CDP Processes you would like to see added next

[j-p-higgins]

The plan is to add as many CDP processes to SoundThread as possible, however right now as I develop the UI I am keeping the number of processes to a small curated list. If you have particular processes you would like me to add next, please comment this here rather than opening an issue!

[coryncomposer]

Hi Jonathan

Great work. Requests - a lot of them I know - but most of all formants and vowels...
SHUFFLE
SPREAD
SUPPRESS
EXAG
ANALFILE
FOCUS ANALFILE
FOLD ANALFILE
FREEZE
STEP ANALFILE
SUPERACCU
BAND
BLTR
PLUCK TRACE
VOWELS

DIFF
MAKE
SPECTWIN
GET ANALFILE
GETSEE ANALFILE
PUT ANALFILE
SEE FORMANTFILE
VOCODE ANALFILE
ALTHARMS
TRANSP
SPECTRUM
GLISS
INVERT SHIFT
SQUEEZE

[dimitrissound]

Jonathan, this is amazing work, man!

One important feature for me is the ability to create automation using code, similar to the option available in Sound Loom's automation settings. This is quite important for sonification processes.

Another useful feature in Sound Loom, if I recall correctly, is stack processing — the option to process multiple files simultaneously. That’s something I’d love to see implemented in the future.

[j-p-higgins]

How would you want this to work? One of the reasons automation is a bit vague in soundthread is because it is not sound file specific as each node doesn't know how long the sound file it is going to receive is so all timings are in percent. Would precise parameter values be enough or would you need precise timings?

[dimitrissound]

Hi Jonathan, sorry for the late reply. If it’s not possible for Threads to read the file duration, then maybe there could be a message like: “Start with X values, and over X milliseconds gradually change to these values,” etc. Does this make sense?

[kijhome]

When I press the "Run Thread" button, the "Processes failed with exit code:1" message pops up and no output is generated, what should I do?

[j-p-higgins]

Please see issue #72

[kijhome]

I did the "cdprogs" folder path correctly. I still get that kind of message..

[j-p-higgins]

Ok I think I know what might be happening here. I think the space " " in your cdprogs path between windows and (2) might be upsetting the command line, can you rename that folder to remove the space and then rechoose your cdprogs folder location in the settings? I think the error it is trying to give you is this:

and windows is treating everything after the space as a new command

Also can you let me know what language you have your system set to? Does it use non-latin characters? I want to update the font in the console so that you get actual useable output in future versions.

[kijhome]

I solved it with your answer. Thank you :) My system language is Korean :)

[aftabhussainprojects]

Amazing work Jonathan, thanks for making and sharing this!

As for feature requests I'd love to see the spectral morphing process implemented: https://www.composersdesktop.com/docs/html/cmorph.htm

Cheers!

[j-p-higgins]

These are pretty high on my wish list although they may not be in the next major update (0.3.0) as the processes require multiple input files and that requires a bit of a rewrite to how SoundThread handles building commands for CDP and a fair bit of behind the scenes processing to handle mono/stereo mixes and potentially different sample rates. Having said that though, I have made some reasonable progress on that already, the development version now allows more than one input file in a thread, so maybe I can get them in for the next update, but if not at least some of these will be the one after (0.4.0).
Current development test build:

[secretkiillerofnames]

Morphing and Texture stuff is definitely on my wishlist when and if you get to it.
Some of the processes that allow you create an analysis of a file and then apply that analysis to another (be it a filter, envelope, spectral profile etc..) seem relatively unique and powerful.
And Modify Convolve!

[j-p-higgins]

Thanks these are all high on my personal wishlist! They won't be in the next release (0.3.0) but will be the focus of the 0.4.0 release see milestones for a rough timeline for that

[electronic-tiger]

I'm very happy that you've started this project.

Not sure if these suggestions are in the right spot, or if they're already coming. Or if I missed something. But I'd love to have...

A keyboard shortcut to add new process modules.

For example in Blender, a free and open source 3d editor, I can use "shift + A" and start typing to add the node that I want. I'd love to have a similar functionality in SoundThread. Personally I find "shift + A" a very convenient shortcut as it leaves my hands mostly in place to start typing right away. Minimizing hand movement. Especially between switching between keyboard and mouse. As is currently the case in SoundThread.

And another feature that could perhaps be lifted from Blender, dragging a module over a wire could insert it between the connection of two existing modules. Equally so, in Blender you can right click the mouse and drag over a wire to disconnect it. That'd be a welcome feature as well. Then maybe also the ability to delete a module in between two modules. With it automatically restoring the connection between the outer two modules.

I realize I may be biased towards the way Blender operates. But I do find these specific controls to be quick and intuitive.

[j-p-higgins]

Thanks for the great feedback and ideas!

I have opened an enhancements issue #101 to implement some keyboard controls in the search menu to make patching more efficient. I probably won't use shift+a for the keyboard shortcut to open the search as all other keyboard shortcuts in SoundThread are ctrl+something and I would like to keep the functionality consistent across all shortcuts. If you have any more thoughts on how the search could be better optimised for keyboard use, please do add some comments on that Issue!

I am planning to try and revisit how cables work by 1.0.0 there is an issue open for it here: #38. The reason this is being left so late in the development is that the patch cords are a default function in Godot (the software I am building SoundThread in) and by default there is no way to easily implement any changes to their behaviour. It is possible to override and develop my own patch cord system however, it is going to be a lot of work so I have decided to focus on getting the core functionality of audio processing in first before trying to rework the patch cords.

