richhop

I've been making this jazz patch obsessively and always want to make it better, even though it is pretty good already, i'm not sure when to stop, as I have no deadlines and a seemingly endless, unquenchable curiosity for augmenting theoretical and physical ability in a musical context.

The idea is kind of similar to a chord memoriser, except you have two dependancies, the root note that triggers the chord (as usual) and a harmonising melody note that selects the chords add on/extended harmony notes. The chord function also triggers an appropriate scale to jam over the chord, which is played in using a 16 pad drum controller.

I made the patch as I love extended harmonies in chords and jazz music but can't play the piano or solo on the guitar over the complex key changes or think about it fast enough to play it, so I've reduced the theory behind jazz into an algorithm that allows control with no knowledge of music theory or instrument playing ability.

I write this long ass post as I am at a cross roads of the direction I should take next with it, should I delve into the dark depths of MSP and make some sort of live input machine listening reharmoniser? Or shall I make the chord reharmonisation be controlled by the current melody note from the drum pad? I'm not sure what will end up sounding better. Or does anyone have a perspective from someone who can actually play and knows their theory/instrument like this and where I could possibly go next to include?

Or shall I just call it a day? Instead of obsessively trying to make every little part behind the scenes as tight as I can! even though it works as intended; I'm not really sure when to stop!

Here's an example of what it can do: http://youtu.be/NdNX5Qz4g5I

The idea is kind of similar to a chord memoriser, except you have two dependancies, the root note that triggers the chord (as usual) and a harmonising melody note that selects the chords add on/extended harmony notes. The chord  function also triggers an appropriate scale to jam over the chord, which is played in using a 16 pad drum controller. 

I write this long ass post as I am at a cross roads of the direction I should take next with it, should I delve into the dark depths of MSP and make some sort of live input machine listening reharmoniser? Or shall I make the chord reharmonisation be controlled by the current melody note from the drum pad? I'm not sure what will end up sounding better. Or does anyone have a perspective from someone who can actually play and knows their theory/instrument like this and where I could possibly go next to include? 

Or shall I just call it a day? Instead of obsessively trying to make every little part behind the scenes as tight as I can! even though it works as intended; I'm not really sure when to stop!


does-anyone-else-not-know-when-to-stop

sounds like you're at the point where i'd usually switch from 'developing software' to 'practicing music' with it(and you'll still end up correcting/editing things along the way).

now observe as it lives and breathes for a bit, it'll tell you what's next when it's ready.

sounds like you're at the point where i'd usually switch from 'developing software' to 'practicing music' with it(and you'll still end up correcting/editing things along the way).
now observe as it lives and breathes for a bit, it'll tell you what's next when it's ready.

now observe as it lives and breathes for a bit, it’ll tell you what’s next when it’s ready.

I've been doing exactly that for the last month or more, constantly tweaking and making music I never had the ability to make before, it's an actually crazy experience being able to skip hours of practice and understanding, go straight to the heart of making and all live. 

What I heard when I listened to the patch was it telling me to track a vocal line's pitch and use that as the harmonising note, but I'm sure it's gonna be hard to do with pitch tracking live input vox. But then again auto tune tracks pitch in real time, so maybe it's not as unfeasible as I thought.

I've been doing exactly that for the last month or more, constantly tweaking and making music I never had the ability to make before, it's an actually crazy experience being able to skip hours of practice and understanding, go straight to the heart of making and all live. https://soundcloud.com/irichop

What I heard when I listened to the patch was it telling me to track a vocal line's pitch and use that as the harmonising note, but I'm sure it's gonna be hard to do with pitch tracking live input vox. But then again auto tune tracks pitch in real time, so maybe it's not as unfeasible as I thought.


What I heard when I listened to the patch was it telling me to track a vocal line’s pitch and use that as the harmonising note, but I’m sure it’s gonna be hard to do with pitch tracking live input vox.

ah, nice! (a vocal can still be easier to track than an instrument playing chords, but it's true, tracking noisy 'sibilants' do make it difficult...) other than searching here for 'pitch track' i'd recommend seeing the 'fzero~' helpfile if you need starting ideas.

Takes deep breath, this is might be a world I'm not quite ready for haha! I thought it wouldn't be too fast actually, as I read somewhere you need a big enough window size, or detect a certain amount of continuous cycles to give an accurate reading, but then is it not gonna be bang on the exact frequency of this arbitrary log scale in equal tempered tuning? Would you determine thresholds or boundaries for out of tune singers? or accounting for the slight fluctuation in pitch like vibrato seems like crazy stuff to me!

It's either MSP or go insane trying to inject awareness into the chords themselves and make a tangled hierarchy feedback loop! Harnessing the random and weighted probabilities seem equally fruitful ports of call next also, now I have the voices and all the scales in every key mapped out, making a decision making engine that's self aware seems less scary than the DSP stuff haha! I don't know when to stop; I do wonder whether it would be easier to go into code for the neural net though and make an actual max object (I'll see you in ten years) I think i'm completely obsessed. It's amazing max is the infinite instrument.

Thing is, having an infinite instrument with no virtuosity built up doesn't help.

Take the time to play it. Take the time to find your style and your approach to wrangling the beast you've made.

Worst case? You get some great idea fodder for when you start patching again. Best case? You realize the patch is all you've ever wanted, more or less.

/Andreas realizes he's recommending NOT patching anymore, and goes to take a nap.

Thing is, having an infinite instrument with no virtuosity built up doesn't help. 

Take the time to play it. Take the time to find your style and your approach to wrangling the beast you've made.
Worst case? You get some great idea fodder for when you start patching again. Best case? You realize the patch is all you've ever wanted, more or less.

/Andreas realizes he's recommending NOT patching anymore, and goes to take a nap.


just chipping in my own 2c. On the wider creative topic, if my patch functions as I had originally intended, in parallel with unexpected emergent features, then I consider it complete, and I try to move onto something else. Option anxiety is a creative force for both good and ill, unless you know how to channel it towards your original goal. You know the expression "it does exactly what it says on the tin"? It's become my design mantra. But having said that, MaxMSP is 

 a great tool for improvising, if you know your way around the 'instrument'. Yes, I've talked myself into a philosophical conundrum. (EDIT: and for those reasons, my short answer to your post topic would be, Yes, I also sometimes find it difficult to stop, when an idea presents itself in mid-patching).

Looking around at the work of others is also a helpful process (reading your post and watching the video made me think of Richard Garrett's nwdlbots for some reason).

Raja's recommendation for [fzero~] is a good one, and Andreas Wetterberg's recent thread on recursive algorithms might be related to your later musings.

Hi
just chipping in my own 2c. On the wider creative topic, if my patch functions as I had originally intended, in parallel with unexpected emergent features, then I consider it complete, and I try to move onto something else. Option anxiety is a creative force for both good and ill, unless you know how to channel it towards your original goal. You know the expression "it does exactly what it says on the tin"? It's become my design mantra. But having said that, MaxMSP is itself a great tool for improvising, if you know your way around the 'instrument'. Yes, I've talked myself into a philosophical conundrum. (EDIT: and for those reasons, my short answer to your post topic would be, Yes, I also sometimes find it difficult to stop, when an idea presents itself in mid-patching).

Quick, true story that I tell almost all of my students:

When I was a kid, I built this kick-ass go-cart. It was literally four wheels attached to a board with a piece of string to steer the front two wheels. We lived on a hill and we'd all race our go-carts. Mine won the first time out, and kept winning.

To celebrate the win, I started "accessorizing" the thing: I added lights and a brake. Still pretty cool. And I still won. I painted it orange...still won.

I kept adding to it, bit by bit, but when I shaped a wire mesh to the front and built it up with plaster to make it look more like a race car, I stopped winning. I figure it must have been too heavy.

Whatever the case, the moral is pretty obvious: it was perfect until I started messing with it. I added to it until it basically stopped being functional the way I had intended and needed it to be.

When writing a piece of music, or writing a max patch, I keep that in the back of my mind. When it's doing what you want it to do, leave it alone. Build your next go-cart, don't f*ck up the one you have.

When I was a kid, I built this kick-ass go-cart. It was literally four wheels attached to a board with a piece of string to steer the front two wheels. We lived on a hill and we'd all race our go-carts. Mine won the first time out, and kept winning. 

To celebrate the win, I started "accessorizing" the thing: I added lights and a brake. Still pretty cool. And I still won. I painted it orange...still won. 

