Login / Register  0 items | $0.00 New

PaulStretch3 preview(11) release

Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Mon Nov 13, 2017 4:08 pm Re: PaulStretch3 preview(10) release

Some progress with the plugin, it can now capture the audio coming from the host. (Automation of parameters works too but I forgot to show that in this video. edit : Sadly the automation will suffer from poor timing accuracy, especially with larger FFT sizes, and there isn't much that can be done about that.)
https://www.youtube.com/watch?v=tQ7fMX7 ... e=youtu.be
Last edited by Xenakios on Mon Nov 13, 2017 7:21 pm, edited 1 time in total.
lexr
KVRer
 
8 posts since 31 Jan, 2017

Postby lexr; Mon Nov 13, 2017 5:09 pm Re: PaulStretch3 preview(10) release

Awesome!
Kalamata Kid
KVRAF
 
2862 posts since 26 Jul, 2001, from Tarpon Springs, Florida, USA

Postby Kalamata Kid; Mon Nov 13, 2017 7:38 pm Re: PaulStretch3 preview(10) release

Wow!
Fantastic!
VST or VTSi?
touchable?
multi-touchable?

Thanks Xenakios for all the great work you are doing!
23" 10-touch Monitor and iPad Air as Midi Controllers, iConnectMIDI2+, Windows 10-64bit, Xeon W36900, 12 GB RAM.
Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Tue Nov 14, 2017 8:31 am Re: PaulStretch3 preview(10) release

Kalamata Kid wrote:VST or VTSi?
touchable?
multi-touchable?

Probably just an effect type "generator" plugin, making it an instrument plugin unfortunately probably doesn't make sense because of the high CPU load, high latency and high memory use that would be required per voice. The CPU load and memory use could be mitigated by only allowing one voice etc, but the latency is something that maybe can't really be solved I suppose. The prebuffering could maybe be removed, but then the host's audio threads would be taxed with spikes of CPU activity. Longer FFT sizes (around over 0.5 seconds, which sound the nicest IMHO) wouldn't probably work at all in typical hosts...

But I have to do some more testing, it appears the CPU performance overall is worse now in the plugin version than it was in the stand alone app version. I might have just made some mistake somewhere...

I don't have a desktop/laptop with a touch screen so I don't know how it behaves with such devices. It should have whatever (multi)touch screen support JUCE has.
Neon Breath
KVRAF
 
2523 posts since 15 Sep, 2010

Postby Neon Breath; Tue Nov 14, 2017 8:46 am Re: PaulStretch3 preview(10) release

@Xenakios, just out of curiosity here. If you go forward and finally make PaulStretch as VST plugin, will you be able to own part of the product and sell it? Or you’ll have to make it free because of the code that was distributed and the creative commons rights ? If you have to give it away, you should at least have a Paypal account so people can make a donation and drop you money, you freakin deserve it!!! :hail:
__________________________________________________
http://soundcloud.com/red-fog
http://soundcloud.com/sleeping-on-lotus-ashes
__________________________________________________
Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Tue Nov 14, 2017 9:04 am Re: PaulStretch3 preview(10) release

Neon Breath wrote:@ Xenakios, just out of curiosity here, if you go forward and finally make PaulStretch as VST plugin, will you be able to own part of the product and sell it? Or you’ll have to make it free because of the code that was distributed and the creative commons rights ? Of you have to give it away, you should at least have a Paypal account so people can make a donation and drop you money, you freakin deserve it!!!

I could attempt selling it, the GPL licensing doesn't prevent that. It however wouldn't make much sense because anyone could grab the source code, build it and then themselves distribute it for free or in turn attempt selling it.

I do have a Paypal account, there's a donate button in my blog, but I haven't encouraged people to give donations for the standalone PaulStretch3 application because so much of the interesting parts of the code were done by Paul, not me. But the plugin version could turn out to be such an involved thing to do, I might encourage donations at least a little bit... ;)
Neon Breath
KVRAF
 
2523 posts since 15 Sep, 2010

Postby Neon Breath; Tue Nov 14, 2017 10:29 am Re: PaulStretch3 preview(10) release

Xenakios wrote:I do have a Paypal account, there's a donate button in my blog, but I haven't encouraged people to give donations for the standalone PaulStretch3 application because so much of the interesting parts of the code were done by Paul, not me. But the plugin version could turn out to be such an involved thing to do, I might encourage donations at least a little bit... ;)


Totally. Telling you, you ALREADY deserve it for all you did.

Cheers
__________________________________________________
http://soundcloud.com/red-fog
http://soundcloud.com/sleeping-on-lotus-ashes
__________________________________________________
Kalamata Kid
KVRAF
 
2862 posts since 26 Jul, 2001, from Tarpon Springs, Florida, USA

Postby Kalamata Kid; Tue Nov 14, 2017 3:24 pm Re: PaulStretch3 preview(10) release

Xenakios wrote:
Kalamata Kid wrote:VST or VTSi?
touchable?
multi-touchable?

