OpenMPT Issue Tracker - Issues

0001042: Render pattern channels to separate buffers

Fri, 29 May 2026 07:07:01 +0000

Suggested by emoon (and some other people in the past): Provide a libopenmpt interface for rendering each channel separately, e.g. so that you can draw scopes for each pattern channel.
OpenMPT could also benefit from this.
Plugins make this more complicated because if two channels feed their output into one plugin, we cannot tell them apart anymore. Renoise "solves" this by only showing the plugin output on one of the channels (the last one that triggered a note, I think).
We can already render channels to different buffers (e.g. for plugins or reverb), so the infrastructure is there, it just cannot be used to its full extent at the moment.

0001957: libopenmpt API for MIDI output data

Wed, 27 May 2026 15:34:14 +0000

It might be desirable to, in addition to the rendered PCM output, provide access to realtime MIDI data that would be sent to external devices or unavailable plugins.

0001972: XM instrument rendered significantly quieter with libopenmpt than inside OpenMPT

Wed, 27 May 2026 11:55:35 +0000

I noticed a specific instrument/note sounding much clearer/louder while playing in OpenMPT (Windows) than on a app (actually 2) using libopenmpt.
I have a website that plays mods, this issue occurs there, in any OS.
The global volume of the song seems similar but this particular instrument sounds different.
I already turned the world around but couldn't find the cause.

0000779: Allow certain playback-related flags to be set as load-time ctl or during playback

Wed, 27 May 2026 11:53:02 +0000

Some internal flags should be able to be chosen by the user (e.g. the currently unimplemented flag for playing Amiga MODs as mono, force-enabling VBlank timing for ProTracker MODs or super-smooth ramping for XM files).
It would be rather simple to allow those things to be specified through ctl values, but ultimately it would be very useful to also be able to get and set them in real-time (e.g. through openmpt_module_*_render_param), as e.g. a module player should be able to change some of these settings in real-time.

0001971: Looping problem in OpenMPT

Sun, 24 May 2026 05:37:51 +0000

Hi )
Now available problem and as I think not correct work OpenMPT plugin , what I want mean...
If I not want use repeating and looping file I check repeat never , with this setting all files will play once and next will stoped
If I want do looping playlist , me need check this setting and check repeating or looping playlist in player and this logical . but if module available jump to back with looping zone - I can not hearing correct result if do set repeat never , if I set repeat once - I get one replay and all , but what if I want not do looping all modules , but want recive corretly forever looping separate modules ?
Examples :
https://disk.yandex.ru/d/9O5NXOqieNhnFA - this without looping
https://disk.yandex.ru/d/WsIPhuqDNAOtMA - this file must be looping , but in currently settings absent option auto , I mean sayd what OpenMPT can not detect is it necessary to loop the module?
Please , add option besides Never , Forever , Once , method Auto , I offer do auto as automatic detect is there any need looping in data of tracker

0001772: Move MIDI output off audio thread

Sat, 16 May 2026 15:10:05 +0000

While working on [Beyond](https://sagamusix.de/download/beyond/) my ASIO driver was struggling due to the MIDI driver taking its time to process outgoing MIDI messages to seven different devices. I modified the MIDI code locally to run on a separate thread (one per device actually), but that solution is not quite in a merge-able state yet (there are still some random deadlocks when closing a module). But the direction is right and should be used for more stable MIDI output.

0001968: Accessibility issues with new sample grid settings

Tue, 12 May 2026 16:01:36 +0000

The new sample grid options have a few accessibility and keyboard usability problems:
1. The "grid segments" static text is wrongly presented by screen readers as belonging to the segment size edit field.
2. The "Divide every" radio button seems to be keyboard-unreachable.

0000612: Sample Tab: Allow separate selection of sample channels

Sat, 25 Apr 2026 17:55:41 +0000

It should be possible to edit the right and left channel of a sample individually. Right now, it is only possible to edit both channels at the same time.

0001967: UltraTracker bug

Tue, 21 Apr 2026 21:16:09 +0000

Available bug , yes , all in UltraTracker , if you play this melody from 2:10 , so you can hearing as on 2:30 - 2:38 not accurately reproduced note with pitch shifting , pitch rises too sharply

Source melody https://disk.yandex.ru/d/V65d4Ly1uh-RAw

Modplug Player 1.46 https://disk.yandex.ru/d/QOk0U2JLxT6Dtw

OpenMPT plugin 8.6 for XMPlay https://disk.yandex.ru/d/SwO6uLKYzAn-cA

0000917: Virtualize MIDI I/O

Sun, 19 Apr 2026 14:53:42 +0000

Currently each MIDI I/O plugin instance as well as the separate MIDI recording feature directly access RtMidi or the WinAPI. This makes it impossible to re-use a device, e.g. to create to MIDI I/O instances reading MIDI data from the same MIDI device and then passing it on to different plugins.

