keropi wrote:

I have uploaded the mp3 on tinyupload: http://s000.tinyupload.com/index.php?file_id= … 248816631417444

I haven't noticed malware on mediafire or had any complaints before - then again I do use an adblocker on chrome so even if there was a pop-up it did not open.
Also I don't have an old-style Roland mpu, all I have atm is 401AT/SCC-1/LAPC-I

I hadn't noticed any malware on mediafire before, either, which is why I was willing to click on the download link. I also have Adblock Plus installed on Chrome, and I did it from Chrome...so...yeah...

And it wasn't a pop-up. It was the actual download link. It downloaded the mp3 and opened the scam tab simultaneously when I clicked on it. Makes me wonder if mediafire got hacked.

Since you're using the new(er) LA Roland devices, I would imagine that's why it worked so well for you. The gen 2 MT-32s onward got rid of the buffer overflow bug.

oh! now I see what you mean - I will try mediafire again

just to clarify , I am using the old-version MT-32 and a MQ mpu - the roland mpus are in storage and I have not tested with any of them

edit: tried the mediafire download, did not get anything other than the file... maybe it was hacked or an ad was malware and they disabled it who knows...

Any luck?

keropi wrote:

oh! now I see what you mean - I will try mediafire again

just to clarify , I am using the old-version MT-32 and a MQ mpu - the roland mpus are in storage and I have not tested with any of them

edit: tried the mediafire download, did not get anything other than the file... maybe it was hacked or an ad was malware and they disabled it who knows...

When I got home yesterday, I went back to that .mp3 download page and tested it a few times. It appears to have been a malicious ad, and I have not been able to get it to come up again. The spammy tabs that open when clicking on those download links are clearly on some kind of rotator or random(?) selector, so the malicious tabs I encountered might still open for other people.

It makes me severely distrust mediafire.com for allowing advertisers to put hooks into their download links, especially ones that can evade ad blockers. God, I hate Internet-based advertising. At least with TV, radio, and billboard ads your computer will never be infected or the security team at your job sicced on you. =)

(I'm not blaming you -- AT ALL -- for what happened, btw.)

tpowell.ca wrote:

Any luck?

Haven't tried, yet. My family and extended family showed up at my house yesterday right after I got home from work, which occupied my time until I was too tired to do anything useful. I'm definitely going to try your hack when I get home tonight -- in fact, I'm quite excited to try it out.

If it works, you will have come up with something brilliant that many other people in my situation could find very useful, and it would need to be spread out there for the fans of Prince of Persia and probably other games.

Even if it doesn't work, it's a damned good idea. =)

aquishix wrote:

Haven't tried, yet. My family and extended family showed up at my house yesterday right after I got home from work, which occup […]
Show full quote

tpowell.ca wrote:

Any luck?

Haven't tried, yet. My family and extended family showed up at my house yesterday right after I got home from work, which occupied my time until I was too tired to do anything useful. I'm definitely going to try your hack when I get home tonight -- in fact, I'm quite excited to try it out.

If it works, you will have come up with something brilliant that many other people in my situation could find very useful, and it would need to be spread out there for the fans of Prince of Persia and probably other games.

Even if it doesn't work, it's a damned good idea. =)

Yeah, it worked great in my case for those problematic games like Wing Commander 1+2 where Mt-32 is supremely sensitive to the speed at which those games send SysEx data.

tpowell.ca wrote:

Any luck?

Unfortunately, no.

I tested it by using the GUS's PLAYMIDI.EXE player with the -mpu option. Messages were clearly making it to the MT-32 because I saw the green LED on the front panel of my MT-32 light up and it took a while to finish playing the file...but then when I went to play Prince of Persia, it sounded exactly as it does when I unplug or turn off the MT-32 during the initial loading. I.e., no instrument changes, but no garbage and lock up either.

From what you said in your other post, your technique captures the instrument reprogramming MIDI messages. Therefore, I don't know what to make of this.

keropi wrote:

just to clarify , I am using the old-version MT-32 and a MQ mpu - the roland mpus are in storage and I have not tested with any of them

This is making me think that the MQ mpu card is actually different in implementation from the MIF-IPC-B + MPU-401 chain. I don't know how else to explain that your old version MT-32 works perfectly on a 386SX33 with cache enabled. I'm using a 386DX40 with cache *disabled*, and holding in the turbo button, and it's still screwing up. The only significant difference in our setups is the interface to the MT-32. =(

I even tried running the game from a floppy disk drive to see if that would make any difference; it didn't.

just to make sure - there is no way of a port conflict right? have you tried with no soundcard installed and only the mpu present?

keropi wrote:

just to make sure - there is no way of a port conflict right? have you tried with no soundcard installed and only the mpu present?

I'm (100-epsilon)% sure I did eliminate that possibility, but I will make 100% sure when I return from my vacation (~1 week from now).

That doesn't seem like a possible explanation anyway, because literally every other game I've tried works just fine on all 3 of these systems including SMI, KQ4, SQ3, ROTD, etc.

aquishix wrote:

Unfortunately, no. […]
Show full quote

tpowell.ca wrote:

Any luck?

Unfortunately, no.

I tested it by using the GUS's PLAYMIDI.EXE player with the -mpu option. Messages were clearly making it to the MT-32 because I saw the green LED on the front panel of my MT-32 light up and it took a while to finish playing the file...but then when I went to play Prince of Persia, it sounded exactly as it does when I unplug or turn off the MT-32 during the initial loading. I.e., no instrument changes, but no garbage and lock up either.

From what you said in your other post, your technique captures the instrument reprogramming MIDI messages. Therefore, I don't know what to make of this.

Is it possible that the GUS midi player doesn't support SysEx data?
Try using a proper midi file player like this one: http://www.midimusicadventures.com/qs/midi-zi … ies/playmid.zip

And use something other than the GUS card. Try your MPU-401 cards or soundblaster (16+) if possible.