VOGONS


Reply 41 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Kekkula wrote on 2023-01-27, 17:25:

Just a thought....
But have you actually tested that the simm is working.
If it randomly loads banks it could just be marginal module.

I'll try another module, but this one worked fine in Windows aside from the quirks with Dream's bank loading software... but that had problems with another stick of memory as well. The bank loader from EWS could load huge banks with no issues at all in Windows 98SE, and they sounded perfect.

Some other updates...
I have tried the SIMM jumpers in different configurations and it didn't seem to help any in DOS.

I also have tried some other sound banks (from the 94b pack), and they seem to all have similar glitches when loaded in DOS.

One thing I have noticed is that the time that 94DINIT.EXE takes to load the firmware and sound bank doesn't seem to change at all, no matter how big the bank is. In Windows, the larger banks take a couple seconds longer to load.

Is it possible that it's some kind of odd speed or timing issue? I have tried with a Pentium II 266... I can try going back to the PIII 850 to see if there's any change. Being that the card was designed and revised from 1995-2001 I doubt either of these CPUs would be a problem, but it's worth a shot.

Maybe a memory mapping issue? I don't even know if that is relevant in this situation... that's one area in which I have very little knowledge sadly.

Now for some blitting from the back buffer.

Reply 42 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Boy, this is confusing and not super useful for my situation... but I'm going to record all of this stuff so that no one else has to do all this trial and error. Well, and so that I don't have to figure all this out again years from now if I tinker with this thing in the future. 🤣

Okay, so, I was curious about using the S=ROM switch when running 94DINIT, which was mentioned in a couple different places. It seemingly "resets" the sound bank RAM by reloading the Dream chip's firmware and coffee it to just use a bank that is stored in ROM. This card doesn't have a bank ROM, where as the MaxiSound64 does, but it doesn't seem to matter as long as the 94DINIT program from that MaxiSound (~1996) is being used. I can run that program with that switch on my card and it seems to initialize the card with no bank loaded into RAM.

So I can do this:
94DINIT.EXE F=97PNP2 S=GMBK9708 P=330

...to load a bank. Then do this:
94DINIT.EXE F=97PNP2 S=ROM P=330

And the card will be re-initialized with no bank loaded. However! If I try this with the 94DINIT.EXE from 1998 it gives me an error that says that no ROM.94b was located... and nothing changes on the card at all.

So, apparently a lot changed with this program from 1996 to 1998. I would love to know if there are any newer or different versions out there.

I wonder if anyone involved with Dreamblaster would have access to anything like this?

(And for the record, resetting the bank RAM in DOS doesn't seem to change my MIDI issues.)

Also, trying to load the MaxiSound firmware (94PC4M1) using either version of 94DINIT gives me an error saying it can't initialize UART mode... so it looks like the 9707's firmware is definitely different than the 9407's.

Also tried swapping back to the PIII 850, took out some RAM, added some stuff to config.sys that was missing (files=30, buffers=30)... and the music still sounds weird with anything but a basic 1MB GM bank loaded in DOS. I will say, the glitched music does sound different some times. It is very strange.

Now for some blitting from the back buffer.

Reply 43 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

So, another interesting discovery that will at least make testing faster.

The Hoontech Sound Track 97 series of cards from Korea (mentioned by mattw earlier) are apparently very close to a stock SAM9407 configuration, so the DOS drivers seem quite flexible afterall. What is nice is that there is a program in that package called PNP97SET.EXE, and it is basically just an easy to use GUI for selecting RAM size (probably not applicable to this card since it is jumpered), Firmware file, Bank and IO Address which also automates the process of running 94DINIT (and the Hoontech mixer which seems to not work). Very cool!! I don't have to type these crazy file names any more if I want to keep testing things.

