Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add feature objects for slicers #35

Merged
merged 14 commits into from
Mar 29, 2022

Conversation

weefuzzy
Copy link
Member

@weefuzzy weefuzzy requested review from tremblap and jamesb93 March 28, 2022 17:29
@jamesb93 jamesb93 added this to the beta6 milestone Mar 29, 2022
@jamesb93
Copy link
Member

Help file documentation done for this now

@tremblap
Copy link
Member

I see @jamesb93 is on fire - should I wait for a ping to review?

@jamesb93
Copy link
Member

go for it now! I struggled to replicate even a fraction of what Max could show with waveform and multislider but the guts are there...

@tremblap
Copy link
Member

indeed waveform is not done yet (and has the added pleasure of not having all the needed bits in vanilla)

@tremblap
Copy link
Member

@weefuzzy I get a bug - the ampfeature outlet is considered an audio one by Pd . loading the helpfile provided by @jamesb93 yield an error and once I disconnect it, it refuses to reconnect:
Screenshot 2022-03-29 at 15 06 39

@tremblap
Copy link
Member

noveltyfeature and onsetfeature are working fine in list output. maybe it is a 'feature' of the object to output audio rate (which would make sense) and I'll fix the helpfile...

@weefuzzy
Copy link
Member Author

It's an audio output yes

@tremblap
Copy link
Member

ok so that means that this help never worked. I'm fixing them and @tedmoore is showing me how to push on the PR. stay tuned!

@tremblap
Copy link
Member

ok the RT version of the spectral ones generate a lot of 0s between otherwise what I guess are the right onset values. I reckon it is to do with hop size vs (like for all descriptor?) but maybe we could check that fluid.onsetfeatures~ gives the same live stream in Max. At the moment I'm sorting the other helpfile but if @weefuzzy @tedmoore or @jamesb93 can confirm on 'nicol' with the setting in the recently updated pd help generate a lot of zeros. multislider type patch is now in pd so that will help to see what I mean

@jamesb93
Copy link
Member

ok the RT version of the spectral ones generate a lot of 0s between otherwise what I guess are the right onset values. I reckon it is to do with hop size vs (like for all descriptor?) but maybe we could check that fluid.onsetfeatures~ gives the same live stream in Max. At the moment I'm sorting the other helpfile but if @weefuzzy @tedmoore or @jamesb93 can confirm on 'nicol' with the setting in the recently updated pd help generate a lot of zeros. multislider type patch is now in pd so that will help to see what I mean

onsetfeature generates a lot of zeroes but spikes in the right places, but noveltyfeature with a higher kernelsize should give you something smoother looking.

@tremblap
Copy link
Member

Same behaviour on noveltyfeature, now updated too. I still get a load of zeros between the values despite kernel being big.

@jamesb93
Copy link
Member

Same behaviour on noveltyfeature, now updated too. I still get a load of zeros between the values despite kernel being big.

This is on the oboe sound:

image

Which is somewhat smooth. Do you get something like that?

@tremblap
Copy link
Member

try on drums, and put your min to -1 so you can see the many 0s between values. Once I'm done checking the rest I can send the updated values, on drums, which make them even more apparently strange.

@jamesb93
Copy link
Member

image

blue is change filtered, red is the normal one

domain is -1 -1

@tremblap
Copy link
Member

after updating the bufnovelty version, I can confirm there is a discrepancy of output between the rt and the nrt in pd - if I use a small number of frames on the buf I don't get zeroes between components. @jamesb93 remove the change in your patches (although you would get the interleaved 0s I'm talking about) so @weefuzzy there seem to be a Pd problem here

@jamesb93
Copy link
Member

after updating the bufnovelty version, I can confirm there is a discrepancy of output between the rt and the nrt in pd - if I use a small number of frames on the buf I don't get zeroes between components. @jamesb93 remove the change in your patches (although you would get the interleaved 0s I'm talking about) so @weefuzzy there seem to be a Pd problem here

It's occluded by the lack of a third dimension but change is running on one copy (blue) and the unfiltered is red.

