Timing Jitters - esp on 16th hi-hat sounds

Help, Support & Bugs
1

Idk, am I imagining, or is it real, but…

During playback, I can hear a timing jitter - especially on 16th hi-hat type of sounds.

Both in real-time song playback and in rendered wav. Both Asio(Asio4All) and Windows DirectSound and WASAPI

It should all be ‘ts-ts-ts-ts-ts’, but occasionally it goes something like ‘ts-ts-ts-tr-ts-rs’. Sounds like it skips something for a millisecond.

I have included an example song and rendered file(inside the song, second sample).

Windows 10, Lenovo T460, Built-in sound-card. Renoise 3.4.2.
rendersynchrotest.xrns (2.7 MB)

Edit: the original file had some effects and mastering options activated - I deleted/deactivated them all - it is now just one track raw:
rendersynchrotest02.xrns (2.8 MB)

2

It’s subtle but it does sound like there’s an “analog flavor” going on…

I ran it through an oscilloscope (Psyscope) and the waveform does change ever so slightly in Renoise for every run but same experiment in Live and Reaper it stays the same !?

3

Idk about the analog, it’s just the sample and that’s it.

And I can hear the issue both during live real-time play and in rendered wav.

4

Yeah, does have a nice analog feel. There are clear, cyclic, mysterious, fluctuations around 5.1k.

5

So… my laptop is possessed :wink: :smiley:

Maybe it’s the CPU/Bus throttling that does it?

6

spectrum-view
spectrum-view1116×508 171 KB

The spectrum view

7

It’s perhaps a kind of chorus flanging effect caused by the jitter

8

if you use a sinewave instead of the hat you can with Psyscope see there’s a pulsing pattern to the change

1 Like
9

A line in the tracker grid may occasionally be 1 sample longer if the sample rate and tempo do not match the LPB settings perfectly. Otherwise it would run out of time when running for a while (just like leap years are added).

e.g.

SamplesPerBeat = SampleRate * 60.0 / BPM;
SamplesPerLine = (SamplesPerBeat / LPB);

// with SampleRate 44100, LPB 8 and BPM 120:
// -> SamplesPerLine = 2756.25
// So a line is 2756 samples by default and every 4th line one sample longer: 
// 2757 to add up the fraction.

So this is not random, it’s a fixed pattern.

Similar things will happen in other daws with beat timing though, as they also must round down the (fractional) beat times to (integer) sample times at some points. The “leap” patterns may be different though.

Analog hardware does not suffer from this :slight_smile:

4 Likes
10

Please have a look at the video below where just changing the BPM turns the sinewave plucky and becomes louder. Is this expected behavior?

I’m sorry the sound wasn’t captured but here’s the XRNS
timing jitter v1b.xrns (59.9 KB)

11

Thanks for the clarification!

So… are there some BPMs that are less jitter prone to a given Sample Rate and LPBs? Something like in this table?

renoise-bpm-spl
renoise-bpm-spl1192×868 128 KB