Reply to thread

Message: [QUOTE="Shamiso, post: 92503, member: 160"] Tiefkraft uses both multiple velocity layers and round-robin sampling to vary how its sampled basses respond to MIDI. Those features often get lumped together as “humanization,” even though they fix different problems. Velocity layers change what happens when you play harder or softer, while round-robin changes which recorded takes fire when you repeat a similar note. Easy mistake. When you edit MIDI instead of playing the part live, [B][URL='https://goldmidi.com/community/threads/pfundstein-introduced-tiefkraft-eight-bass-plugin.77886/']Tiefkraft’s sampled bass collection[/URL][/B] can respond very differently to two edits that both look like harmless variation in the piano roll. Push velocity far enough, and you may reach another recorded dynamic layer, while repeated notes can already rotate through alternate samples without needing that shove. [HEADING=2]Velocity layers can change the whole attack[/HEADING] A velocity layer is a separate recording made at a different playing strength and mapped to part of the MIDI velocity range. On a sampled bass, a soft fingered note is not simply a loud note turned down. Attack, fret noise, harmonic balance, and sustain can all shift when the player hits the string harder. Blanket velocity randomization can therefore make a good bass line worse. Move a note far enough, and the sampler may leave one recorded layer and trigger another with a noticeably different attack. An edit meant to add tiny variation can become an accent, a harder pluck, or a tonal jump that the phrase never needed. MIDI velocity numbers also do not promise the same acoustic result from one instrument to another. One library might divide its layers across broad ranges, another may use narrower zones or crossfades, and velocity 90 can feel completely different after you switch patches. Your ears matter more than a supposedly natural spread of values. [HEADING=2]Round robins stop copied attacks giving the game away[/HEADING] Round-robin sampling keeps several recordings of the same basic note and playing condition, then rotates among them as that note repeats. A real bassist does not reproduce the same attack, finger noise, and decay every time a fret gets played. Reusing one recording makes those tiny details repeat as well. Repeated eighth notes expose it fast. The MIDI timing can be solid, and the velocities can make musical sense, yet one identical transient firing again and again still sounds mechanical. Round robins tackle that repetition without forcing the phrase to lurch through different dynamic levels. You can leave four repeated notes at similar velocities and still get variation from the underlying takes. This is the bit generic “randomize your velocities” advice tends to miss. If the problem is copied attacks, changing playing intensity is a clumsy fix because it may solve repetition by creating a completely different performance problem. The beginning of the note deserves attention here. [B][URL='https://doi.org/10.18154/RWTH-CONV-239465']Work on attack transients and timbral brightness[/URL][/B] shows that onset characteristics can influence brightness judgments even when sounds have similar spectral-energy measures. Cycling genuinely different recorded attacks changes more than level, which is exactly why repeated notes can feel less stamped out. [HEADING=2]Humanization is cleaner when each control has one job[/HEADING] Shape velocity for musical emphasis first. Strong beats, pickups, ghosted notes, and phrase endings can use different strengths where the line calls for them. Keep repeated notes in roughly the same dynamic neighborhood when they are supposed to feel equally strong. Let round robins handle much of the variation inside that neighborhood. Four steady notes do not need four wildly different velocity values just to avoid repetition. Alternate recordings can supply the small changes while velocity keeps the intended dynamic contour intact. Timing belongs in its own lane too. Moving a bass note slightly early or late changes groove, not playing strength, so edit it because the rhythm needs a push or pull. Random timing, random velocity, and random note length piled onto every event usually sound less human, not more. Fast passages make the separation especially useful. Decide which notes deserve accents, listen for repeated attacks that still feel copied, then touch timing only where the groove needs help. A stubborn repeated-note figure can stay dynamically steady while the sample engine supplies take-to-take variation underneath it. One practical benefit shows up when the arrangement changes. If a chorus needs more aggression, you can raise selected velocities and deliberately move into harder-sounding layers without wrecking the round-robin behavior that was already keeping repetitions varied. The edit has a musical reason instead of being random motion for its own sake. [/QUOTE]

Name