PaulXStretch - has an update....

[essejlc]

There's a bug with the latest installer (Windows 10) - the 1.5.0 version installed the .vst3 file in the folder requested; the current one uses the .vst3 name as a folder in the location requested, and can't create a folder if you're updating, since the existing .vst3 file exists. Easy to solve (delete the old file), but no obvious reason why a .vst3 folder is needed anyway...

Actually the bug was in the old version, the VST3 format is supposed to be in a folder! That’s why I said to remove the abortive and incorrect vst3 file before installing 1.5.1.

You expect me to read all the instructions?

Personally, I really hate VST3s that create folders for, er, one file that has a .vst3 extension. It seems messy.

Apparently that's the "modern" way that steinberg wants people to use going forward (and thus the way that the JUCE build system make them), but it is a little silly when it isn't being made use of, I mostly agree. The idea is that multiple architectures could go in the same one...

[acousmod]

Nice, but Xenakios version was supporting 32 channels in the standalone and 16 in the plugin: would it be possible that you remove the 8 channels limit?

[greententacle]

chrash studio one 5 here, win 10

[essejlc]

Nice, but Xenakios version was supporting 32 channels in the standalone and 16 in the plugin: would it be possible that you remove the 8 channels limit?

I will remove the limit.

[essejlc]

chrash studio one 5 here, win 10

And how about the standalone on the same system? What is the exact CPU identifier on that system?

[acousmod]

Nice, but Xenakios version was supporting 32 channels in the standalone and 16 in the plugin: would it be possible that you remove the 8 channels limit?

I will remove the limit.

Thank you very much!

[pbelchamber]

Thanks essejlc for keeping this up to date – something this unique should be kept alive.

Unfortunately I am having issues - the standalone will not open on my machine (Win 10, i7 4820, 16GB Ram). The plugin also instantly generates a serious error message in both Cubase 12 and Ableton 11.1, and will not load in Bitwig 4.16, which gives an error saying “Plugin host died: Could not read async reply: End of stream: Broken pipe”.

I would be interested to know if anybody else has it working in those hosts - it could just be my machine, but usually it is very stable.

I should be able to send you crash dump files if that would be helpful – if so let me know where to send them.

[essejlc]

(I fixed the windows crashing issues!)

Grab the latest 1.5.2 for all platforms at https://sonosaurus.com/paulxstretch

- Increased maximum number of channels to 32
- Optimized FFTW build to use available hardware
- Fixed crash issue some people were seeing on older windows systems
- Added option (now default) to end recording automatically when chosen buffer size is reached
- More extremely small window size tweaks

[pbelchamber]

Fantastic - the 1.5.2 release seems to have fixed all the issues I was having. Thanks!

[kperry]

This may be a dumb question (and I can't get the tutorial on https://sonosaurus.com/paulxstretch/ to actually appear), while it works fine in Cakewalk By Bandlab, I can't actually see how to get the processed audio *back* onto a track (that's when using it as a plugin - I can obviously run the standalone and do the processing offline).

[Xenakios]

I can't actually see how to get the processed audio *back* onto a track

The plugin itself can't replace the audio on the host's track. The offline rendering feature is only part of the standalone application version. (But *maybe* that could be somewhat useful to have in the plugin version too?)

With the plugin, you will need to use whatever facilities the host has to offline render (or record) audio from the track/project. I am not familiar with how Cakewalk works but I would assume it has some kind of "render/export offline" or "glue/apply plugin" feature to make new audio files from the audio processed by the plugin.

[CinningBao]

Hi essejlc, I'm responsible for the initial push to get paulstretch working in 64-bit OSes, with of course deep thanks to xenakios for doing the actual work!

He listened to our outlandish requests and implemented ultra-long buffer sizes, and a few other things not initially present in the first edition. The end product gave us great numerical feedback and 'that' sound.

With that in mind, can I ask that you consider implementing a similar feature set? Looking at it alongside the Xenakios' version, can I ask that you

- put in the octave and 5th additional pitch-layer levels (in the Pitch section) back in. This is a really great way to thicken up the stretched sound, and add an extra layer of harmonics, and we all love extra harmonics
- larger FFT window - the previous version had a maximum of around a minute (at 44.1kHz) and this feature shouldn't be discarded
- better visual numerical feedback regarding
- FFT size (should be in milliseconds not samples, and scaled to seconds when value exceeds 1000)

Screenshot 2022-04-25 at 12.08.42.jpg

- the previous version gave us the option alter the 'range of available sizes from a menu item

Screenshot 2022-04-25 at 12.19.28.jpg

- estimated length (the output values in the GUI seem to be capped at 24 hours, so if the stretched length is 1 day and 4 minutes, only the 4 minutes is displayed - perhaps add a 'day' parameter to that value),
- add back the Space bar as a 'play/stop' controller

And a question: what does the red bar in the waveform area represent? It doesn't seem to change with any parameter adjustment so I'm struggling to determine what it is showing us.

Other than those issues, it does look great, so no qualms there!

[kperry]

I can't actually see how to get the processed audio *back* onto a track

The plugin itself can't replace the audio on the host's track. The offline rendering feature is only part of the standalone application version. (But *maybe* that could be somewhat useful to have in the plugin version too?)

With the plugin, you will need to use whatever facilities the host has to offline render (or record) audio from the track/project. I am not familiar with how Cakewalk works but I would assume it has some kind of "render/export offline" or "glue/apply plugin" feature to make new audio files from the audio processed by the plugin.

Yes, but it produces silence when you bounce.

[Xenakios]

Yes, but it produces silence when you bounce.

Is there any audio previously on the track or did you just import a file into the plugin? In the case there is no audio on the track the host might think there's no audio to bounce and optimize it out. You might need to do something like add a dummy audio file into the track that forces the host to render. That certainly wouldn't be an ideal workflow, but I don't think there's much the plugin can do about that. (Perhaps there's some kind of "plugin is audio generator" flag available, but I haven't myself looked into that...)

[kperry]

I've tried both with and without an audio clip, and silence reigns in both cases. There is such a flag (which I set) and I don't think it made any difference - I'm going to go through more methodically later.

There is one setting which does seem to be necessary for playback of the effected audio beyond the boundaries of a clip in the track ("Stop At Project End" - disable) but that is a pain to leave like that as the transport never stops rolling.