Copy link
Member

@tremblap tremblap left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok this works now as expected. thanks!

@jamesb93 jamesb93 merged commit 12c4bf9 into flucoma:dev Mar 29, 2022
jamesb93 pushed a commit that referenced this pull request Mar 30, 2022
* nightlies workflow

* use PD_PATH flag

* use the pd binaries on msp

* dont use powershell in unix

* fix borked ls

* point to the right sdk

* make release builds on windows

* remove pdb file filter

* use dev core

* delete release before uploading the new one

* abstraction added to convert string back to float

* giving credit (#11)

* rename tag

* provide the proper fluid path for mac builds

* provide the proper fluid path for the other 2 CCE's

* clone core not sc... 🤦‍♂️

* Add handling for MLP read methods to wrapper (#10)

* cleaned structure and copy of abstractions

* add docs to PD CI

* add pd path

* use a macos runner for docs

* correct the name for Windows .zip

* [CI] Refactor and use composite actions (#13)

* use flucoma actions instead of horrid code

* use more composite actions where possible

* remove extraneous -name:

* try a different folder structure out

* move one level up for packaging

* demoaudiofiles abstraction for the helpfiles

* added fluid.demosound~

* move to v4 of actions

* also update docs

* cosmetic changes to improve readability

* chroma help update

* remove weird magic number fixes #18

* correct order of variables in PDBufferAdaptor error message fixes #17

* corrections in progress

* chroma new format

* bufchroma done

* changing spaces for underlines for enhanced readability

* use new action for building docs

* use ubuntu to build docs

* update release workflow

* add support for core #88

* copy to datasets path

* ignore data

* copy resources data on cmake configure

* change source1 to sourcea and source2 to sourceb

* add workflow dispatch to nightly

* correct docs install source

* fix the missing ; at the end of the updated file names

* Add feature objects for slicers (#35)

* Add feature objects for slicers

* onsetfeature help file

* rt feature help files

* remove thing about pianos

* add buffer feature help files

* @Algorithm

* fixing the helpfiles

* expanded onsetfeature

* norveltyfeature fixed

* ampfeature fixed

* fixed bufampfeature

* bufnoveltyfeature corrected

* fix bufonsetfeature

* corrected range for onsetfeature viewer
jamesb93 pushed a commit that referenced this pull request Jun 14, 2022
* nightlies workflow

* use PD_PATH flag

* use the pd binaries on msp

* dont use powershell in unix

* fix borked ls

* point to the right sdk

* make release builds on windows

* remove pdb file filter

* use dev core

* delete release before uploading the new one

* abstraction added to convert string back to float

* giving credit (#11)

* rename tag

* provide the proper fluid path for mac builds

* provide the proper fluid path for the other 2 CCE's

* clone core not sc... 🤦‍♂️

* Add handling for MLP read methods to wrapper (#10)

* cleaned structure and copy of abstractions

* add docs to PD CI

* add pd path

* use a macos runner for docs

* correct the name for Windows .zip

* [CI] Refactor and use composite actions (#13)

* use flucoma actions instead of horrid code

* use more composite actions where possible

* remove extraneous -name:

* try a different folder structure out

* move one level up for packaging

* demoaudiofiles abstraction for the helpfiles

* added fluid.demosound~

* move to v4 of actions

* also update docs

* cosmetic changes to improve readability

* chroma help update

* remove weird magic number fixes #18

* correct order of variables in PDBufferAdaptor error message fixes #17

* corrections in progress

* chroma new format

* bufchroma done

* changing spaces for underlines for enhanced readability

* use new action for building docs

* use ubuntu to build docs

* update release workflow

* add support for core #88

* copy to datasets path

* ignore data

* copy resources data on cmake configure

* change source1 to sourcea and source2 to sourceb

* add workflow dispatch to nightly

* correct docs install source

* optional args: add to wrapper

* Wrapper: Add choices param (symbol list -> bitset)