Probably just an effect type "generator" plugin, making it an instrument plugin unfortunately probably doesn't make sense because of the high CPU load, high latency and high memory use that would be required per voice. The CPU load and memory use could be mitigated by only allowing one voice etc, but the latency is something that maybe can't really be solved I suppose. The prebuffering could maybe be removed, but then the host's audio threads would be taxed with spikes of CPU activity. Longer FFT sizes (around over 0.5 seconds, which sound the nicest IMHO) wouldn't probably work at all in typical hosts...

But I have to do some more testing, it appears the CPU performance overall is worse now in the plugin version than it was in the stand alone app version. I might have just made some mistake somewhere...

I don't have a desktop/laptop with a touch screen so I don't know how it behaves with such devices. It should have whatever (multi)touch screen support JUCE has.


Ok, got it! It will be a VST fx not a VSTi.
I had no idea about the issues on a VSTi.

As for Juce I understand it makes the plugins multi-touch automatically. Yea!!! For multi-touch the knobs and controls work better if the are more like faders, vertical motion. But I may be the only one here concerned about multi-touch.

I will wait until out of beta and check it out. I am really looking forward to the release. :love: Thanks Xenakios for all the hard work. :)
23" 10-touch Monitor and iPad Air as Midi Controllers, iConnectMIDI2+, Windows 10-64bit, Xeon W36900, 12 GB RAM.
Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Tue Nov 14, 2017 4:25 pm Re: PaulStretch3 preview(10) release

Kalamata Kid wrote:
It will be a VST fx not a VSTi.

It's not outside the realm of possibilities it could work as an instrument (at least for sounds that don't require great rhythmic precision), but at the moment the code is in such a stage I don't want to put in any MIDI note messages handling and so on.
acousmod
KVRist
 
418 posts since 27 Jan, 2004, from France

Postby acousmod; Wed Nov 15, 2017 6:15 am Re: PaulStretch3 preview(10) release

Hello Xenakios,

With the current version, PaulStretch crashes when the stretch value goes below 1.0 with multichannel files.

As for multitouch, the standalone version supports it perfectly. We will just need a little more space for the fingers, and/or perhaps that vertical faders would be more appropriate ?

Thanks a lot !
Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Wed Nov 15, 2017 6:59 am Re: PaulStretch3 preview(10) release

acousmod wrote:Hello Xenakios,

With the current version, PaulStretch crashes when the stretch value goes below 1.0 with multichannel files.

As for multitouch, the standalone version supports it perfectly. We will just need a little more space for the fingers, and/or perhaps that vertical faders would be more appropriate ?



Couldn't reproduce the crashing behavior on Windows with 4 output channels and a 4 channel input file. What channel counts did you try? What FFT size? Was the huge prebuffering mode enabled and an extra large FFT size used? Was the file an uncompressed wav file or a compressed format like flac or mp3?

Yeah, the GUI layout is a bit too tight even for mouse use, I need to look into that...Not sure about the vertical faders thing, because then getting the parameter name labels placed gets a bit problematic. (Too bad I have to do the GUI layouting at the moment by altering the C++ code, so every change takes a relatively long time to test...)
acousmod
KVRist
 
418 posts since 27 Jan, 2004, from France

Postby acousmod; Wed Nov 15, 2017 7:22 am Re: PaulStretch3 preview(10) release

No crash with 4 channels, but always with 6 and more, in Windows 10, uncompressed 24 bits files, prebuffering medium, doesn't seem to depend on other values.
Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Wed Nov 15, 2017 7:26 am Re: PaulStretch3 preview(10) release

acousmod wrote:No crash with 4 channels, but always with 6 and more, in Windows 10, uncompressed 24 bits files, prebuffering medium, doesn't seem to depend on other values.

OK, got it to crash with 8 channel input file, I'll investigate. Thanks for the report!
acousmod
KVRist
 
418 posts since 27 Jan, 2004, from France

Postby acousmod; Wed Nov 15, 2017 7:32 am Re: PaulStretch3 preview(10) release

Thanks !
I have also the same behaviour on OSX with the defaults settings.
Xenakios
KVRian
 
1004 posts since 9 Sep, 2005, from Oulu, Finland

Postby Xenakios; Wed Nov 15, 2017 7:35 am Re: PaulStretch3 preview(10) release

This is a very weird bug since it indeed seems to only happen with stretch amounts lower than 1.0...And it's about the channel counts going wrong somewhere. Those should not be related in any way. This might take some time to figure out... edit : Well, actually they were related, with the shortening mode a code path is hit that has rarely happened when just slowing down... :)
PreviousNext

Moderator: Moderators (Main)

Return to Effects