PNP97SET.EXE creates and modifies "STRACK97.INI" , which stores all the relevant settings that would normally be directly typed in as switches with 94DINIT.EXE
If you run PNP97SET.EXE without the .INI file, it will show the GUI and allow you to edit it. Once the .INI file exists, you have to launch PNP97SET.EXE /E to see the gui, otherwise it just runs the default configuration.

So, for the easiest config to use\test\configure this card in DOS, you'll need:
PNP97SET.EXE (Hoontech SoundTrack 97)
94DINIT.EXE (Hoontech SoundTrack 97)
Firmware File
Bank File

Interestingly, if I load the Hoontech Firmware HOON4D.BIN to this card it actually initializes properly... but it has the same odd playback issues with large banks. Despite the firmware files being much different in size and age, they both seem to work identically. So, this tells me it isn't likely to be a firmware issue necessarily.

Man, I'd love to get my hands on the latest DOS files direct from Dream. I'd be surprised if I didn't already have the latest firmware for the 97PNP2, but there is probably a slightly newer version of that 94DINIT out there somewhere.

I wish the download link on this page wasn't dead:
https://m.blog.naver.com/PostView.naver?isHtt … &logNo=55230060
https://blogattach.naver.net/dc49c073613f38e4 … dls-spc9999.zip

"hoonx44b.bin : New Firmware supporting GM+XG+DLS for 97PnP" 🤯 🤯 🤯 🤯 🤯

Anyone know any Korean sound card enthusiasts, or Korean FTP or file sites? 😮

Now for some blitting from the back buffer.

Reply 45 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
Nikita Lita wrote on 2023-01-28, 02:23:

uhm, I was able to download it fine. Here it is.

What!?

Wow, the link does not work for me. 🤯

Thank you!

EDIT: Downl0aded it and it looks like the file dates are from May of 1997... so it isn't newer, but I'll test it anyway.

EDIT2: No dice. Basically the same exact issues with either hoonx44 firmware. Also, it isn't related to the SIMM RAM because I can remove the SIMM and load the 1.7MB GM8DLS.94B bank from the Korean site above and the music still sounds really screwy. Though it's more recognizable than when loading the larger banks.

My next thing will probably just be to dump all of the available GM 94b banks I can find into the folder and test to see if anything other than the single 1MB GM bank works properly in DOS. Seems odd that it alone would be fine, but nothing else works... and yet in Windows all the banks sound totally normal.

Now for some blitting from the back buffer.

Reply 46 of 72, by darry

User metadata
Rank l33t++
Rank
l33t++
Ozzuneoj wrote on 2023-01-28, 02:38:
What!? […]
Show full quote
Nikita Lita wrote on 2023-01-28, 02:23:

uhm, I was able to download it fine. Here it is.

What!?

Wow, the link does not work for me. 🤯

Thank you!

EDIT: Downl0aded it and it looks like the file dates are from May of 1997... so it isn't newer, but I'll test it anyway.

EDIT2: No dice. Basically the same exact issues with either hoonx44 firmware. Also, it isn't related to the SIMM RAM because I can remove the SIMM and load the 1.7MB GM8DLS.94B bank from the Korean site above and the music still sounds really screwy. Though it's more recognizable than when loading the larger banks.

My next thing will probably just be to dump all of the available GM 94b banks I can find into the folder and test to see if anything other than the single 1MB GM bank works properly in DOS. Seems odd that it alone would be fine, but nothing else works... and yet in Windows all the banks sound totally normal.

a) Does the onboard RAM actually get disabled when a SIMM is installed (or is the onboard RAM always used in addition to any installed SIMM based RAM) ?
b) Could the onboard RAM be defective or have one or more of its higher address lines stuck or disconnected resulting in corruption when using bigger soundbanks ?

Reply 47 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
darry wrote on 2023-01-28, 05:55:
Ozzuneoj wrote on 2023-01-28, 02:38:
What!? […]
Show full quote
Nikita Lita wrote on 2023-01-28, 02:23:

uhm, I was able to download it fine. Here it is.

