Such a thing as a CV "humaniser"?

[chimp_spanner]

Hey everyone! So I'm looking to great a sort of loose, organic, staggered feeling between my drum machines. I wondered if there's a way I could introduce small, random amounts of delay to a CV signal in order to do this? Any suggestions? Thanks!

[Loque]

I think there is a Player from the Panda that can help. You can try CV delays like Lolth or a real delay with cv to audio and back. But all works only forward. A real humanizer forward and backward with shuffle would be interesting, for cv and as a player.

[Electric-Metal]

First thing that comes to mind is using this nice little RE

https://shop.propellerheads.com/rack-ex ... -splitter/

and slightly modulate the delay amount/time with some random LFOs.

[guitfnky]

I took a whack at trying to build something like this earlier this year. it works fairly well, but requires a number of paid REs. it adds adjustable, small amounts of randomization to both note velocities and timing. also some conditional probabilities.

viewtopic.php?f=4&t=7506591&hilit=behemoth

[Loque]

You can try this one here viewtopic.php?p=393976#p393976

It adds random strumming for chords. Can't remember exactly, but it should work for single notes too.

[Auryn]

I've also been playing with this idea, although I've yet to find a proper solution for randomizing the timing. Actually robotic bean's acoustic clapper RE does this with the claps, so I've considered asking him to slap on some CV outs on that device just for this purpose

[Zac]

https://shop.propellerheads.com/rack-extension/later/

Later is a cv delay by The Chronologists, their Randrome RE might be interesting too.

[dvdrtldg]

I took a whack at trying to build something like this earlier this year. it works fairly well, but requires a number of paid REs. it adds adjustable, small amounts of randomization to both note velocities and timing. also some conditional probabilities.

viewtopic.php?f=4&t=7506591&hilit=behemoth

Holy shit that's great

[pushedbutton]

What about the built in 'ReGroove Mixer'?

[Prints]

Ableton has a randomizer built in that only requires adjusting the percentage desired.

FL Studio has a shift control (that shifts the midi note forward) that can easily be modulated by a synced random algorithm using one of its built in modulation plugins. It’s super easy to just shift all the notes backwards (or forwards) to adjust where the randomization lands on the grid.

I use this feature (subtlety) all the time with my hihats and snares. It was really disappointing for me to not find an easy way of doing this in Reason. If I were to try to do this in Reason again now, I’d try to maybe find a way of modulating this RE:

https://shop.propellerheads.com/rack-ex ... ple-delay/

[selig]

I took a whack at trying to build something like this earlier this year. it works fairly well, but requires a number of paid REs. it adds adjustable, small amounts of randomization to both note velocities and timing. also some conditional probabilities.

viewtopic.php?f=4&t=7506591&hilit=behemoth

I would think that adding probabilities is a must if you’re looking to “humanize”, because while no human is “perfect”, it’s also true that no human is “random”.

Random timing just sounds “random” to me. Drummers (like myself) are far from perfect, speeding up at certain points and pushing one part of the kit while pulling another (all the while “thinking” we’re playing tight).

So a true “humanized” likely requires machine learning to include characteristics and probabilities of real drummers, if your intent is to get your programmed drums to sound more “human”.

So, apologies for not have a product to suggest, maybe someone else knows of such a device? As for me, I simply play the parts in live - instant “humanize”!


Sent from some crappy device using Tapatalk

[guitfnky]

I took a whack at trying to build something like this earlier this year. it works fairly well, but requires a number of paid REs. it adds adjustable, small amounts of randomization to both note velocities and timing. also some conditional probabilities.

viewtopic.php?f=4&t=7506591&hilit=behemoth

I would think that adding probabilities is a must if you’re looking to “humanize”, because while no human is “perfect”, it’s also true that no human is “random”.

Random timing just sounds “random” to me. Drummers (like myself) are far from perfect, speeding up at certain points and pushing one part of the kit while pulling another (all the while “thinking” we’re playing tight).

So a true “humanized” likely requires machine learning to include characteristics and probabilities of real drummers, if your intent is to get your programmed drums to sound more “human”.

So, apologies for not have a product to suggest, maybe someone else knows of such a device? As for me, I simply play the parts in live - instant “humanize”!


Sent from some crappy device using Tapatalk

out of curiosity, did you look at what I built, or at the explanation?

the reason I ask is, I think you’ve made a similar comment in response to this combi before, and I vaguely remember responding in much the same way then. your response seems (and I could be wrong about this; it’s just how it seems to me) like you’re presupposing what it does without looking at it to understand how it’s built/how it works.

anyhoo, I took all of those things into consideration (to the extent I could) in the build. you can push and pull the kick ahead of or behind the other kit pieces. the random timing isn’t just random. it used a slow sine as a sort of “carrier” that provides the biggest variation in timing, but it’s further modulated by a smooth random LFO. the result would look like a distorted sin wave, if you were to look at it with a CV analysis tool.

the end result is the timing of each kit piece slowly moves ahead of and behind the beat, and back, with a bit of micro-randomness on top. similar to—as you suggest—the performance of a real drummer, gradually pushing and pulling the beat, alternatively.

I think you’re right that there’d need to be some sort of advanced AI algorithms to really do it well, but barring that, with the tools available, this is about as close as I can get. and why bother at all? for the same reason you suggest in your second last paragraph—I don’t know of any existing product that even makes an attempt at it.

[gloriatyler]

Hi everyone,

Thanks for the tips. Using CV delays or random LFOs can definitely add an organic feel. You might also find text humanizer useful to explore tools that make text sound more natural, especially if you’re working on descriptions or notes for your music. Keep experimenting to see what works best.

[Toki]

Would Ochen K’s probability drum trig work?

https://www.reasonstudios.com/shop/rack ... m-trigger/

Or since you specifically asked about CV

https://www.reasonstudios.com/shop/rack ... v-trigger/

Edit: not sure if this just controls the probability of a trigger or when that trigger occurs (e.g. fwd/back in time). But I’m sure it can shuffle or groove to give realistic hits.

[selig]

I would think that adding probabilities is a must if you’re looking to “humanize”, because while no human is “perfect”, it’s also true that no human is “random”.

Random timing just sounds “random” to me. Drummers (like myself) are far from perfect, speeding up at certain points and pushing one part of the kit while pulling another (all the while “thinking” we’re playing tight).

So a true “humanized” likely requires machine learning to include characteristics and probabilities of real drummers, if your intent is to get your programmed drums to sound more “human”.

So, apologies for not have a product to suggest, maybe someone else knows of such a device? As for me, I simply play the parts in live - instant “humanize”!


Sent from some crappy device using Tapatalk

out of curiosity, did you look at what I built, or at the explanation?

the reason I ask is, I think you’ve made a similar comment in response to this combi before, and I vaguely remember responding in much the same way then. your response seems (and I could be wrong about this; it’s just how it seems to me) like you’re presupposing what it does without looking at it to understand how it’s built/how it works.

anyhoo, I took all of those things into consideration (to the extent I could) in the build. you can push and pull the kick ahead of or behind the other kit pieces. the random timing isn’t just random. it used a slow sine as a sort of “carrier” that provides the biggest variation in timing, but it’s further modulated by a smooth random LFO. the result would look like a distorted sin wave, if you were to look at it with a CV analysis tool.

the end result is the timing of each kit piece slowly moves ahead of and behind the beat, and back, with a bit of micro-randomness on top. similar to—as you suggest—the performance of a real drummer, gradually pushing and pulling the beat, alternatively.

I think you’re right that there’d need to be some sort of advanced AI algorithms to really do it well, but barring that, with the tools available, this is about as close as I can get. and why bother at all? for the same reason you suggest in your second last paragraph—I don’t know of any existing product that even makes an attempt at it.

Answering this 6 years later…
My opinion is based on over 40 years of having “randomize” available for MIDI sequencing and never using it, combined with 30 years of editing performances of studio musicians (timing, mostly), added to 50 years of playing drums/keys!

What I discovered early was “random” is not all that musical. Even AI isn’t (yet) all that musical, and it’s doing far more than a random algorithm can do. Next I found patterns to what professional musicians do with time, either accidentally or on purpose. Many musicians rush the tempo when going from section to section in a song, then settle down for a while when little changes - no “random” algorithm accounts for this. Musicians also drift ‘off time’ rather gradually, then correct their timing rather quickly in comparison, something no random algorithm does. Musicians also tend to have occasional ‘hiccups’ rather than general and constant ‘randomness’ as an algorithm does.

Best you can currently do (although I expect this to change in the near future) is slap a “groove template” of a few bars over your track, which doesn’t account for any of the above natural alterations but DOES come closer to human feel.

Bottom line, humans are not random, context is king, and natural timing alterations are WAY more complex than you may think (and each of us does “random” differently). In the future, I expect more contextual tools that are more aware of the overall song arrangement, which could mean a real “humanize” feature that follows the song structure as well as allowing personalization of the timing quirks that could only come from a real human (pure speculation, I have no inside information here).

[motuscott]

Selig appreciation response.
Dude seems to know hella lot about the musics.

[guitfnky]

out of curiosity, did you look at what I built, or at the explanation?

the reason I ask is, I think you’ve made a similar comment in response to this combi before, and I vaguely remember responding in much the same way then. your response seems (and I could be wrong about this; it’s just how it seems to me) like you’re presupposing what it does without looking at it to understand how it’s built/how it works.

anyhoo, I took all of those things into consideration (to the extent I could) in the build. you can push and pull the kick ahead of or behind the other kit pieces. the random timing isn’t just random. it used a slow sine as a sort of “carrier” that provides the biggest variation in timing, but it’s further modulated by a smooth random LFO. the result would look like a distorted sin wave, if you were to look at it with a CV analysis tool.

the end result is the timing of each kit piece slowly moves ahead of and behind the beat, and back, with a bit of micro-randomness on top. similar to—as you suggest—the performance of a real drummer, gradually pushing and pulling the beat, alternatively.

I think you’re right that there’d need to be some sort of advanced AI algorithms to really do it well, but barring that, with the tools available, this is about as close as I can get. and why bother at all? for the same reason you suggest in your second last paragraph—I don’t know of any existing product that even makes an attempt at it.

Answering this 6 years later…
My opinion is based on over 40 years of having “randomize” available for MIDI sequencing and never using it, combined with 30 years of editing performances of studio musicians (timing, mostly), added to 50 years of playing drums/keys!

What I discovered early was “random” is not all that musical. Even AI isn’t (yet) all that musical, and it’s doing far more than a random algorithm can do. Next I found patterns to what professional musicians do with time, either accidentally or on purpose. Many musicians rush the tempo when going from section to section in a song, then settle down for a while when little changes - no “random” algorithm accounts for this. Musicians also drift ‘off time’ rather gradually, then correct their timing rather quickly in comparison, something no random algorithm does. Musicians also tend to have occasional ‘hiccups’ rather than general and constant ‘randomness’ as an algorithm does.

Best you can currently do (although I expect this to change in the near future) is slap a “groove template” of a few bars over your track, which doesn’t account for any of the above natural alterations but DOES come closer to human feel.

Bottom line, humans are not random, context is king, and natural timing alterations are WAY more complex than you may think (and each of us does “random” differently). In the future, I expect more contextual tools that are more aware of the overall song arrangement, which could mean a real “humanize” feature that follows the song structure as well as allowing personalization of the timing quirks that could only come from a real human (pure speculation, I have no inside information here).

so, no. six years later the answer is no. you didn't look at what I built, or the explanation.

even more bizarre, it doesn't sound like you even read the post you were responding to. weird thing to "respond" to, years later, if you're mostly just responding to the thoughts in your head.

[selig]

Answering this 6 years later…
My opinion is based on over 40 years of having “randomize” available for MIDI sequencing and never using it, combined with 30 years of editing performances of studio musicians (timing, mostly), added to 50 years of playing drums/keys!

What I discovered early was “random” is not all that musical. Even AI isn’t (yet) all that musical, and it’s doing far more than a random algorithm can do. Next I found patterns to what professional musicians do with time, either accidentally or on purpose. Many musicians rush the tempo when going from section to section in a song, then settle down for a while when little changes - no “random” algorithm accounts for this. Musicians also drift ‘off time’ rather gradually, then correct their timing rather quickly in comparison, something no random algorithm does. Musicians also tend to have occasional ‘hiccups’ rather than general and constant ‘randomness’ as an algorithm does.

Best you can currently do (although I expect this to change in the near future) is slap a “groove template” of a few bars over your track, which doesn’t account for any of the above natural alterations but DOES come closer to human feel.

