Tempo is also one of the best expressive devices for performance. The mind massaging works in reverse! Lots of great choruses bump the bpm significantly, and the prevalence of recording and performing to click tracks has made it rare to have the kind of fluidity that you can get without them. You want to be able to manipulate your internal metronome expressively. I'm not a drummer, but definitely something to work on for every musician.
the prevalence of recording and performing to click tracks has made it rare to have the kind of fluidity that you can get without them.
Performing to a click is not nearly as lifeless as a hard-quantized MIDI production without live instruments. Sure, it may not have drastic tempo changes like a drummer who isn’t listening to a click, but it still has human variation. Making MIDI programming sound convincing and realistic is not an easy task
" Sure, it may not have drastic tempo changes like a drummer who isn’t listening to a click, but it still has human variation. Making MIDI programming sound convincing and realistic is not an easy task"
Not knowing anything about music at all I've wondered about this a few times. Isn't it just a matter of adding some stochasticity from an appropriate distribution? What is it that makes human imperfection different from machine generated imperfection?
At the best of my crappy math abilities, the way we feel about movements (let's say music is a movement) is full of layers of continuity. Simple randomness won't fit the bill IMO.
When I play I think about diffeq and nurbs .. but that's a highly subjective analogy.
Think about what makes a groove or feel. You can play quarter notes that feels heavy metal or funk, just by altering the impact and accent. Also I believe that our ears perceive a lot more than discrete events. A single hit has a duration and is more like a tiny bell curve than a zero width line on a chart. Same for singing, if you hit the right note at the right time nobody will like it, it's all in the way you get there in between, even if you're off a little, the subtleties are there to make it pleasing to our ears.
That is beyond my level of math understanding, haha.
In addition to differences in the performance data (MiDI) versus an audio recording of someone playing a real instrument, I find that MIDI instruments do not sound as realistic as a recording of someone playing acoustic drums or a real piano. Perhaps that is because VSTi plugins usually sample perfect hits and close-miced drums; while a recording of someone playing live drums might have many more imperfections and room noise... I don’t know. It’s certainly possible to program realistic parts with MIDI, but they always sound much more polished to me than real recordings.
>It's no secret that every drum machine has its own vibe - whether analog or digital, they rush, drag, swing, shuffle, and funk like no other. Some are extremely accurate, and others are completely off the grid. An MPC60 will sequence your sounds like a band is playing them, and the LM1 (responsible for the first sequencer shuffle of all time) will give you that incredible early 80s Prince feel. Whatever the machine, they all have a unique groove.
There are ways similar to what you've described and certainly doesnt even have to be that clever. Just adding a small random offset using uniform white noise helps.
People go to immense ends to get their programmed drums to sound good though - and the techniques that are coupled with this go beyond a statistical approach.
tldr: there's been some research into it – turns out that if you have a few parts (instruments), the randomness you add to each of them needs to be correlated:
> "The first [example] has had completely random timing errors inserted [...] with no link between the errors in different parts. The result sounds unmistakably unmusical and inhuman."
someone built a "humanizing" plugin for Ableton Live based on that research.
here's some quotes i liked:
> "[...] the timing of each individual note is dependent on every single note that both players had already played – a minor timing hiccup near the start of a piece will continue to affect every single note after it, up to the last notes. And when you play a duet every note your partner plays affects your playing, and every note you play affects your partner [...]"
> "[...] if everything is recorded together in the same take then quite large variations in timing are no problem – they don't sound like errors, just the natural movement of the music. But if the parts are multi-tracked, or sequenced parts are mixed with human parts, then the timing errors are glaringly obvious, they sound wrong because they are unnatural, and our capability to identify the uncanny marks them out as unpleasant and undesirable [...]"
About the correlation between musicians. I don't know if you ever played drums in a band. But there's such a weird coupling between everybody when the rhythm is solid. A few ms off from the drums and everybody in the room will have a hiccup.
Not related to the content of your comment itself, but congratulations on owning the unique Hacker News comment whose numerical ID corresponds to the YYYYMMDD date of its posting:
A classic example is Roxanne by The Police. ~132bpm verse, ~138bpm chorus by my rough measurement. You can really feel the rhythm section wind down after the first chorus, which wouldn’t have worked if the guitar hadn’t waited two bars to come in and resolve the shifting tempo with precision jabs right as the band hits the original pace. Great moment.
