Dear all
To create sustained sounds I like to set wave playback to loop>pingpong.
Unfortunately, regardless of where I set the loop point and which wave I use, OT produces a high-pitched noise artifact upon changing playback direction.
The effect is sometimes more pronounced and sometimes less, but it is always there.
I haven’t figured out how to mitigate this issue. Does anybody know the solution to this problem?
I won’t post a file here - the issue is simple to reproduce.
Thanks to all
not sure why this is happening but i got high pitched whine noises early on when first getting to know the OT.
no one knew why there were these noises.
[edit the noises were there at the end of a loop, like if the loop was playing and looping, there would be a high cheerful chirp sound at the last note. fascinating, unexplainable, and irritating. hardly any one else had experienced this. suggestions were made regarding audio file format, etc, to no avail].
solution: i had called the hard drive a funny name and called my Set a curious title and also included a non-standard letter here and there in all the namings. things like @ & * - _
so i wiped the drive (media card) reformatted and reinstalled the OS.
called everything standard names, with nothing other than letters or numbers.
no more high pitched whine noise.
You need to think about what happens with reference to my comments in the other thread on single cycle waves. What happens when a wave is reflected. You would need both loop points to fall on the 0 of a cosine wave or the 90 of a sin for it to work in fwd and pingpong and sound continuous. Obviously depends on material. Then what do elektron do at the rebound is it a b c d e e d c b a or is it a b c d e d c b a and so on. I find it tricky to get it usable on a pickup machine. But I don’t think it means it is a bug either. If you have nothing at the end of your loops I doubt it’ll glitch at those points but it won’t sound continuous. What’s a bug and what’s just physics essentially. I was a bit moody about regular loop clicks, but even when I tried to do it manually in a software environment it was surprisingly difficult to get rid of the glitches to certain sustaining sounds. Getting the ideal length of a sample is not a given. Having said all that it does seem more usable in other soft/hardware. Esp looping pedals ! The ping pong artefacts are particularly grating though
avantronica, thanks for your message.
I understand the 0 crossing argument and I take great care to position the loop points accordingly.
Even if this is done almost perfectly, I still feel that the looping point artifacts are beyond the mere ‘click’ level, but, as you say, particularly grating.
I am not sure if this is a valid argument, but I’d say for a sample based machine like the OT, this should not happen.
avantronica, I agree with fachperson. I’m familiar with ping-pong looping from other tools (trackers, Roland S550) and the type of artifacts that can occur in the situations you describe. If this were a matter of the loop point phase not being set properly, I’d expect the discontinuity to be a short blip or click lasting no more than one or two cycles.
This is different, it’s more of a “squeak” that is very audible and sustained for about 5-10x longer than I’d expect. It also has a fairly uniform pitch - even if you change the pitch of the sample being looped it stays more or less the same but has a kind of greasy or squeaky linear frequency quality like ring modulation, downsampling, or syncing a static oscillator to a dynamic one.
I actually think it sounds kind of cool sometimes but I could see it being very annoying in some situations.
I need to make some recordings of this. I don’t have time at the moment but it was pretty easy to reproduce for me. I’m not sure if it’s something about the sample I’m using. Anyway, I just load it in, set it to ping-pong, and mess with the “length” parameter. Try adjusting other parameters to see what I mean about the pitch linearity. It’s interesting with the rate parameter because it extends the length of the glitch but its pitch is mostly intact, or at least it’s harmonics off the same base frequency. Obviously it’s more apparent on shorter samples just because you hear it more often.
EDIT: Another issue is that if you set the length short enough, it will stop looping. This is inconsistent/intermittent too, you can set it pretty short and it might play for a while but stop after maybe 20 iterations of the loop.
fwiw, i mean zero degrees on a cosine, theoretically speaking, which equates to peak, not zero amplitude, from the top whichever way you go is valid and the same, from a zero crossing point, you could go up or down at the rebound but it won’t be the same both ways !! - all this doesn’t mean that there’s not actually a bit of tidying to be done, but it’s very source material sensitive
i’ll knock up a single cycle cosine wave, that may work fwd & pingpong in theory, but will it in practice ? !!
a pic from the net >
obviously if using samples like this, you do them sample accurate and do not select snap to zero crossing when slicing !
i hear that bit grating reversal fairly clearly with pickup machines, it is horrible, i do agree - but will it do it when fed content that may well play nice when played either way - i’ll try the cosine thing out !
I really think there is a bug in the looping algorithm - it sounds like it does a micro-loop at the end. It sounds like once it gets to the end and plays backwards, it switches back to forward after just a few milliseconds, then goes backward again at the end, and repeats this for several iterations.
it’ll all be evident if it gets re-sampled and viewed in an editor, i’ll try when i get a mo
I put up a video. It’s unlisted so as not to make Elektron look bad.
Audio is quiet (phone mic) but you can still hear the artifacts pretty clearly.
This is kind of weird, but the sound actually reminds me of something. Specifically, in college, my roommate had this stereo with a 5-CD changer with a weird door. You could actually open the drawer and run your finger over the CD while it was playing. Kind of like a turntable except it was digital so instead of a smooth slow-down, you’d get all these digital error noises with a very similar quality to these glitches.
it’s definitely a bug. think it’s been reported. I believe you can actually see the marker looping a tiny section of the sample in the audio editor when it happens.
I wonder when the next update for the Octatrack will happen, can’t wait to see all those little bugs squashed !
two points to add to the debate
firstly, the shortish single cycle i tried to pingpong would not loop at all !
it would play fine in fwd direction
secondly i constructed a test tone with a very aggressive fade, then added a silence, the silence got chopped (somehow) and the glitch was very very evident even though i had content at the end of the fade that was virtually zero
more annoyingly - it was inconsistent, sometimes it would actually loop seamlessly, then the glitch would get more noticeable ! - inconsistency is especially annoying
so i agree, whatever the ins and outs of certain material - it seems to be a bug, and inconsistent 
glad someone posted this and it has been reported, I ran into it the other day and spent a ton of time trying to figure out what was going on
figured it had to be something I was doing wrong…guess not 
I had these artifacts too… but I don’t mind them, actually I like how they sound 
Please Elektron if you fix this, make the fix optional
reading that, i fear that you are actually being serious 
here’s my tip then - just sample it, then you can have it as much as you like
glitchy noises are great, but only when you want them !
if they retained the bug as an option, you could have my OT, that’s how confident i am that won’t happen
Heh. Was really surprised to hear such artifact. I’m long-time Renoise user and i use pingpong loops alot. So, after getting the OT i decided to tweak loop modes in the first place and spotted this artifact immediately 
Was this sound present in earlier firmwares?
By the way, for pingpong loops i prefer to set loop points in such fashion:
