RetroGamer4Ever wrote on 2023-07-25, 13:09:

I was using the XG VST and 32-bit FSMP (6.1.1) wouldn't play the files it flagged as 32-channel after trying to load them, simply moving to the next 16-channel MIDI file. I'm testing it again right now and it seems to have been a hiccup, cause none of them would play when I was using the XG VST earlier and they work fine now.

As I said with the above linked FSMP (version 6.2) none of these Midi files are detected as 32-channel anymore (since the tracks that refer to port B actually do not contain any note events). But I re-tested with 6.1.1 and could not reproduce your problem on 6.1.1 either.

Maybe not an important news for everyone, but VST MIDI driver now also works on Win NT 4 SP 6 and Windows 2000.
FSMP works with reduced functionality on Win NT 4 because of the lack of Directsound 8 (so there is no BassMidi/BassVSTi) thus working VST MIDI driver is a nice addition to Win NT 4. Although 32-bit float formats and 4-channel mode are not working, basic 16-bit stereo WaveOut works nicely .
BTW, both S-YXG50 and VSC plugins work perfectly on NT 4.

Version 1.6 released:
https://github.com/Falcosoft/vstdriver/releases

1. Added Windows NT 4 SP6 and Windows 2000 support. On Windows NT 4 32-bit float formats and 4-channel mode are not working in case of WaveOut mode but basic 16-bit stereo WaveOut works nicely.

2. Added sample format selection. The default stayed 32-bit float but 16-bit int can also be selected. On older Windows versions 16-bit int can be more compatible but even on modern Windows versions (Vista+) 16-bit int option can be somewhat still relevant. It's because in case of 32-bit float output modern Windows versions apply a limiter/compressor automatically which on the one hand prevents clipping but on the other hand can result in wavy volume response and makes it harder to calibrate proper synth output level.

3. Fixed multiple error dialogs when in Waveout mode a device is explicitly selected (instead of Microsoft Sound Mapper) and device options are changed in audio control panel of Windows during playback. Nevertheless selecting Microsoft Sound Mapper is the preferred way to use the driver since if it's selected the driver automatically can follow the selected default output device even during playback and no error occurs even if the device options are changed during playback.

4. Some other minor fixes and enhancements.

Works with Windows NT4 SP6/2000/XP/Vista/7/8/10/11 (Both Waveout and ASIO output)

PS: Back to FSMP 😀

Dang, Win 10 won't let me install the latest version since it complains about (erroneously) detecting a virus.

Trelokk wrote on 2023-07-29, 07:07:

Dang, Win 10 won't let me install the latest version since it complains about (erroneously) detecting a virus.

Are you sure that Windows itself (Windows Defender) is the one that prevents you from installing this version? It was interesting that the last time vsthost32.exe was missing on your system. It's likely that some anti-virus stuff removed it after installation.
I have tried to install this version on both Windows 10 and 11 with updated Windows Defender definitions and none of them prevented me installing it. The biggest problem was downloading it, since all browsers complain that this file is dangerous since not many downloaded this file before (catch-22).

So please check what anti-virus stuff is on your system and report the file as false positive.

It's just flagged as "weird and unusual" by web browser security AI and because it is an executable and has no software information attached, the OS security flags it as potentially harmful.

Unfortunately many Virustotal engines flagged the installer as malicious:
https://www.virustotal.com/gui/file/6bf205ad5 … 0b171f49117f625

Interestingly Version 1.5 has even more false positive detection... I hate this and I repeteadly have to face this problem with FSMP also.
I have already reported the false positive issue to vendors that have online file submission forms but it is a tedious work and I'm not sure it's worth the hassle.
I do not think I can do more than this. If you have the mood you can also report false positive issue if your engine detect the installer/driver as malicious.

If you do not trust me you can compile the driver for yourself from the repository.

PS: I have checked and it seems it's not my special build process that causes the problem. The last Aracula verison of the driver installer has the same number of false positive detection:
https://www.virustotal.com/gui/file/16db83ca7 … 2ec50d0db7be7b7

Well, the root problem is that you don't have a software publishing certificate, which is basically a seal of approval/security recognized by Microsoft Windows and most security software, but that's costly and burdensome for you to deal with, as an individual developer. Dege has the same issue with his dgVoodoo releases and because they are apparently found all over malware-infested third-party sites, they are automatically flagged as security threats by AV software, regardless of whatever dialogue he opens with the security software makers.

RetroGamer4Ever wrote on 2023-07-29, 11:46:

Well, the root problem is that you don't have a software publishing certificate, which is basically a seal of approval/security recognized by Microsoft Windows and most security software, but that's costly and burdensome for you to deal with, as an individual developer. Dege has the same issue with his dgVoodoo releases and because they are apparently found all over malware-infested third-party sites, they are automatically flagged as security threats by AV software, regardless of whatever dialogue he opens with the security software makers.

I succeeded in getting 0 false positive with my last FSMP package despite the fact that none of the executable files are signed:
https://www.virustotal.com/gui/file/572386b77 … 778b1d898b1a126
The new test version is not so lucky, it has 4 false positive detections currently:
https://www.virustotal.com/gui/file/796196068 … 6466f3a6be99be7

Considering the above results I'm still somewhat suprised by the 14 false positive detections of VST Midi driver. Yes, signing could help but it seems it's not the key factor. I think the NSIS installer itself plays a big part in the problem. Arakula's version and my current version differ a lot now, yet we got almost the same false positive detections. I think it's because both versions use the same NSIS installer.
I would bet that "Gen:Variant.Adware.Nemesis.140" is an adware that uses NSIS installer 😀 .

PS: It seems we are not alone:
https://nsis.sourceforge.io/NSIS_False_Positives

For the record: I'm not using any external AV, only Defender.

Maybe there's a different installer you could try, even though none comes to mind. Inno or something. Would have to be freeware and lightweight, I guess. NSIS doesn't cut deeply into the registry, which is something I've always appreciated.

Trelokk wrote on 2023-07-29, 20:04:

For the record: I'm not using any external AV, only Defender.

Maybe there's a different installer you could try, even though none comes to mind. Inno or something. Would have to be freeware and lightweight, I guess. NSIS doesn't cut deeply into the registry, which is something I've always appreciated.

FSMP does not use any installers yet it is regularly flagged by various AV engines. So there is no guarantee. Even if the installer framework was changed there is a high chance that the problem would remain( maybe a little fewer false detection, but it would not be a game changer). I do not think it is worth the effort.
BTW, you should try again in a few days. I have reported the false positive detections to many vendors. Let us see if it has any effects.

Would you share a little more detail about what Windows did or told you exactly?
Maybe a screenshot would be the best.

@Edit:
Now, we are under 10 😀 9 remained:
https://www.virustotal.com/gui/file/6bf205ad5 … 7f625?nocache=1

Seems your report really helped. I tried to install the driver again this morning, with latest Defender updates. Even though there was a Smartscreen warning again (= unknown file), this time it continued uninstalling the previous version and I was able to finish the entire process.

Let's hope this was the final word regarding the matter.

Trelokk wrote on 2023-07-31, 05:23:

Seems your report really helped. I tried to install the driver again this morning, with latest Defender updates. Even though there was a Smartscreen warning again (= unknown file), this time it continued uninstalling the previous version and I was able to finish the entire process.

Let's hope this was the final word regarding the matter.

I'm glad it worked. Now the false positive detections are reduced to 6 on Virustotal. Unfortunately I do not think we will ever reach 0.
Some AV vendors have no submission forms but expect email reports at best case and then no response at all from them (not to mention how hard is to send even password protected zipped exe files in email nowadays). Some AV vendors business practices are rather questionable so personally I do not expect more from them. E.g. here is an interesting thread on Malwarebytes forum where MaxSecure's represenatives ask Malwarebytes to remove MaxSecure files from their malware list 😀
https://forums.malwarebytes.com/topic/274793- … -on-virustotal/

Honestly I do not agree with Virustotal's policy towarding new AV vendors. I would only pick up a new vendor if it has good reputation and forms to report problems.
BTW, expecting that new versions will have no such problems is not realistic.

If it takes a few days or weeks until this gets whitelisted in common AVs, so be it. For me, latest version is already worlds better than I ever thought it could be, so I would survive some waiting time till I can update.

Hi,
version 1.7 of VST MIDI driver is released.
https://github.com/Falcosoft/vstdriver/releases/tag/v1.7.0

The situation seems to be better than last time regarding false positives (only 5 false detections without reporting anything to AV vendors so far)
https://www.virustotal.com/gui/file/ea82d577f … 93678834ccbb467

Changes:

1. Added Always on Top option to port A/B editor dialogs.
Sometimes editors opened up behind the client that opened the port. This cannot happen anymore. If always on top status is disturbing for some reason you can switch it off from the new "status bar" area located right at the bottom of the editor dialog.

