Roland User wrote on 2024-06-02, 16:16:

And got a question )
When I see structure BASSMIDI , I see this :
maybe it makes sense compile in one and add files as resources ?

No , it does not make any sense since:
1. Antivirus engines make it hard to release new versions of any binaries for small developers even without any shady practices.
Embedding dll files as resources that are extracted and executed at runtime would quadruple the amount of false positive detection since it is considered a shady practice by most AV engines.
I suggest that you should try this practice and make your tests on Virustotal before recommending this to anyone...
2. It would make debugging and testing more difficult.
3. It has no real world practical advantages but it can generate such errors in edge cases that are hard to foresee (e.g. extraction on a read-only drive etc.).

PS:
This was the last time I answered to any of your ideas, suggestions, etc. Your quota has run out.
From now on I will only answer to real bug reports from you.

Why if set big wave buffer and play this file , will be this ? Big wave buffer is 8192 samples and more )
https://disk.yandex.ru/i/8Rj0gM5gp3cH6A
Why WinMultMIDI lost ing messages ?
Problem in WinMultMIDI because is happening partial internal freeze , and next is happening fully frize
And yes , I use this plugin not so , as was intended ) , I use this plugin for other target ) and I will be use this plugin as use now , that's why I and write to you questions and ideas )
Why only to you ? All simply , I not known other VST plugin and developer with this functional who I can contact ) especially if is freeware )

Roland User wrote on 2024-06-03, 17:27:

Why if set big wave buffer and play this file , will be this ? Big wave buffer is 8192 samples and more ) https://disk.yandex.r […]
Show full quote

Why if set big wave buffer and play this file , will be this ? Big wave buffer is 8192 samples and more )
https://disk.yandex.ru/i/8Rj0gM5gp3cH6A
Why WinMultMIDI lost ing messages ?
Problem in WinMultMIDI because is happening partial internal freeze , and next is happening fully frize
And yes , I use this plugin not so , as was intended ) , I use this plugin for other target ) and I will be use this plugin as use now , that's why I and write to you questions and ideas )
Why only to you ? All simply , I not known other VST plugin and developer with this functional who I can contact ) especially if is freeware )

Re: Falcosoft Soundfont Midi Player

This MT-32 midi file contains more than 20KB Sysex data. With your large buffer all the Sysex data come in 1 block and this can cause problems. In case of VSTi plugins the buffer size not only affects audio data/latency but also the frequency/amount of the incoming midi data.

And which audiobuffer you can recomended if your recomendation will be diffirent of zero ?

Roland User wrote on 2024-06-03, 18:37:

And which audiobuffer you can recomended if your recomendation will be diffirent of zero ?

The lower the better. But the right buffer size is also sample rate dependent.

In case of 48000 Hz (buffer size in samples):
240 (FSMP uses this with real-time API)
480 (FSMP uses this with Stream API)
960
1920 (max. reasonable value)

in case of 96000 Hz (buffer size in samples):
480 (FSMP uses this with real-time API)
960 (FSMP uses this with Stream API)
1920
3840 (max. reasonable value)

Thank you ) will try
If return a discussion about WinMM Multi Port VSTi and my offers , so I can not understand one thing in FSMP this have , but in WinMM Multi Port VSTi you very not want and resist adding what already exists in FSMP ) why so ? I say about Select Custom SysEx ) it would remove close all questions about SysEx in WinMM Multi Port VSTi and not only from me )
I realy can not understand you in this )
Please do not get angry with me

Roland User wrote on 2024-06-03, 19:19:

Thank you ) will try If return a discussion about WinMM Multi Port VSTi and my offers , so I can not understand one thing in FS […]
Show full quote

Thank you ) will try
If return a discussion about WinMM Multi Port VSTi and my offers , so I can not understand one thing in FSMP this have , but in WinMM Multi Port VSTi you very not want and resist adding what already exists in FSMP ) why so ? I say about Select Custom SysEx ) it would remove close all questions about SysEx in WinMM Multi Port VSTi and not only from me )
I realy can not understand you in this )
Please do not get angry with me

There is no return to any previous discussions.
The point is you do not have to understand this, you simply have to accept this.
Do you remember our deal?:
Re: Falcosoft Soundfont Midi Player + Munt VSTi + BassMidi VSTi

