Christoph Gohlke
Christoph Gohlke
Thank you. I will review this when libjpeg-turbo 3.1 is released.
I have tested this with libjpeg-turbo 3.1 beta and it works as expected. The changes will be in the next release of imagecodecs along with some tweaks to make tests...
The plan is to do a release before Python 3.13, this or next weekend. Hopefully there are no major issues on macOS. I am attaching the current [_jpeg8.pyx](https://github.com/user-attachments/files/17077959/_jpeg8.zip), which should...
Fixed in imagecodecs 2024.9.22.
The released wheels are built against libjpeg-turbo 3.0.4. I'll wait for version 3.1. No, the system JPEG library is not used. The dynamic JPEG library used is in the `imagecodecs/libs`...
Thanks for reporting the issue. I can reproduce the warnings and traced them to the latest changes in the encoder of [_jpeg2k.pyx](https://github.com/cgohlke/imagecodecs/commit/080cc2c08e21e168e97c929e8457d3c83701bba5), not due to any OpenJPEG changes.
Ouch. Turns out the encoded buffer is not cropped correctly in some cases and trailed with uninitialized memory, which sometimes interferes with decoding. Not sure how the tests could pass...
Fixed in version 2024.9.22.
> data is in another zenodo > the data Alex gave to us Where can we find those data?
Do I understand correctly that the harmonics are expected in the last dimensions of the phasor coordinates? When implementing `phasor_from_signal` function, we settled that harmonics, if any, are in the...