I try to separate projects into two camps - the first are those where I know exactly what they need to do, I'll make them and then be finished. The other camp is made up of patches which never had an end goal, endless journeys which can change totally from the start to whenever I put them down. I've made peace with them never being finished or having any real goal, but those are usually the patches I get the most enjoyment from. The ones that keep me up at night with thoughts of 'maybe if I connected that to there...'

This thread is so great to read, thanks everyone for posting, I've almost got a general consensus from people just saying play the damn thing! Which is probably the kick I need to stop looking for things to fix, or break as the great parable from Lowenstern put so eloquently!

"Option anxiety is a creative force for both good and ill, unless you know how to channel it towards your original goal"

Completely agree with that sentiment, and the in relation to the Lowenstern point as well, I have actually made versions more complex than at present, that actually just confused me when I looked back and couldn't progress further, so I stripped it back a bit and works like a force now!

The problem I got to with it was wanting to do everything at the same time but not having enough hands for it, for instance in the video I played it in cross handed with one hand triggering a note for the chord and one hand triggering the drum pad, so changing key at the same time was quite unfeasible. I also had an arp'er and bassline you could play in with the same controller as the Chords, but just got too complex to keep track of, or jam fluently with.

Also the most random thing did happen to me last night, and think I solved all my problems, when I wasn't thinking and just doodling on graph paper, just drawing shapes and think I found the perfect graphical interface for this chord device. And then I googled tessellation and got way more inspired to make crazy interfaces than I should have, the magic of shapes!! I think max has connected me to some alien slip stream of nonsensical creative bliss!!

"I try to separate projects into two camps – the first are those where I know exactly what they need to do, I’ll make them and then be finished. The other camp is made up of patches which never had an end goal, endless journeys which can change totally from the start to whenever I put them down."

I've got the max/msp book by Giri now and think that's how I'm gonna approach learning with that book. I have no clue what I want to do with the DSP stuff, so will be a great opportunity to learn/experiment/get inspired, instead of trying to make something concrète (pardon the pun).

Oh and btw the nwdlbots thing is pretty cool, really like that markov chain tutorial knocking about as well, you get some great realistic random music coming out. I wonder if the nwdlbot could benefit from some probabilistic weightings fed from real musicians, as it almost seems too random to be useful in certain contexts. That was very much the consideration when building my patch, making something that is useful for me in composition situations. Really get off on exploring how to connect algorithmic processing with elements of control that result in synergistic effects and end up augmenting ability! Too much fun!!

This thread is so great to read, thanks everyone for posting, I've almost got a general consensus from people just saying play the damn thing! Which is probably the kick I need to stop looking for things to fix, or break as the great parable from Lowenstern put so eloquently! 

Completely agree with that sentiment, and the in relation to the Lowenstern point as well, I have actually made versions more complex than at present, that actually just confused me when I looked back and couldn't progress further, so I stripped it back a bit and works like a force now! 

"I try to separate projects into two camps – the first are those where I know exactly what they need to do, I’ll make them and then be finished. The other camp is made up of patches which never had an end goal, endless journeys which can change totally from the start to whenever I put them down." 

I've got the max/msp book by Giri now and think that's how I'm gonna approach learning with that book. I have no clue what I want to do with the DSP stuff, so will be a great opportunity to learn/experiment/get inspired, instead of trying to make something concrète (pardon the pun). 

Oh and btw the nwdlbots thing is pretty cool, really like that markov chain tutorial knocking about as well, you get some great realistic random music coming out. I wonder if the nwdlbot could benefit from some probabilistic weightings fed from real musicians, as it almost seems too random to be useful in certain contexts. That was very much the consideration when building my patch, making something that is useful for me in composition situations. Really get off on exploring how to connect algorithmic processing with elements of control that result in synergistic effects and end up augmenting ability! Too much fun!!


Two cliches come to mind, both of which apply

1 - regarding music - "songs are never finished, they are merely abandoned"

2 - regarding software - "perfect is the enemy of good enough"

Two cliches come to mind, both of which apply
1 - regarding music - "songs are never finished, they are merely abandoned"
2 - regarding software - "perfect is the enemy of good enough"


Does anyone else not know when to stop...?