Yes , I remember this , just wanted to see improvements in this plugin ) and thinked what this not will problem ) but I was wrong )

I want clarify , you say about timings interconnected with Sample Rate
as need write loader if have this condition

On Error Resume Next
Set WshShell = CreateObject("WScript.Shell")
WshShell.RegWrite "HKCU\Software\Seib\WinMultMIDI\Settings\MidiIn", "MIDI Input", "REG_SZ"
WshShell.RegWrite "HKCU\Software\Seib\WinMultMIDI\Settings\WaveOut", " ", "REG_SZ"
WshShell.RegWrite "HKCU\Software\Seib\WinMultMIDI\Settings\SampleRate", 0, "REG_DWORD"
WshShell.RegWrite "HKCU\Software\Seib\WinMultMIDI\Settings\WaveOutBufSize", 0, "REG_DWORD"

Can use in here 0 as samplerate and buffersize or it is forbidden?
Or force need set any samplerate and wavebuffer = 1/100 for samplerate to streaming ?

Roland User wrote on 2024-06-03, 19:29:

Can use in here 0 as samplerate and buffersize or it is forbidden?
Or force need set any samplerate and wavebuffer = 1/100 for samplerate to streaming ?

You are the SAVIHost user not me so you could check it for yourself, but OK:
It's not forbidden but it is not valid either so it results in default values (44100 sample rate , and 2205 buffer size).
It seems SAVIHost can only use 44100 Hz sample rate when no Wave out device is selected. But still you had better use explicit valid values: 44100 sample rate and 210 buffer size (it should work well for both real-time and stream modes).

I can not understand why WinMM Multi Port VSTi if I often replace MIDI output devices on one port , plugin freeze , why so ?
This will need when I compare diffirence with versions of synth , me need fast switch synths for compare , and in this mode WinMM Multi Port VSTi freeze )
Freezin fully , closin only via taskmanager , why so ?
I understand what I say about use mode which you didn't plan

Roland User wrote on 2024-06-04, 18:51:

I can not understand why WinMM Multi Port VSTi if I often replace MIDI output devices on one port , plugin freeze , why so ? Thi […]
Show full quote

I can not understand why WinMM Multi Port VSTi if I often replace MIDI output devices on one port , plugin freeze , why so ?
This will need when I compare diffirence with versions of synth , me need fast switch synths for compare , and in this mode WinMM Multi Port VSTi freeze )
Freezin fully , closin only via taskmanager , why so ?
I understand what I say about use mode which you didn't plan

I have no idea since I cannot reproduce it. Try to stop playback before changing Midi out devices on ports.

BassMidi VSTi 1.6 has been released:
https://falcosoft.hu/softwares.html#bassmidi_vsti

version 1.6

1. Added option to disable Sound Blaster hardware limits emulation.
2. Added option to use bank 127 in XG mode as XG drum bank.
3. Added SF3 format support to file select dialog.
4. Fixed Bassmidi regression that resulted in wrong preset names of bank 126 in XG mode.
5. Fixed rare visual glitch in right click context menu of reset button.
6. Added basswv.dll to support WavPack compressed samples.
7. Updated other Bass libraries.

Thank you )
Only parameter Use bank 127 for XG Drums work in inverted mode If I select this , in XG I can select MT-32 Patches , and if I deselect this parameter I can use XG Drums , but if I select this parameter in GS mode or any other , nothing not is happening
Video https://disk.yandex.ru/i/LwLQGbOHV330pA

Roland User wrote on 2024-06-09, 09:16:

Thank you )
Only parameter Use bank 127 for XG Drums work in inverted mode If I select this , in XG I can select MT-32 Patches , and if I deselect this parameter I can use XG Drums , but if I select this parameter in GS mode or any other , nothing not is happening
Video https://disk.yandex.ru/i/LwLQGbOHV330pA

No, it does not work 'inverted'. You just simply misunderstand this option.
1. As the option's name suggests it only has effect in XG mode. It does nothing in any other modes.
2. In XG mode you need an XG compatible soundfont in order this option to work properly (that is a soundfont which has 2 drum banks: an XG compatible drum bank at bank 127 and a standard GS drum bank at bank 128).
If you have a standard GM/GS soundfont you should never enable this option.
Without this option always bank 128 (the standard SF2 drum bank) is mapped to bank 127 in XG mode and also bank 128 is used on channel 10 (drum channel).
When this option is enabled then the real bank 127 of the SF2 soundfont is used when bank 127 is selected and also bank 127 of the SF2 soundfont is used on channel 10 (drum channel) in XG mode.

Here is an XG compatible soundfont so you can test this option:

Ps:
Here is the documentation about what this option does (BASS_MIDI_FONT_XGDRUMS flag):
https://www.un4seen.com/doc/#bassmidi/BASS_MI … I_FontInit.html
So if you do not like how it works and you have suggestions about a different behavior then you should convince Ian not me (as I have already told you previously).

Thank you ) I understand )
Also you sayd what added option "Do not emulate SB limits" what are the restrictions this option removes ? )

Roland User wrote on 2024-06-09, 10:34:

Thank you ) I understand )
Also you sayd what added option "Do not emulate SB limits" what are the restrictions this option removes ? )

Mostly LFO and filter cutoff limits but you should ask Ian:
https://www.un4seen.com/forum/?topic=20425.ms … 43029#msg143029

I have a question about this software to Falconsoft.

I've used this MIDI player on modern machine (with VST) without any problem, untill I tried using it on other machine with REAL hardware midi synth. Reason is I've got an older PC with two HARDWARE SYNTH Yamaha PCI soundcards installed (SW1000XG and one of YMF7x4 reference soundcards with XG midi synth onboard). PC itself is rather fast. It has pentium G2120 dualcore, ssd onboard and runs windows xp really smoothly. It is my, say, dedicated MIDI PC to use those great MIDI soundcards. The problem is that the MIDI player tends to experience some lags with hardware MIDI playback.
Tempo may vary during playback, there may be subtle hangs and glitches. Some sort of lags and jitter with MIDI events being played by synth.
This must have something to do with the clock, sync and other such things, because MIDI devices are pure hardware.
I'm no programmer, but I understand that such things cannot happen because of 'slow cpu', bad Windows and any issues of this kind in such a case.
If I had such a PC back in 90s I wouldn't believe it's real. I don't really know how powerful it is in comparison to PCs in 1997-2003. 10x as powerful? 100x?
I've tried some other MIDI players and they don't tend to have such issues.

Galova wrote on 2024-06-26, 20:45:

I have a question about this software to Falconsoft. […]
Show full quote

I have a question about this software to Falconsoft.

I've used this MIDI player on modern machine (with VST) without any problem, untill I tried using it on other machine with REAL hardware midi synth. Reason is I've got an older PC with two HARDWARE SYNTH Yamaha PCI soundcards installed (SW1000XG and one of YMF7x4 reference soundcards with XG midi synth onboard). PC itself is rather fast. It has pentium G2120 dualcore, ssd onboard and runs windows xp really smoothly. It is my, say, dedicated MIDI PC to use those great MIDI soundcards. The problem is that the MIDI player tends to experience some lags with hardware MIDI playback.
Tempo may vary during playback, there may be subtle hangs and glitches. Some sort of lags and jitter with MIDI events being played by synth.
This must have something to do with the clock, sync and other such things, because MIDI devices are pure hardware.
I'm no programmer, but I understand that such things cannot happen because of 'slow cpu', bad Windows and any issues of this kind in such a case.
If I had such a PC back in 90s I wouldn't believe it's real. I don't really know how powerful it is in comparison to PCs in 1997-2003. 10x as powerful? 100x?
I've tried some other MIDI players and they don't tend to have such issues.

I'm activley using Midi Player on Win9x and Win XP with real hardware regularly (Athlon XP 1700 from 2002 and P1 MMX 166 from 1997 + SC-55) and I do not experience any such problems. Midi Player works perfectly. I really do not think your problem is clock precision related. Most likely the problem is with either video or audio hardware. You can try a few things to determine what can cause your problem:
1. Disable audio recording (that is always active by default so that the volume meter to work). Device Settings -> Rocording Source -> Input Port -> press the minus sign.
2. Try to shrink Midi Player to minimal view so that no controller knobs/virtual piano are visible and close channel view/visualization dialog. That is:

Test these and report back.

Falcosoft wrote on 2024-06-26, 21:59:

Test these and report back.

Well, shrinking the player interface seems to help. Maybe not perfectly but it does for sure. I've also disabled recording device the way you pointed by removing it from the list clicking the minus button, just in case.
My PC has a cpu integrated intel graphics as it can be seen in cpu name. It is Pentium G2120. G goes for graphics core obviously. I've just haven't noticed such issues even with youtube videos watched from that PC which should be a lot more resource hungry than a MIDI player. So it seems like player animation or something like that makes it lag. First I thought it may be related to multicore support, because some software tends to have issues with it. My PC runs dual core CPU and Windows XP doesn't fully support multicore CPUs. It supports dual core only and it is not as full as in later Windows versions. So this may have issues sometimes. Programms running in single-threaded mode for example may fully load some core while keeping other cores idle. Windows 98 doesn't support it at all.
Is it possible to optimize it some way?
Should I install some discrete video card instead of using integrated one? What is your opinion? May it help? I didn't know that integrated graphics may have issues with some software. I just have no idea where to get a working windows 98/XP compatible video card these days. I used to have one, but I'm afraid it's dead. Since lags depend on visualization it must be GPU or GUI related.

I would also kindly ask you to add a feature allowing to show and hide animated piano keyboard without hiding everything else with it, because I really like your player and do not want to use another one.

Galova wrote on 2024-06-26, 23:28:

Well, shrinking the player interface seems to help. Maybe not perfectly but it does for sure. I've also disabled recording devic […]
Show full quote

Falcosoft wrote on 2024-06-26, 21:59:

Test these and report back.

Well, shrinking the player interface seems to help. Maybe not perfectly but it does for sure. I've also disabled recording device the way you pointed by removing it from the list clicking the minus button, just in case.
My PC has a cpu integrated intel graphics as it can be seen in cpu name. It is Pentium G2120. G goes for graphics core obviously. I've just haven't noticed such issues even with youtube videos watched from that PC which should be a lot more resource hungry than a MIDI player. So it seems like player animation or something like that makes it lag. First I thought it may be related to multicore support, because some software tends to have issues with it. My PC runs dual core CPU and Windows XP doesn't fully support multicore CPUs. It supports dual core only and it is not as full as in later Windows versions. So this may have issues sometimes. Programms running in single-threaded mode for example may fully load some core while keeping other cores idle. Windows 98 doesn't support it at all.
Is it possible to optimize it some way?
Should I install some discrete video card instead of using integrated one? What is your opinion? May it help? I didn't know that integrated graphics may have issues with some software. I just have no idea where to get a working windows 98/XP compatible video card these days. I used to have one, but I'm afraid it's dead. Since lags depend on visualization it must be GPU or GUI related.

I would also kindly ask you to add a feature allowing to show and hide animated piano keyboard without hiding everything else with it, because I really like your player and do not want to use another one.

I could also test with Turion64 X2 and Phenom 2 X4 under WinXP and found no problems. So there is no general multi-core problems on Win XP either.
But you can test your theory by selecting Main menu -> Compatibility Settings -> Use only 1 CPU core option. You can also try Main menu -> Priority -> Above Normal.
And you should also test one more thing: Select a different standard WinMM Midi out port (e.g. MS GS Softh Synth if there is no other) to determine if the problem is somehow port related.
You can also test BassMidi/VSTi mode since actually these modes are more CPU hungry than WinMM Midi Out modes.
Unfortunately I cannot optimize if I do not know what to optimize. So without being able to reproduce the problem my only chance is that your testing can tell what is the edge case that is responsible for your problem.

PS:
To analyze the situation further please check the "TGT Interval" with the attached program. It should be 1 when Midi Player's playback is active regardless of used ports.

@Edit:
I have made a test video with a Phenom II X4 from 2009 at 2.5 GHz that is comparable to your system in terms of performance (actually Pentium G2120's single core performance is much better) .
I used SB Hardware Midi synth port as Midi Out (PCI SB Audigy 2). Screen capturing on XP is rather demanding but as you can see/hear timing and precision were perfect:
https://youtu.be/enX5c3_yoWE
So something strange must be happening on your system.