Bottom line, humans are not random, context is king, and natural timing alterations are WAY more complex than you may think (and each of us does “random” differently). In the future, I expect more contextual tools that are more aware of the overall song arrangement, which could mean a real “humanize” feature that follows the song structure as well as allowing personalization of the timing quirks that could only come from a real human (pure speculation, I have no inside information here).

so, no. six years later the answer is no. you didn't look at what I built, or the explanation.

even more bizarre, it doesn't sound like you even read the post you were responding to. weird thing to "respond" to, years later, if you're mostly just responding to the thoughts in your head.

Not sure your point, as we already agreed AI would do this and other suggestions would not, did we not?
Do we not still agree?
And if you’re really wanting my feedback on your work there are probably better ways to ask…

[guitfnky]

so, no. six years later the answer is no. you didn't look at what I built, or the explanation.

even more bizarre, it doesn't sound like you even read the post you were responding to. weird thing to "respond" to, years later, if you're mostly just responding to the thoughts in your head.

Not sure your point, as we already agreed AI would do this and other suggestions would not, did we not?
Do we not still agree?
And if you’re really wanting my feedback on your work there are probably better ways to ask…

my point is just that it still doesn’t seem like you’ve read any of what I said, yet felt the need to respond literally years later. my point is that I don’t see the point of doing that. sorry if that gets my hackles up, but I find it kind of insulting when people respond to stuff that the person they’re responding to didn’t say. call it a personal failing of mine.

I don’t even use the combinator anymore, because there ARE good algorithmic tools that do this better than a fun, complex little project I did more than half a decade ago. so, no, I’m not looking for feedback. six years ago, I was.

[selig]

Not sure your point, as we already agreed AI would do this and other suggestions would not, did we not?
Do we not still agree?
And if you’re really wanting my feedback on your work there are probably better ways to ask…

my point is just that it still doesn’t seem like you’ve read any of what I said, yet felt the need to respond literally years later. my point is that I don’t see the point of doing that. sorry if that gets my hackles up, but I find it kind of insulting when people respond to stuff that the person they’re responding to didn’t say. call it a personal failing of mine.

I don’t even use the combinator anymore, because there ARE good algorithmic tools that do this better than a fun, complex little project I did more than half a decade ago. so, no, I’m not looking for feedback. six years ago, I was.

That's not accurate or fair, I didn't "feel the need to respond". I felt I owed you a response. Sorry if I have bothered or offended you in doing so.

[guitfnky]

my point is just that it still doesn’t seem like you’ve read any of what I said, yet felt the need to respond literally years later. my point is that I don’t see the point of doing that. sorry if that gets my hackles up, but I find it kind of insulting when people respond to stuff that the person they’re responding to didn’t say. call it a personal failing of mine.

I don’t even use the combinator anymore, because there ARE good algorithmic tools that do this better than a fun, complex little project I did more than half a decade ago. so, no, I’m not looking for feedback. six years ago, I was.

That's not accurate or fair, I didn't "feel the need to respond". I felt I owed you a response. Sorry if I have bothered or offended you in doing so.

our exchange, in a nutshell:

sr: hey check out this thing I made that maybe sorta helps!
s: humans aren’t random—randomness doesn’t sound human
sr: I took that into account—did you read what I said? (I’d said it’s largely not random)

6 years go by…

s: in my experience, humans aren’t random—randomness doesn’t sound human
sr: you still haven’t read any of what I’ve said

if that’s inaccurate, or misrepresentative, I’ll eat my wallet. regardless, I know neither of us want to litigate this any further, so I’ll leave it at that—no hard feelings.

[selig]

That's not accurate or fair, I didn't "feel the need to respond". I felt I owed you a response. Sorry if I have bothered or offended you in doing so.

our exchange, in a nutshell:

sr: hey check out this thing I made that maybe sorta helps!
s: humans aren’t random—randomness doesn’t sound human
sr: I took that into account—did you read what I said? (I’d said it’s largely not random)

6 years go by…

s: in my experience, humans aren’t random—randomness doesn’t sound human
sr: you still haven’t read any of what I’ve said

if that’s inaccurate, or misrepresentative, I’ll eat my wallet. regardless, I know neither of us want to litigate this any further, so I’ll leave it at that—no hard feelings.

Everything is correct but the part about me not reading what you wrote. If you simply disagree with me, that's fine, and I'm guessing that's why you didn't think I actually read what you wrote?

Either way, I'm not sure how a back and forth about human rhythm turned into accusations about responding without first reading a post...
Carry on, I hope we can still discuss these things around here in the future!