What!?

Wow, the link does not work for me. 🤯

Thank you!

EDIT: Downl0aded it and it looks like the file dates are from May of 1997... so it isn't newer, but I'll test it anyway.

EDIT2: No dice. Basically the same exact issues with either hoonx44 firmware. Also, it isn't related to the SIMM RAM because I can remove the SIMM and load the 1.7MB GM8DLS.94B bank from the Korean site above and the music still sounds really screwy. Though it's more recognizable than when loading the larger banks.

My next thing will probably just be to dump all of the available GM 94b banks I can find into the folder and test to see if anything other than the single 1MB GM bank works properly in DOS. Seems odd that it alone would be fine, but nothing else works... and yet in Windows all the banks sound totally normal.

a) Does the onboard RAM actually get disabled when a SIMM is installed (or is the onboard RAM always used in addition to any installed SIMM based RAM) ?
b) Could the onboard RAM be defective or have one or more of its higher address lines stuck or disconnected resulting in corruption when using bigger soundbanks ?

a) The tools in Windows show roughly 16MB of total memory, so it appears that the card is reporting to the drivers that it has disabled the onboard RAM. Whether the card is functioning properly and that is happening as it's supposed to... I can't say.

b) I was thinking that might be the case, but again it works fine when loaded with very large 10MB+ banks in Windows. They play perfectly fine, and as long as I use the less buggy tools from EWS, any bank I want to use will load without errors.

This is one that I think could probably be solved by someone with experience designing hardware like this (like the guys at Serdaco), but I just don't know enough about how all of this stuff interacts to understand what could cause the music playback to be so screwed up. The sound corruption appears to be limited either to specific instruments or to specific types of instruments. I played the Descent setup MIDI track with a large GS bank in DOS and the instruments were mostly static noises or off-key bleeping techno-insanity except for the drums which were extremely clear, realistic and loud. So it's like only the percussion loaded or was being read properly... and yet loading this bank in Windows works 100% perfectly.

I think it has something to do with the software being used to load the sound bank into the card, because the firmware version doesn't seem to make a difference in the sound as long as it's compatible (97PNP2 firmware from CD and any from Hoontech work the same in DOS). The method used by the Windows driver at startup, and used by EWS's BANK.EXE appears to be loading the sound bank properly every time... where as the 94DINIT program seems to be loading it improperly in a lot of cases, with the newer 1998 version doing slightly better than the 1996 version. The 1998 version is capable of loading a basic 1MB bank with no problems as well. 🤯

It looks like the source code in the SOFTWARE folder on the CD may be helpful in figuring out how the sound bank loading works, but I'm not a programmer, and without the source for 94DINIT I would never be able to guess what they are doing differently.

Now for some blitting from the back buffer.

Reply 48 of 72, by mattw

User metadata
Rank Oldbie
Rank
Oldbie
Ozzuneoj wrote on 2023-01-27, 18:32:

So, apparently a lot changed with this program from 1996 to 1998.

what did not change is that all those cards Hoontech, Terratec, Guillemot use SAM9407 and not SAM9707 chip. So, basically, what I am saying is that there is not a single known case of 94DINIT working with SAM9707, which essentially your tests confirm as well. the only other card, at least that I know, which uses SAM9707, is "Guillemot Maxi Studio ISIS (Interactive Sound Integration System)", but it's no longer ISA, but PCI card and that's why they put SAM9707 behind ISA-to-PCI bridge making DOS initialization very different - in fact, I believe, but I am not sure, maybe someone with such card can give us more details - SAM9707 is not supported by Guillemot on their Maxi ISIS card in DOS and they support just the ESS audio sound chip there (the other audio chip found on the card). They even go that far to define two modes - one they call "game mode" (only ESS audio chip is utilized) and the other Studio mode (SAM9707 is utilized and is Windows only). Even the open-source Linux driver for SAM9407 failed to make SAM9707 on "Guillemot Maxi Studio ISIS" work in Linux. So, I don't think 94DINIT is at fault, we're just rediscovering what was probably a known fact 20 years ago - that 94DINIT is not suitable for SAM9707.

