VOGONS


Munt Reloaded - Development

Topic actions

Reply 181 of 964, by collector

User metadata
Rank l33t
Rank
l33t

Sure, but I need to iron out a few things first. I tested it on XP and worked as expected, but it is proving to be a little inconsistent on Win7 64-bit. It seems to be a system32/SysWOW64 issue. I also want to make sure that it uninstalls cleanly. Once I get that sorted out and get it cleaned up/ commented, etc., I'll get it to you.

The Sierra Help Pages -- New Sierra Game Installers -- Sierra Game Patches -- New Non-Sierra Game Installers

Reply 183 of 964, by Jiri

User metadata
Rank Member
Rank
Member
wd wrote:

Why are you using nsis and not an .inf file?

I think that NSIS is good open source system to create installers and very suitable for this task. Some less experienced users could have problems to use .inf file while install/unistall with NSIS installer can easily handle anybody.

Reply 185 of 964, by collector

User metadata
Rank l33t
Rank
l33t
wd wrote:

Why are you using nsis and not an .inf file?

In part because they couldn't get the inf install to work on 64-bit Windows.

@ KingGuppy, Check your PM. I have commented it where it might not be so clear. Most of the required includes should come with the default install of NSIS, except for possible "Registry.nsh", which can be found on the Wiki. I also have in the script to install Putzlowitschs Vista-MIDIMapper, but it is commented out as I do not believe it is GPL.

If you have any questions or need any help with it, just ask. If I am not around here, I can be found on my message board on the Sierra Help Pages.

Reply 186 of 964, by OmerMor

User metadata
Rank Newbie
Rank
Newbie

I noticed from your commits that you use digital samples for tuning the emulation code.
Can you give a rough estimate on how much more tuning is needed to achieve 100% accuracy?
Also, KingGuppy said that he's trying to get the MT-32 chips scanned for perfect reverse engineering. Where does this effort stand right now?

Reply 187 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

I can't give you a percentage or accuracy estimation 'cause I ain't interesting in such statistics 😀 But I won't be too far from trues saying that:
- PCM partials ~100% accurate
- synth partials generated ~100% accurate
- TVA, TVP, TVF are very close
- MIDI message and SysEx processing seem to be accurate including emulation of model-specific behavior but not all the bugs
- reverb mode 3 must be accurate but isn't thoroughly tested for now
- other reverb modes are too far from accurate 😀

Reply 188 of 964, by wd

User metadata
Rank DOSBox Author
Rank
DOSBox Author

In part because they couldn't get the inf install to work on 64-bit Windows.

I've played with the ZMBV codec installer back then, but that's aiming at installing
the 32bit files on a 64bit system (WoW64 stuff) so that may not be of interest to you.

If the nsis solution is quite straightforward that may be nice to consider for zmbv as well.

Reply 189 of 964, by KingGuppy

User metadata
Rank Member
Rank
Member
OmerMor wrote:

I noticed from your commits that you use digital samples for tuning the emulation code.
Can you give a rough estimate on how much more tuning is needed to achieve 100% accuracy?

As SergM said, almost everything except reverb is now close to 100% accurate, to the extent that you can compare real and emulated output sample-by-sample and see differences of 0 much of the time, and +/- 2 or so in the worst case for basic partials (though I may have to tweak the TVA multiplier a little more for synth partials). Note that the real device won't necessarily produce the exact same output each time for the same parameters, largely due to asynchronous communication between the microcontroller and the LA32.

We have many GB of digital captures (though we'll need many more) and plan to go through and automatically compare them in order to find anything that stands out as inaccurate in the emulation. We'll probably publish some statistics at that point (I *do* like useless stats 😀 ). I'm also working on some documentation of how the MT-32 works internally and how accurately we model each part.

SergM is the "reverb guy" right now, and I'll be providing him with more digital captures this weekend and helping out in any way I can. I think that the emulator is going to sound indistinguishable from the real thing in casual listening very, very soon.

OmerMore wrote:

Also, KingGuppy said that he's trying to get the MT-32 chips scanned for perfect reverse engineering. Where does this effort stand right now?

The die scan guys have had an MT-32 since July 2010. Last I heard there were some problems with the decapping, but they'll try again. Obviously, we're not holding off on development for this...

Reply 190 of 964, by collector

User metadata
Rank l33t
Rank
l33t
wd wrote:
I've played with the ZMBV codec installer back then, but that's aiming at installing the 32bit files on a 64bit system (WoW64 st […]
Show full quote

In part because they couldn't get the inf install to work on 64-bit Windows.

I've played with the ZMBV codec installer back then, but that's aiming at installing
the 32bit files on a 64bit system (WoW64 stuff) so that may not be of interest to you.

If the nsis solution is quite straightforward that may be nice to consider for zmbv as well.

Even on 32-bit XP you had to go through the add hardware wizard. On Win7 the wizard exists, but is no longer accessible via the control panel. You have to run the exe directly. Try telling noobs how to start the hardware wizards when the Runbox is by default hidden and the Windows directory's contents are also hidden.

The codec installs just on a right click, which is not that difficult for beginners, but still may be nice to add as an option to the DOSBox installer. I have not looked at the ZMBV INF to see what is needed, but I assume that zmbv.dll gets copied into the System32 directory, gets registered and possibly has a registry entry made. All can be easily added as a selectable option to the DOSBox installer.

Reply 192 of 964, by KingGuppy

User metadata
Rank Member
Rank
Member
Varka wrote:

Just to be clear, do you mean only the mt32emu.dll MIDI driver will be abandoned soon?

Will the mt32emu.exe still be supported?

I'm not even 100% sure what this is - SergM's Win32 app? - it's never been officially supported.

There will be a new application that will run the emulator with an optional GUI and expose the emulated device(s) via RTP-MIDI. You can then use an RTP-MIDI driver on your OS to get native access for applications. We'll do a full announcement and documentation when it's ready.

Reply 194 of 964, by ponytear

User metadata
Rank Newbie
Rank
Newbie

Sergm: if I may, I have a (small) request regarding the win32app. Would it be possible to display in the app window the LCD text that's supposed to be displayed on the MT-32 screen? That would be cool! 😀 Thanks!

Reply 195 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

Actually, I have no plans of win32app further development. It's for mt32lib tests only. For production use, a cross-platform application will be developed with all necessary interface features you ever imagine (as KingGuppy once told me). So, very soon it will be available I suppose 😀

Reply 197 of 964, by sergm

User metadata
Rank Oldbie
Rank
Oldbie

Basically, the project is a cross-platform library. It includes Win and ALSA drivers for now, but we want to extend support for other systems. To achieve this goal, we're moving to Qt (user interface) and RTP-MIDI.

Reply 198 of 964, by collector

User metadata
Rank l33t
Rank
l33t

The driver can be made to work on x64, but not by the INF install. I wrote an NSIS script that will install the driver on x64 as well as Win32. KingGuppy has the NSIS script, but it is up to him as to what to do with it. I don't know if he has had a chance to look at it yet or not.

Reply 199 of 964, by Mok

User metadata
Rank Newbie
Rank
Newbie
sergm wrote:

- reverb mode 3 must be accurate but isn't thoroughly tested for now

It requires some tweaking. Try checking some games that heavily use mode 3 reverb, for example in The Humans level 1 music, echo is hardly audible compared to previous version (and LAPC recordings from http://cosmic-dreams.net).