and the lines get broken, which can happen even if the plotted coordinates are within limits.
It’s easy to boost a clean recording. Like @reeloy says, gainstaging is the thing.
Also, it’s not the end of the world if you get some clipping. Just something to be considered.
As for what’s stored where, the OT keeps a little preference file with volume settings etc., and the .wav file has its own volume baked in.
Yep, I usually look for orange flickering on peaks, occasional red is ok.
Sample attributes have their own gain parameter, set to +12 by default.
If you open a sample in the audio editor, fx1 button brings you to attributes, remember to save your changes. Fx2 button brings you to file operations.
It is stored at slot level if modified. You can also save sample settings, which modify/create a file (sample.ot) associated with the sample (sample.wav). This stores Trim/Slices/Attributes settings, and can be recalled.