Reply 49 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
mattw wrote on 2023-01-28, 08:56:
Ozzuneoj wrote on 2023-01-27, 18:32:

So, apparently a lot changed with this program from 1996 to 1998.

what did not change is that all those cards Hoontech, Terratec, Guillemot use SAM9407 and not SAM9707 chip. So, basically, what I am saying is that there is not a single known known case of 94DINIT working with SAM9707, which essentially your tests confirm as well. the only other card, as least that I know, which uses SAM9707, is "Guillemot Maxi Studio ISIS (Interactive Sound Integration System)", but it's no longer ISA, but PCI card and that's why they put SAM9707 behind ISA-to-PCI bridge making DOS initialization very different - in fact, I believe, but I am not sure, maybe someone with such card can give us more details - SAM9707 is not supported by Guillemot on their Maxi ISIS card in DOS and they support just the ESS audio sound chip there (the other audio chip found on the card). They even go that far to define two modes - one they call "game mode" (only ESS audio chip is utilized) and the other Studio mode (SAM9707 is utilized and is Windows only). Even the open-source Linux driver for SAM9407 failed to make SAM9707 on "Guillemot Maxi Studio ISIS" work in Linux. So, I don't think 94DINIT is at fault, we're just rediscovering what was probably a known fact 20 years ago - that 94DINIT is not suitable for SAM9707.

Right, it is very possible that there is just no way to make the SAM9707 do more than what we've gotten it to do so far in this thread, but I don't agree that there isn't a known case of it 94DINIT working in DOS with the SAM9707... in fact it works perfectly with a 1MB GM sound bank under DOS, as the testing here has shown today. The fact that simply using a different version of the 94DINIT program is what took it from not playing music at all to playing music perfectly with a 1MB bank (and playing music wrongly with other banks) still makes me think that a proper 94DINIT for this specific card could fix it. If that doesn't exist and no work-arounds can be figured out, then yes, this is as far as it's going to go.

DOS support in general is undocumented for this chip, as you said, and yet we've managed to get it to work perfectly with a 1MB bank in DOS. If I was content with it having that limitation, it would actually be completely usable right now in DOS as-is.

I wouldn't compare it much to the ISIS, being that that is a much later PCI card and connects through an ESS Maestro chip that certainly makes it harder to get working in DOS (versus something like an ESS Solo-1 or YMF-724 which can, in some cases, operate with minimal software like an ISA card). So far, everything on this card actually does initialize easily with the correct programs and utilizing no TSRs or other complicated stuff. So, basic DOS support is certainly there... it's just a matter of figuring out if it's possible to get around the current limitations. I will probably poke around in the Maxi Studio ISIS drivers now that you mention it though... just in case there is some latent DOS support or left over programs in there. 😀

The PCB on this card itself is labeled 94PNP2, so the design of the card is almost entirely the same as the previous model that did use the SAM9407.
The SAM9707 datasheet says:

Pin and function compatible with SAM9407, with additional features for professional
use :
— Up to 8 channels audio in
— Improved digital mix levels and digital overflow handling
— Improved tuning accuracy
— Additional DSP micro-instructions and data path for more efficient audio processing
algorithm coding

... so, Pin and function compatible seems like it should at least do what the 9407 could do. Which is why I felt it was worth tinkering with, and that has actually worked out really surprisingly well.

Anyway, I hope it's possible to get more out of it eventually, just because I like to experiment in an attempt to maximize the usefulness of stuff like this... but for now, I'm happy that we've at least managed to get it working in DOS, even if there is a limit on sound banks. It really is pretty cool to have been able to get this far at least, since it seemed impossible a few days ago. I am grateful for the help you've offered too. 😀

EDIT: oo... interesting. In the Maxi Studio ISIS drivers there is an ISISINIT.EXE program, and it appears to be a 16bit app. Real briefly I just opened the program up in a text editor to look for any plain text that might indicate whether it's a DOS or Windows application and I found a line of plain text that says it requires DOS 2.10 or later... sooo... the ISIS comes with an "INIT" program for DOS.