The actual MIDI I/O should be moved to a separate component (using RtMidi in the background) that sends the inputs to all listeners and processes all outputs.

0001966: sizeof(chunk) in SetChunk() measures span object size instead of PluginChunk size on 64-bit

Fri, 17 Apr 2026 16:04:35 +0000

In DigiBoosterEcho.cpp:210 and SymMODEcho.cpp:244, the SetChunk() function uses sizeof(chunk) where chunk is a const ChunkData& (i.e. mpt::const_byte_span&). On 64-bit platforms, sizeof(mpt::const_byte_span) is 16 bytes, but the target m_chunk (PluginChunk) is only 8 bytes (verified by static_assert in the header).
This causes the condition chunk.size() == sizeof(chunk) to require 16 bytes of input data, and the subsequent memcpy(&m_chunk, data, chunk.size()) writes 16 bytes into an 8-byte struct, overflowing 8 bytes into adjacent class members.
The correct check should be chunk.size() == sizeof(m_chunk), which is already used correctly in RestoreAllParameters() in the same file (line 98).
Affected files:
- soundlib/plugins/DigiBoosterEcho.cpp line 210
- soundlib/plugins/SymMODEcho.cpp line 244
Note: On 32-bit platforms sizeof(span) may happen to equal sizeof(PluginChunk) (both 8), masking the bug.

0001964: Vst3Shell 1.4 not working in OpenMPT

Sat, 11 Apr 2026 22:15:44 +0000

I installed a plugin called Misstortion2, and installed it via the installer that you get when you download it from a website. It only comes in Vst3 and I found out OpenMPT doesn't support Vst3 after failing to load Misstortion2 directly. So I do some looking and decide to try Vst3Shell as it looks the simplest and easiest to use as I like keeping things minimal.
I have Misstortion2 installed to C:\Program Files\Common Files\VST3, and all of Vst3Shell's components installed to C:\Program Files\Common Files\VST3\vst3shell_v1.4, however when I add Vst3Shell to OpenMPT as a plugin (after scanning the vst3shell_v1.4 folder in OpenMPT for new plug-ins) the "Editor" button is grayed out and I can't open it or use it.
Nothing happens afterwards, I installed my plugins into what I believe is the standard location, I don't know what I'm doing wrong.

Thanks for the help, means a lot

0001962: VS code generation bug leads to crash in plugin editor

Wed, 08 Apr 2026 07:17:31 +0000

I'll try to use one plugin called 'SQ8L' and for some reason when I select the 2nd option on it on my PC and as soon as I press a button. It suddenly crashes and shows window I have a screenshot of.

0001929: Main toolbar spin buttons disappear when locking PC on Windows 11

Mon, 30 Mar 2026 16:22:22 +0000

This somehow doesn't happen on Windows 10, but I see it on Windows 11. After coming back to OpenMPT from the Windows lock screen, the main toolbar's spin buttons disappear. They re-appear when moving / adjusting the toolbar. This is new in OpenMPT 1.32 and doesn't happen when not setting the corresponding edit fields as buddy controls.

It would be nice if this could be fixed without having to back to not setting buddy controls, as this looks better and is more correct.

0001960: crash with read-only mptrack.ini

Tue, 24 Mar 2026 11:54:23 +0000

If simply upgrade OpenMPT Tracker to latest version and leave mptrack.ini from previous version , so OpenMPT Tracker not runing , process in task manager available , but work window does not appear

0001953: UltraTracker 16-bit sample loops clamped incorrectly

Sun, 22 Mar 2026 18:17:53 +0000

Hi )
In UltraTracker available bug , very short samples in this melody

0001017: Tick boundary rendering (was: Provide access to next play position)

Sat, 21 Mar 2026 12:55:40 +0000

PoroCYon asked on IRC if it's possible to retrieve the next play position (row/pattern). We already have this information, but it is currently not exposed.

0001952: Retrigger note volume adjustment is not applied to MIDI velocity

Sat, 21 Mar 2026 11:34:04 +0000

As the title states, using Qxx with a value other than Q0x on a MIDI instrument does not impact the velocity of the outputted notes with each re-trigger.

0001951: Keep loop pattern on when using Space to play song

Wed, 11 Mar 2026 22:11:26 +0000

When using Space to play a track with Loop Pattern turned on, the loop keeps disabling. There's some more information here: https://www.reddit.com/r/OpenMPT/comments/1k16k60/comment/o1q6o7y/?context=1 - as explained, if using pause instead of stop, the loop pattern doesn't turn off, but there should be a way to force Loop Pattern to stay on persistently.