* fix the missing ; at the end of the updated file names

* Add feature objects for slicers (#35)

* Add feature objects for slicers

* onsetfeature help file

* rt feature help files

* remove thing about pianos

* add buffer feature help files

* @Algorithm

* fixing the helpfiles

* expanded onsetfeature

* norveltyfeature fixed

* ampfeature fixed

* fixed bufampfeature

* bufnoveltyfeature corrected

* fix bufonsetfeature

* corrected range for onsetfeature viewer

* ignore more permutations of build folder

* PDWrapper: deal with some warnings

* FluidTensorView copy updates

* changing preset and interface for noveltyslice examples

* new overview including all data objects

* PDBufferAdaptor: Add redraw() and call on release()

* remove the data folder being copied in the dev packaging (#27)

* concataudiofile and it help merged in dev

* `ControlIn` objects
– properly adjust `ac` when adjusting `av` 
– ensure mListSize is always intialized to something

* Enhance/max params (#44)

* CMake: Centralise C++ version and set to 17

* Wrapper: Handle new LongRuntimeMax parameter type

* Wrapper: unused variable

* Wrapper: add `max<blah>` parameter to PD object for `LongRuntimeMaxT` 

fixes #45

* updated helpfiles for scalers to include and test inversetransform(point)

* Wrapper: fix argument count and warning printer

* Wrapper: update for `maxFFTSize`

* Wrapper: fix reset for shared objects

* Wrapper: use raw maxFFT in FFT setter

* Wrapper: fix logic for max params

* Wrapper: sort max FFT logic, fixes #48

* Enhance/generate stubs (#49)

* CMake: Generate .cpp stubs

* CMake remove old cpp stubs and scripts

* CMake add Utiltiy for making a working folder of symlinks for dev work

* Placate MSVC

* CMake: properly invoke docs

* CMake: install prefix is a cache variable

* CMake: Tidy up

* CMake: typo

* CI: update nightly workflow

* CI: remember to use correct branches

* CI: remove reference to docs job

* CMake: belatedly add branch selection for flucoma deps upon which CI relies

* CI: Correct name of Linux build artefact

* help upgrade: first batch Martin

* help upgrade: Martin's 2nd batch

* help review: martin 3rd batch

* help - typo in filename

* CMake: fix installation path for externals, fixes #50

Repeat after me: don't just paste from the Max one

* CMake: Pick up client tags for finer tilde control

* Enhance/managed objects (#47)

* Wrapper: Add managed buffers and message outptus

* Wrapper: Update outlets for list input

* Wrapper: Fix outlet reference for buffer completion

* Add header for bufinfo

* Add C++ stub for bufinfo

* Managed arrays: shrink collection when needed

* Update include/FluidPDWrapper.hpp

Remove leftover debugging `post()`

* BufferAdaptor: Ensure dirty flag is properly initialised

* Wrapper: Stop hosted buffers vector reallocating and breaking pointers

* updated helpfiles for new interface

* bufinfo help

* PD Wrapper: Quite possibly fix a double free problem for multiple inlets.

* PDWrapper: data objects respond from right outlet for read, write etc

* dataset objects: update with right outlet behaviour

Co-authored-by: tremblap <[email protected]>

* Wrapper: Cheer up MSVC

* PDWrapper: Try to swerve around GCC linking sadness

* CMake: Add local target for `fluid.bufinfo`

* FluidBufInfo: Missing include

* PDWrapper: Redo inlet proxies for less crashing

* kdtree - added 2nd tab

* kdtree 3rd tab added

* kdtree fix

* BufNNDSVD -> BufNMFSeed (#57)

* change filename

* change content

* typo in help report of bufferadaptor

* token skmeans entry and addition to overview (#59)

* [Enhance] Update Release Workflow (#60)

* update extension to be normalised

* cleanup

* update release

* remove workflow disptach variables

* fix release

* Update release.yml

* Update release.yml

* Update release.yml

* Update release.yml
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants