personal_username

do you have general suggestions on how to use [dict] at its best?

I have a huge patch with over 1000 variables, and need to manage structural elements such as midi, music composition elements (melodies, delays, pitch, velocities etc...), audio setups, vsts, and other 10 or so bpatchers.

Being my patch a composition tool to handle 2+ hour music it can get quite stuffed with recallable presets, and as mentioned I'm migrating to [dict] for that.

Right now I need to decide if the best use of it is to:

1) have several [dict]s, (es: [dict midi], [dict variables], [dict composition_elements] etc..) OR

2) one big [dict ALL], where I grab all the info from and redirect to those midi, variables, composition_elements etc...

Right now I'm using 1, but I would like to explore the idea of a unique, all-controlling [dict ALL].

My only concern in proceeding with 2 is that I fear that it might get quite information-heavy (one preset will contain ALL subdicts, and that might be a lot of text), so I fear that recalling/dumping the data to everything at once might get me slowdowns.

Do you have any suggestion on which approach might fit best 

do you have general suggestions on how to use [dict] at its best?
I have a huge patch with over 1000 variables, and need to manage structural elements such as midi, music composition elements (melodies, delays, pitch, velocities etc...), audio setups, vsts, and other 10 or so bpatchers.
Being my patch a composition tool to handle 2+ hour music it can get quite stuffed with recallable presets, and as mentioned I'm migrating to [dict] for that.

1) have several [dict]s, (es: [dict midi], [dict variables], [dict composition_elements] etc..)  OR
2) one big [dict ALL], where I grab all the info from and redirect to those midi, variables, composition_elements etc...

What would be your strategy?
Right now I'm using 1, but I would like to explore the idea of a unique, all-controlling [dict ALL].
My only concern in proceeding with 2 is that I fear that it might get quite information-heavy (one preset will contain ALL subdicts, and that might be a lot of text), so I fear that recalling/dumping the data to everything at once might get me slowdowns.

Have you experienced anything similar? 
Do you have any suggestion on which approach might fit best for real-time performance?



Hi guys, 
do you have general suggestions on how to use [dict] at its best?
I have a huge patch with over 1000 variables, and need to manage

dict-suggestion-on-how-to-use-it-at-best-for-real-time-performance

I think I would lean toward separate dicts. You can always glob their contents together into one big one later, and quite easily, for “meta” presets or backup and files writing.

it’s probably also worth doing some kind of stress/performance tests to see how much dict info could be recalled at once in your patch. I don’t know how many objects you’re dealing with, but I have run into a situation in M4L with pattrstorage that necessitated breaking one big hierarchy down into components in order to avoid big CPU spikes / audio pops on recall. Its possible (though maybe less likely) that dict may be vulnerable to something similar?

I think I would lean toward separate dicts.  You can always glob their contents together into one big one later, and quite easily, for “meta” presets or backup and files writing.

it’s probably also worth doing some kind of stress/performance tests to see how much dict info could be recalled at once in your patch.  I don’t know how many objects you’re dealing with, but I have run into a situation in M4L with pattrstorage that necessitated breaking one big hierarchy down into components in order to avoid big CPU spikes / audio pops on recall.  Its possible (though maybe less likely) that dict may be vulnerable to something similar?

Thanks for your reply Tyler, I am actytually coming from pattrstorage and I hope to reduce CPU spike with dict. Will make some tests these days to follow up on this here

After some testing (and thought experiments! :-P)) having several dicts is the way to go.

Otherwise redundancies will overkill the data flow. Agree with you Tyler. 

[Mastermind Preset] to [small preset] is the most efficient way, especially when paired with "Model–view–controller" structure. Such as removing View and Control from the main patch and have them a la carte loadable B-patchers, and just leaving the Module within the main patch.

After some testing (and thought experiments! :-P)) having several dicts is the way to go.
Otherwise redundancies will overkill the data flow. Agree with you Tyler. 
[Mastermind Preset] to [small preset] is the most efficient way, especially when paired with "Model–view–controller" structure. Such as removing View and Control from the main patch and have them a la carte  loadable B-patchers, and just leaving the Module within the main patch.


dict: suggestion on how to use it at best for real-time performance