It may just be for the Maestro chip, and it probably won't work, but I will be trying this tomorrow. 😀

Now for some blitting from the back buffer.

Reply 50 of 72, by Nikita Lita

User metadata
Rank Newbie
Rank
Newbie
Ozzuneoj wrote on 2023-01-28, 02:38:

It looks like the source code in the SOFTWARE folder on the CD may be helpful in figuring out how the sound bank loading works, but I'm not a programmer, and without the source for 94DINIT I would never be able to guess what they are doing differently.

I was taking a look at that; unfortunately, it’s really just header files and thunked DLLs that hook into driver calls. Nothing really insightful.

I also took a look at decompiling the instrument creator and bank creator included on the CD, but I’m not seeing anything insightful on how the instrument/bank files are structured.

Reply 51 of 72, by mattw

User metadata
Rank Oldbie
Rank
Oldbie
Ozzuneoj wrote on 2023-01-28, 09:12:
... so, Pin and function compatible seems like it should at least do what the 9407 could do. Which is why I felt it was worth ti […]
Show full quote
Pin and function compatible with SAM9407, with additional features for professional
use :
— Up to 8 channels audio in
....

... so, Pin and function compatible seems like it should at least do what the 9407 could do. Which is why I felt it was worth tinkering with, and that has actually worked out really surprisingly well.

yeah, but doesn't mean 100% compatible on register-level, as well memory organization though. that's why maybe it works up to 1MB, but maybe it has some kind of interleaved memory and the next 1MB are not what are the next 1MB on 9407, who knows...

Nikita Lita wrote on 2023-01-28, 11:07:

I also took a look at decompiling the instrument creator and bank creator included on the CD, but I’m not seeing anything insightful on how the instrument/bank files are structured.

I am still hoping someone with understanding of 94B will appear, because that means we can make 94B bank with the Sounds from real SC-55 ROM, as I mentioned previously:

Re: Dream 97PNP2 - Interesting SAM9707 prototype\sample ISA MIDI card from 2001!

also, in that old thread about Roland SC, I did some work on 94B like understanding loop types, etc - you can read there. In any way, Roland sued and won over 4MB 94B bank and GS sounds in the 6MB ROM of SC-55 unpacked are something like 20MB. So, Dream even they cloned so well the sounding of SC-55, still seems are using inferior sound samples - maybe contrary to the community here, where we actually cracked the ROM encryption and have exact digital copy of those sounds, Dream captured the sounds in some other inferior way, hence the reduced size. again, who knows...

Reply 52 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

Small update. I tested ISISINIT.EXE, and it won't run because it doesn't detect an ISIS card... which is too bad, because the ISIS.BIN and the PCI64.BIN files in the same directory both work as proper firmware files for the SAM9707 when loaded with 94DINIT. So, it's possible that ISISINIT could work, but it is prevented from even trying. If there was a way to bypass this check, it would be worth trying, but I have no idea if it's possible.

I did locate a reference on one Korean site to a 97DINIT application for use with Hoontech Soundtrack 97/128 ISA cards, but it looks like it may be a typo? It says 97DINIT in one place, then 94DINIT in another place. The page seems to be about Linux drivers, but it goes on to use all the same filenames as the ones I've been working with in DOS, so I'm not sure what that's all about.

If anyone stumbles on any Hoontech drivers and locates a 94DINIT.EXE or 97DINIT.EXE anywhere, especially anything with a date newer than October 8th 1998, please upload it here along with any other drivers in the package.