In future if you have any other UI improvement suggestions you can just open an Feature Request directly here. Or, if you would like to chat about potential features and ideas before opening a Feature Request there is a suggestions channel on the discord.

[dtenenba]

Congrats on the beta! I am loving SoundThread. It got me into looking at CDP (which I hadn't really done before) and I discovered CERACU which is amazing. Any chance of adding it? I realize it takes a file containing a list of integers which might require a new widget....

[j-p-higgins]

Processes that require special data files will be the focus of a future update (the one after next). I can't guarentee which processes I will be adding in that update yet but things like that will be coming

[coolsen10000]

Really enjoying this so far, amazing work!

Would it be possible to get a simple stereo-to-mono converter?

[j-p-higgins]

as in left and right are mixed down to mono or the ability extract left/right as a file? Theoretically yes, the processes exist in cdp although the framework to allow them to not throw errors if they are sent a mono file is not implmented in soundthread they are some of the only processes in cdp that require stereo input files so I havent implemented that. What would you want to use that for?

[coolsen10000]

Both mixing down the left & right channels and isolating the left & right channels would be fantastic. Mainly to have better control over the overall stereo image. I have a lot of stereo samples/audio files, and it would greatly help the workflow to be able to convert them into mono on the spot without having to do it in another software.

I also love the idea of running multiple stereo files from multiple input files and having some in mono and others in stereo, or even taking the left and right channels from different instances of the inputs and mixing them up.

[ex0du55utu7e]

I love this great, intuitive access to the power of CDP and have been using it to get more familiar with the tools I could never get a good handle on with earlier CDP GUIs. I feel the one thing that would completely unlock the full power of this approach for me was if you had two process node capabilities:

• a way to generically run a CDP command that doesn't yet have a GUI, like a commandline node. Inputs and outputs dragged to it should be able to generate references that can be mentioned on the commandline, and the commandline should be able to pass parameters. So I could open one of these and connect two frequency domain signal inputs and an output and enter on the text box entry something like "combine cross %i1 %i2 %o1 -i 0.5" and it would do the expected operation. This opens a lot of CDP without requiring all the processes to be implemented.
• a way to edit your own process node GUI. Each of these is largely just connecting up the parameters in useful ways to UI elements and hooking up the right CDP process by connections. There is also type validation and other important stuff under the covers, but it feels like this would free up a lot of your work if others could do this. Editable GUI nodes could even be something that could be "exported" in a format that people could submit back to the project as suggested implementations or shared in other spaces even if they didn't become official.

In other words, I think enabling people access to all CDP processes can be done faster than enabling each one-by-one, and it could be possible to distribute the workload and opensource the opensource a bit more. An approach like that forces you to consider your abstractions and ensure they are robust in ways that only get discovered along the way with a more centralized approach.

[j-p-higgins]

The bulk of my development time is not really adding the processes themselves, I am not making a UI for each one individually they are generated dynamically from a json. It is just having the validation in place for SoundThread to be able to call the CDP commands in the right order, ensuring that the input and output files and the variables all meet the specs of that CDP process and having a way to generate the correct break files that some processes need.

I could theoretically add a command line node but right now if the process isn't already in SoundThread, SoundThread probably will fail to send the right things to the process and you'd be better off just using another UI or the command line directly.

In future I have plans to allow for add-ons similar to externals in max or pd see: #86 this will allow anyone to add whatever they like really to SoundThread. It's worth noting that right now there's nothing really stopping anyone from doing this right now, as long as you put the programs in the same folder as the CDP programs you could add them using the JSON editor scene that is in the source code. But like I say there is a decent chance they won't run correctly right now.

[johnbury1]

Hi

Very much a newby here. tried to run my first thread, it failed with an Exit code: 9. Comments appear to suggest the input file was neither mono or stereo (the sample used is stereo). Any pointers or can you let me know if you need more information

cheers, John

[j-p-higgins]

Hi please could you make a bug report in the issues tab: https://github.com/j-p-higgins/SoundThread/issues using the format provided. In particular I will need a full copy and paste of the console output (you can right click in the console to select all and copy) to be able to diagnose the issue. The input file not being mono or stereo error message can come up for lots of different reasons if things aren't set properly and exit codes don't actually tell me much

[zardizardi]

Hi and first of all
We all appreciate this idea that you made it real
I have also an idea about real time editing audio phrases
Like when the out put is looping we can change the elements in real time
And the end of our wishes we can save the final result without to many audio file
Which that could be an option to turn on or off so all those real time imaginations will be saved

[j-p-higgins]

Hi sorry, the majority of the processes in cdp simply cannot run in realtime, they need to see the full file in order to run. For what you are looking for I would reccomend just using Pure Data or Max Msp

[timbgray]

A way to manage the saved generation file with the output would be cool. I have some worthwhile sounds generated but it’s tough to associate them with the saved source generation file.

[lukenzoner1]

Hey Jonathan,
Many many thanks for the last update.
It really gives some more depth to soundthreads and this pushes CDP to the right direction.
I would like to suggest something like a bunch of function that could be accessed via customizable keyboards shortcuts.
Maybe it's already init and i Missed it out 📦 ...
Cheers
Lucien

[j-p-higgins]

for suggestions you can make a feature request in the issues tab, theres a template for that!

[lukenzoner1]

okay done.
I'm sorry, first time suggesting things here :)

[zellerron]

Mane!!!! PITCH QUANTIZE!!!!!!!!!!

Morph
NewMorph
NewMorph2
MADRID
Motor
Motifs
Newtex
Hilite Vowels
Doublets
Ceracu
CHORD
chordF

[zellerron]

Oh yeah .... and HOVER & HOVER2