v5.95pre13 - August 26 2018

Edgemeal · 08-26-2018, 06:38 PM

v5.95pre13 - August 26 2018
+ Automation items: duplicating items with AIs will obey trim content behind automation items preference
+ JSFX: fix mdct() concurrency issue (bug dated to 2004, diagnosis thanks to Lauri Liinat)
# Paste: fix undo following AI autoremoval [p=2027654]

Full changelog - Latest pre-releases

X-Raym · 08-26-2018, 10:09 PM

2004 ! Never too late :P

clepsydrae · 08-26-2018, 11:03 PM

Is that the record for the oldest-fixed reaper bug now? I remember there was a really old one a few months back... :-)

X-Raym · 08-26-2018, 11:32 PM

From native REAPER, here are the affected files:

filters\mdct-filter
filters\mdct-volsweep
pitch\mdct-shift
sstillwell\thunderkick

I didn't find any Reapack shared JSFX with this function.

Dannii · 08-27-2018, 03:09 AM

Quote:

Originally Posted by X-Raym

From native REAPER, here are the affected files:

filters\mdct-filter
filters\mdct-volsweep
pitch\mdct-shift
sstillwell\thunderkick

I didn't find any Reapack shared JSFX with this function.

Out of curiosity, what exactly does this fix?

Justin · 08-27-2018, 04:45 AM

Quote:

Originally Posted by ReaDave

Out of curiosity, what exactly does this fix?

If you had MDCT-using JSFX on multiple tracks, there was a concurrency issue where the output of one or both could be corrupted (e.g. one track JSFX could have completely silent input and it would start getting signal/noise from the other).

Quote:

Is that the record for the oldest-fixed reaper bug now? I remember there was a really old one a few months back... :-)

I think so! The other old bug was one with the copy/paste code not skipping more than 32 tracks, which was late 2005. This one is older, though the bug likely didn't manifest until 2006 when multiprocessing support was added to REAPER (JSFX and REAPER originally didn't have multiprocessing...)

Dannii · 08-27-2018, 08:52 AM

Quote:

Originally Posted by Justin

If you had MDCT-using JSFX on multiple tracks, there was a concurrency issue where the output of one or both could be corrupted (e.g. one track JSFX could have completely silent input and it would start getting signal/noise from the other).

Ah. Very interesting. This might explain some of the weirdness I've experienced with Trond's JS Ambisonic Toolkit. I've had quite a few occurances where outputs would suddenly just stop working (or the signal would become very distorted) when adding extra JS plugins when also using the ATK. I'm going to have a play around with the ATK again and see if this has cleared up.

This also occured when changing i/o pin configurations.

Justin · 08-27-2018, 09:57 AM

Quote:

Originally Posted by ReaDave

Ah. Very interesting. This might explain some of the weirdness I've experienced with Trond's JS Ambisonic Toolkit. I've had quite a few occurances where outputs would suddenly just stop working (or the signal would become very distorted) when adding extra JS plugins when also using the ATK. I'm going to have a play around with the ATK again and see if this has cleared up.

This also occured when changing i/o pin configurations.

Probably not, unless it uses mdct()

Dannii · 08-27-2018, 10:16 AM

Quote:

Originally Posted by Justin

Probably not, unless it uses mdct()

Hmmm.... Will have to have a look through the code and see. I've managed to hack my way around modding a few JS plugins for third order (16 channel) Ambisonic use but I'm a LONG way from a full understanding of JS code at this stage. I have no idea what mdct() actually does (guess I'll have to do some reading).
It's just that your description sounds pretty much like the exact symptoms I've seen with the ATK so I kinda got my hopes up that this was the issue.

EvilDragon · 08-27-2018, 10:42 AM

https://en.wikipedia.org/wiki/Modifi...sine_transform

D Rocks · 08-27-2018, 10:51 AM

Quote:

Originally Posted by EvilDragon

https://en.wikipedia.org/wiki/Modifi...sine_transform

Pretty complex lol

mikeroephonics · 08-29-2018, 01:31 AM

I can't get some of my MIDI devices to work on 5.95pre13. The devices which won't work have "!!! N/A: Enabled" under their "Mode" column in MIDI hardware settings.