2. Added Save Settings option to VST editor dialog.
In case of more complex plugins it was hard to tune settings without hearing the results. Now you can fine tune and save the settings while the plugin is producing audio.
Be aware that it is possible that you cannot save plugin settings from the editor at runtime while you can from the configuration dialog.
It's because the configuration dialog runs with elevated administrator rights while the driver does not. So make sure to give write permission for normal 'Users' to the plugin's folder ( this usually has to be done for plugins located in Program Files (x86) or Program Files).

3. Fixed DPI adjustment for NT4/W2K/XP that do not use DPI virtualization.

4. High DPI related optimizations for modern Windows.
Added manifests for High DPI support. Since VST2 editors mostly do not handle high DPI at all the default is not aware + GDI scaling in Win10 1809+. The default DPI mode is rather compatible but DPI mode can be changed through registry (HighDpiMode key).

HighDpiMode: -5 (0xFFFFFFFB) = Windows enhanced GDI scaling (Win10 1809+) . This is the default.
HighDpiMode: -1 (0xFFFFFFFF) = Windows normal scaling. This is the most compatible but results in blurry texts. This is the default before Win10 1809 (Vista/7/8/10).
HighDpiMode: -2 (0xFFFFFFFE) = Plugin scaling. This is the less compatible option. Since VST2 editors usually cannot do scaling in 90% of cases results in tiny editors.

Optimizations are mostly visible on fonts/texts and common controls (they are sharper). Unfortunately bitmaps cannot be scaled without blur.

4K 200% before:

4K 200% after:

5. Other minor fixes and optimizations.

1.7 was flagged by my AV as suspicious, so I had to disable the AV, to install it. That's the first time that has happened to me, as the previous versions were only flagged by Windows security due to being unknown software. I'm testing 1.7 right now with the latest FSMP and both seem good to go together, in every way.

RetroGamer4Ever wrote on 2023-08-11, 15:03:

1.7 was flagged by my AV as suspicious, so I had to disable the AV, to install it. That's the first time that has happened to me, as the previous versions were only flagged by Windows security due to being unknown software. I'm testing 1.7 right now with the latest FSMP and both seem good to go together, in every way.

What AV do you use exactly? Would you report them the detection as false positive, please?

Gratefully, latest version installs smoothly without any AV complaints this time (for me). Let's try to keep it that way. ^^

Falcosoft wrote on 2023-08-11, 18:01:

RetroGamer4Ever wrote on 2023-08-11, 15:03:

1.7 was flagged by my AV as suspicious, so I had to disable the AV, to install it. That's the first time that has happened to me, as the previous versions were only flagged by Windows security due to being unknown software. I'm testing 1.7 right now with the latest FSMP and both seem good to go together, in every way.

What AV do you use exactly? Would you report them the detection as false positive, please?

I use McAfee and I'll have to jump through some hoops to submit the VST Driver installer. They require me to extract the contents of the installer executable and include it in a compressed archive with the executable itself, along with a report about what's what. I'll get around to it later this week, but I have a lot of go-here-and-do-this stuff to deal with this week, so it will probably have to wait until the weekend comes around.

RetroGamer4Ever wrote on 2023-08-15, 10:35:

Falcosoft wrote on 2023-08-11, 18:01:

RetroGamer4Ever wrote on 2023-08-11, 15:03:

1.7 was flagged by my AV as suspicious, so I had to disable the AV, to install it. That's the first time that has happened to me, as the previous versions were only flagged by Windows security due to being unknown software. I'm testing 1.7 right now with the latest FSMP and both seem good to go together, in every way.

What AV do you use exactly? Would you report them the detection as false positive, please?

I use McAfee and I'll have to jump through some hoops to submit the VST Driver installer. They require me to extract the contents of the installer executable and include it in a compressed archive with the executable itself, along with a report about what's what. I'll get around to it later this week, but I have a lot of go-here-and-do-this stuff to deal with this week, so it will probably have to wait until the weekend comes around.

Hi,
Thanks, but if it's so complicated then I do not think it's worth the hassle. I'm in the middle of an experiment with self-signed cerificates since it seems it worked for Coolsoft (Midi Mapper, VMS):
https://coolsoft.altervista.org/en/code-signing

I also switched tactic so I have already uploaded the new installer version to VirusTotal and started reporting false positives to vendors without uploading the new release/installer executable to GitHub.
So I will release 1.7.1 soon with not so many new features but hopefully less false positive detections.