Optimising Samples For Digitakt / Normalization

Elektron Gear Digitakt
#1

Hello,

First post here :slight_smile: I have been a long time reader and observer of this fantastic forum.

I own a Digitakt, Analog Rytm and Analog Four amongst other goodies from Vermona, Make Noise, Arturia, Teenage Engineering, Korg, MFB, Strymon, Waldorf etc. You can find some of my music here - www.soundcloud.com/haxeri

I received my Digitakt last Friday and have to say with all limitations considered it may be my favourite Elektron box. I haven’t had this much fun creating loops, mangling samples and jamming since I got an OP-1.

I apologise if my question has already been covered.

I am wondering if anyone knows the optimum level for samples when imported via Transfer to the Digitakt. I’ve played around with -3db, -6db and -1db. As I understand the sample gets normalised when uploaded not only when live samples through the inputs? It obviously needs headroom for processing in the Digitakt (pitching up and down, FX, distortion etc.)

I think it is safe to say I spend way too long when preparing samples for upload :slight_smile:

I usually load a bunch of samples into Cubase, select which channel (L or R) sounds best (rather than let the Elektron box sum the stereo file to mono) and then increase or decrease the gain of the sample to a preferred level instead of normalising.

Any advice on this would be greatly appreciated!

Also, if anyone is interested I’ve recorded some nice clean tight MFB Tänzbar samples, Elektron compatible - 16/48. Happy to share.

Thanks :slight_smile:

1 Like
Create proper samples files in Live for the DT
Transfer issue
#2

I’d be surprised if this was true, the normalization I mean. maybe sample rate correction, stereo to mono, but that could be a step too far

Easy to test anyway, radically reduce the level of one of two similar samples and compare the outcome once inside

The volume of the samples can obviously be adjusted by levels once inside too, i’d stick with levels that you may find most commonly used in other samples, so you can work from the same standpoint with old and new content … i’d have thought -3dB would be perfect, -6 would possibly be too much unless you want to adjust other content down … plus I suppose it depends on what level any automatic sampling normalization comes in at, so if the DT normalizes down a hot signal by -12dB, you may want to avoid the level mismatch of imported vs sampled by importing lower than you’d normally go

just a thought, but I guess most samples you find lying around the place are in the -3dB ballpark … I suppose it also depends on the nature of the material, whether it’s likely to be stacked and sustained or isolated and brief

1 Like
#3

Hi avantronica,

Thanks for your quick reply :slight_smile:

I agree with you about that if samples are normalised when uploaded via Transfer it is a step too far. I’d ideally like the samples to peak at whatever level they are at source. I read somewhere online that samples are normalised on upload but maybe the writer was mistaking live sampling through the inputs.

Thanks for the advice. I will uploaded a bunch of the same samples at different peak levels and see what the results yield. I guess this is a pretty quick way to check if samples do indeed get normalised on upload.

Yeah -3dB sounds like it would be the obvious level.

Cheers.

1 Like
#4

I set out to figure out where Digitakt normalizes samples. I’d see some other discussion of what level to prepare samples for DT, with some people suggesting -1.5dbFS or -3 would be best. I figured I’d learn what Elektron thought was best on the unit itself…

I did was produced a test tone at 0dbFS on my computer, and imported that. I also played that out my computer into the DT inputs, with lots of headroom so I could learn where it would normalize to. I sampled it on the DT. Now I compared the two samples on the DT by playing them out tracks set identically.

Result: DT normalizes input samples so the peak is 0dbFS.

I think this makes a good case for where to level samples for DT import.

3 Likes
#5

Interesting post guys!

I have some (stereo) samples from thatsound I want to import into digitakt
What would you advise? just let the digitakt convert them?

Or convert them myself, to mono 16 bit, 48khz, -3db, and then import them using elektron transfer?

or something else?

#6

convert them into left/right samples then import them.
use 2 x tracks with each using same sound setup but panned left/right.

if theyre not too different , use 1 sample and setup the filters slightly differently , use the lfo to move them around from left/right at slightly different rates.

also , depending on the sample , double the speed and on DT play -12 , ive found very good results doing this and it saves lots of memory , not a new technique but can work well.

4 Likes
#7

Why not 0db? They are going to get attenuated by the DT sound chain anyway - so why not start with as much as you can get? Go for normalizing to 0dbFS.

1 Like
#8

Thanks! I converted, mixed, and normalized them using izotope rx6 and they sound better on digitakt now!

3 Likes
#10

Hey, I’d love to have your Tanzbär samples on my Digitakt.
I hope you still have em
best regards

1 Like
#11

Hi Guys.
Ghia is pretty old topic but I am already bought Digitakt and have similar doubts about samples quality. I love elektron gear but sound is still makes me headache even when use official samples pack from store. Maybe I expect too much but I want hear best possible quality.

Can you advise what is best and fastest way to convert samples and just to sound it apparently better ?Do you recomend any free app to use? I am not profesionalist and do not have budget to invest hundreds of bucks for single app.

Thank you Elektronauts