The "Error opening devices" window displays upon REAPER launching:

The following MIDI inputs could not be opened:
[18i20]_[5-PIN MIDI INPUT]
2- Scarlett 18i20 USB
nanoKONTROL2 1 SLIDER/KNOB
[18i20]_[5-PIN MIDI IN]
Yamaha MOTIF ES-1_[USB MIDI]
nanoKONTROL2
nanoKONTROL2 SLIDER/KNOB
A-PRO 1
4- Scarlett 18i20 USB
A-PRO 2
loopMIDI Port 14
loopMIDI Port 15
[18i20]_[5-PIN MIDI INPUT]
The following MIDI outputs could not be opened:
2- A-PRO
nanoKONTROL2
3- Scarlett 18i20 USB
Yamaha MOTIF ES-1
Yamaha MOTIF ES-2
Yamaha MOTIF ES-3
Yamaha MOTIF ES-5
Yamaha MOTIF ES-6
Yamaha MOTIF ES-7
Yamaha MOTIF ES-8

Justin · 08-29-2018, 03:41 AM

Quote:

Originally Posted by mikeroephonics

I can't get some of my MIDI devices to work on 5.95pre13. The devices which won't work have "!!! N/A: Enabled" under their "Mode" column in MIDI hardware settings.

The "Error opening devices" window displays upon REAPER launching:

The following MIDI inputs could not be opened:
[18i20]_[5-PIN MIDI INPUT]
2- Scarlett 18i20 USB
nanoKONTROL2 1 SLIDER/KNOB
[18i20]_[5-PIN MIDI IN]
Yamaha MOTIF ES-1_[USB MIDI]
nanoKONTROL2
nanoKONTROL2 SLIDER/KNOB
A-PRO 1
4- Scarlett 18i20 USB
A-PRO 2
loopMIDI Port 14
loopMIDI Port 15
[18i20]_[5-PIN MIDI INPUT]
The following MIDI outputs could not be opened:
2- A-PRO
nanoKONTROL2
3- Scarlett 18i20 USB
Yamaha MOTIF ES-1
Yamaha MOTIF ES-2
Yamaha MOTIF ES-3
Yamaha MOTIF ES-5
Yamaha MOTIF ES-6
Yamaha MOTIF ES-7
Yamaha MOTIF ES-8

See: https://forum.cockos.com/showthread.php?t=209673

airon · 08-29-2018, 07:08 AM

Yeah, I removed a bunch of old entries in the list, the type of stuff I thought I'd buried in the garden. Zombie gear.

Everything worked fine after that.

juliansader · 08-30-2018, 06:47 AM

Since the current pre-releases are working on AI moving/copying again, I would like to revisit an issue that was raised a year ago in the AI pre-release thread:

Quote:

Originally Posted by juliansader

At present (v5.60rc), moving or copying media items together with automation items is an uncomfortable mixture of
* The old option "Envelope points move with media items", and
* Selection status of automation items.

If the option is inactive, no automation (not even selected AIs) move with MIs.
If the option is active, selected AIs (even those outside the MI boundaries) move with the MI, but UNselected AIs below the MI also move (even those in hidden envelope lanes).

I think it would be useful if unselected AIs could (optionally?) be excluded from moving/copying/duplicating actions, similar to how unselected MIs are ignored.** This would allow the user to accurately specify which MIs and automation should be moved or copied.

I suggest that *selected* AIs should always move and copy together with selected MIs.

* If "Envelope points move with media items" is ENabled, and all AIs are UNselected (or the project does not contain any AIs), everything will work as in the olden days: all automation that are below the MIs (or within the time selection, depending on the action), whether in AIs or not, are moved/copied along with the MIs.

* If "Envelope points move with media items" is DISabled, *selected* AIs and *selected* MIs move together, copy together, and Ctrl+leftdrag copy/split "selected area within time selection" together -- similar to how selected MIs work together. Non-AI envelopes as well as unselected AIs are ignored.

As mentioned above, this would allow the user to accurately specify which MIs and automation should be moved or copied. The automation may be longer or shorter than the MI, and sometimes only a subset of the visible envelope lanes should be copied.