Reply to thread

Message: [QUOTE="Shamiso, post: 92667, member: 160"] Bitcrusher jitter moves the timing of downsampling events, while sample-rate reduction lowers the effective rate on a regular schedule. Lowering one does not secretly lower the other. Those controls can sound related because both mess with sampling, but they do different jobs. Regular downsampling repeats a predictable pattern, while jitter makes the timing less predictable and turns that steady degradation into something noisier and less locked. In the [B][URL='https://goldmidi.com/community/threads/riverflux-audio-debuted-a-free-bitlayer-multiband-bitcrusher.78061/']eight-band BITLAYER bitcrusher[/URL][/B], bit depth, sample rate, and jitter sit beside each other on every band. Treating them as three strength controls wastes the useful part, because each one attacks a different piece of the signal. [HEADING=2]Bit depth changes amplitude while jitter changes timing[/HEADING] Bit depth controls how many amplitude values the signal can use. Pull it down and smooth level changes get forced onto fewer steps, so low-level detail becomes rougher and quantization noise becomes much easier to hear. Sample-rate reduction works on the other axis. Instead of limiting available amplitude values, it holds or resamples the signal less often, lowering the effective update rate and creating imaging or aliasing when the discarded high-frequency content is allowed to fold back. Jitter changes the schedule rather than simply choosing a lower fixed rate. Ableton describes its version as noise added to the downsampler clock, while other current bitcrushers describe jitter as randomizing the downsampling frequency or accuracy. The exact implementation can vary, but the common idea is irregular timing. A classic [B][URL='https://www.sciencedirect.com/science/article/abs/pii/016516848090002X']sampling-error analysis[/URL][/B] separates quantization error from time-jitter error for the same reason. Quantization changes the represented sample value, while jitter moves the instant at which a sample is taken. A lower rate does not automatically mean a less stable rate. You can run a downsampler at a brutally low but perfectly regular interval. Jitter only enters when those intervals vary, which is why two processors set to the same nominal rate can feel different. Knowing the difference makes the usual bit depth vs sample rate argument much less fuzzy. Bit depth is mainly about amplitude resolution, fixed-rate downsampling is mainly about a coarser time grid, and jitter makes that time grid wobble instead of staying perfectly repetitive. [HEADING=2]Random timing reacts harder to fast-moving audio[/HEADING] Timing error is not equally obvious everywhere in a signal. Sampling slightly early or late matters more when the waveform is changing quickly, because a tiny move along a steep section produces a bigger value difference than the same timing move along a slow section. High-frequency material therefore tends to expose clock jitter more readily than slower material at the same timing error. Bright percussion, hats, noisy synth edges, and sharp transients can turn grainy or splashy quickly, while a smoother low-frequency part may need a stronger setting before the effect becomes obvious. This is where what jitter does in a bitcrusher starts to feel different from plain sample-rate reduction. A fixed low rate creates a stable pattern tied to the chosen rate. Randomized timing keeps disturbing that pattern, so the texture can feel less pitched, less repetitive, and more smeared around the edges. Stereo behavior can change too when a processor uses independent or decorrelated jitter paths. Ableton's current downsampler documentation specifically notes that its Jitter control can increase stereo width, which is a useful reminder not to assume every jitter knob is mono-safe or channel-linked. BITLAYER gives you per-band jitter, but its public release description does not spell out the exact jitter algorithm. Keep expectations tied to what you can hear. A high band may respond very differently from a low band even when both show the same control value. [HEADING=2]Jitter works best after the main degradation is chosen[/HEADING] Start with bit depth and sample-rate reduction doing the obvious work first. If you want coarse amplitude grit, lower the bits. If you want aliasing, metallic side tones, or old-sampler-style stepping, lower the effective sample rate until the character is already close. Bring jitter in afterward when the downsampling feels too static or too neatly pitched. Small amounts can loosen a repeating digital tone, while larger amounts can make the top end dirtier and the timing artifacts harder to predict. Per-band processing makes this easier to control. You can leave lows on a steadier clock while adding more jitter to upper percussion or synth texture, which avoids turning the entire signal into the same unstable wash. Level-match while doing it. Jitter, bit reduction, and downsampling can all change perceived brightness and density, so the louder setting can win a quick comparison even when the texture itself is worse. Pull the output back, listen to the artifact rather than the level, and stop when the randomness starts masking the part you wanted to keep. [/QUOTE]

Name