0000925: Support for external samples in libopenmpt

Mon, 23 Feb 2026 17:03:49 +0000

libopenmpt does currently not support sample data (or any other kind of data) outside the user-provided module file. There are several multi-file formats supported by OpenMPT:

* MPTM (external samples)
* MT2 (external samples)
* ITP (external instruments)
* Startrekker FLT/EXO modules (external synthesized instruments)
* Symphonie Pro (external samples)
* Face The Music (external samples)
* Graoumf Tracker 2 (external samples)
* SoundSmith (external samples)

In the future, we may also support:
* TFMX (external samples, no plans to support this format at the moment)
* Other formats

Apart from the fact that someone may simply want to play such modules using a libopenmpt-based player, there is also the possibility of someone wanting to use external samples e.g. for saving space when composing music for a game.

There are also security implications when accessing external files, and the feature should probably be opt-in (and possibly fully backwards-compatible with previous library versions, e.g. by using a ctl value for providing the module file path).

0001950: have an all modules (without wav and midi) and all supported files (including WAV and MIDI) selection seperately

Fri, 20 Feb 2026 20:31:17 +0000

for the module file open thing in openmpt add an option for all modules and seperately one for all files

like one option that covers all but also not wav and midi cuz i always have to manually filter them out �

0001530: CLAP plugin support

Mon, 16 Feb 2026 10:12:01 +0000

A new plugin format that several well-known plugin and host manufacturers want to commit to.

https://www.kvraudio.com/forum/viewtopic.php?t=574861
https://github.com/free-audio

0001366: LV2 support

Mon, 16 Feb 2026 10:12:01 +0000

LV2 support would be really nice. there are a lot of really good open-source plugins - and more being released all the time - that unfortunately i can't yet use with openmpt. presumably these would have better performance as well, being native code that doesn't have to go through wine?

i haven't seen an issue for this yet, so i figured it might as well be me.

0000783: cross-platform OpenMPT

Mon, 16 Feb 2026 09:27:09 +0000

This is a meta-issue, collecting various other issues related to making OpenMPT itself cross-platform.

0001559: Don't use WinAPI functions for INI reading/writing

Sun, 15 Feb 2026 09:10:06 +0000

Currently OpenMPT uses `ReadPrivateProfile*` / `WritePrivateProfile*`functions provided by Windows for reading and writing its settings files. This has several issues

- It's relatively slow, which is why we only write setting that we know have changed
- It requires hacks to enforce writing unicode files

As a consequence of the first point, it's possible that one OpenMPT instance takes a long time to shut down, and a newly launched instance may not see the settings as saved by the previous instance. One particularly worrying example is the following scenario:

1. Fresh OpenMPT installation, user sees the Welcome dialog and unchecks automatic update checks
2. User closes OpenMPT. This is the first time the INI file is written, so it takes quite a while to write out all settings.
3. User immediately reopens OpenMPT. It may now try to search for updates because the old instance might not have finished writing the settings yet.

A custom INI implementation should avoid this scenario. While the file is being written, other OpenMPT instances should wait until the file is no longer locked, rather than reading incomplete settings.

OpenMPT Issue Tracker - Issues

0001042: Render pattern channels to separate buffers

0001957: libopenmpt API for MIDI output data

0001972: XM instrument rendered significantly quieter with libopenmpt than inside OpenMPT

0000779: Allow certain playback-related flags to be set as load-time ctl or during playback

0001971: Looping problem in OpenMPT

0001772: Move MIDI output off audio thread

0001968: Accessibility issues with new sample grid settings

0000612: Sample Tab: Allow separate selection of sample channels

0001967: UltraTracker bug

0000917: Virtualize MIDI I/O

0001966: sizeof(chunk) in SetChunk() measures span object size instead of PluginChunk size on 64-bit

0001964: Vst3Shell 1.4 not working in OpenMPT

0001962: VS code generation bug leads to crash in plugin editor

0001929: Main toolbar spin buttons disappear when locking PC on Windows 11

0001960: crash with read-only mptrack.ini

0001953: UltraTracker 16-bit sample loops clamped incorrectly

0001017: Tick boundary rendering (was: ~~Provide access to next play position~~)

0001952: Retrigger note volume adjustment is not applied to MIDI velocity

0001951: Keep loop pattern on when using Space to play song

0000925: Support for external samples in libopenmpt

0001950: have an all modules (without wav and midi) and all supported files (including WAV and MIDI) selection seperately

0001530: CLAP plugin support

0001366: LV2 support

0000783: cross-platform OpenMPT

0001559: Don't use WinAPI functions for INI reading/writing

0001017: Tick boundary rendering (was: Provide access to next play position)