Also, I managed to fix the issues with loading sound banks in Windows as well! The "94WBANK.EXE" program that was in the zip that Nikita Lita uploaded earlier appears to be an older but functionally identical version of the bank loader that came on the 97PNP2 CD - Except that it works without crashing Windows! This just makes me more interested in improving the functionality of the card by trying different programs, because this is a Windows application that was intended specifically for a Hoontech card with the SAM9407, and it works better with the SAM9707 than the newer (beta?) version of the program that came on the CD.

EDIT: Found another interesting page... and, big surprise, nearly every link is broken. I don't want to complain about archive.org, because it's wonderful thing... but it is so incredibly rare to have more than one or two links work on a page there. It is so frustrating. 🙁
https://web.archive.org/web/19990125092117/ht … 97.cyso.net:80/

On the menu to the left, the "download" page has links to lots of interesting stuff including tons of .94b sound banks. Probably all dead links, but may be useful for internet searches if any of them aren't available elsewhere.
https://web.archive.org/web/19980119043455fw_ … et/download.htm

There may be some interesting stuff here too:
https://web.archive.org/web/19980119043512fw_ … net/develop.htm

It seems that Hanmesoft was another name for Hoontech. hanmesoft.co.kr may provide some more leads, since all of the email addresses on that page are from there?

Now for some blitting from the back buffer.

Reply 53 of 72, by Warlord

User metadata
Rank l33t
Rank
l33t
mattw wrote on 2023-01-28, 15:10:

I am still hoping someone with understanding of 94B will appear, because that means we can make 94B bank with the Sounds from real SC-55 ROM, as I mentioned previously:

yea after skimming through the prosound.pdf in the SNDEDIT folder from the iso posted by Ozzuneoj. Think that it gives us all the information we need to make that a reality. The other files there are helpful too.

Theres a 10mb GM 94b floating around which should sound closer to the real real SC-55 ROM. I can't remember where I downloaded it but I think it came from retronn.de ftp server.

Reply 54 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

SUCCESS!!!!!!!!!

Turns out, I was on the right track. Finding the absolute latest 94DINIT.EXE I could find has FIXED the MIDI playback in DOS!! The Dream 97PNP2 can now load huge sound banks with zero issues in pure DOS with nothing but 94DINIT.EXE, 97PNP2.BIN firmware and a sound bank!

🥳🥳🥳🥳🥳

It's late. I will post more later, but the solution came when Googling the same stuff I'd searched for a dozen times before. When worded a little differently, I found a post right here on vogons about an "SC8500":
Re: SC8500 sound card - what does it do?
By chance, the maxi64dos.zip that elianda posted in that thread in 2009 contained a "94DINIT2.EXE" that was 36KB, versus the 70KB of the one from 1998 that I had been using previously (and 35KB for the first one). I don't know exactly what year it was made, what version it is or where these files came from, but the file is dated October 22nd of 2004. There was also a "GM94.BIN" firmware that was dated September 9, 1998, but I didn't bother using that since the proper 97PNP2 firmware seems to be working flawlessly.

The music seems a bit loud, so I may still be in a situation where the mixer volume needs to be adjusted somehow, but at this point I'm just so excited to have gotten this Windows-only developer sample pro-MIDI card from 2001 to work in pure DOS. I have noticed that loading sound banks definitely takes a couple seconds longer than it did with the incompatible 94DINIT versions, but it is still impressively quick. I think it just wasn't loading them in the correct place in memory before.

Anyway...

Woohoo! 😁

EDIT: Just to clarify why this is actually useful...

This card does not use any resources other than Port 330 for MIDI and IRQ 9, so you could use this in addition to another sound card as a dedicated MIDI card that works in pure DOS and loads sound banks easily. Think of it like a Creative AWE Upgrade card that works in pure DOS with any game that can send GM music to Port 330.

Now for some blitting from the back buffer.

Reply 56 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t
appiah4 wrote on 2023-01-29, 10:08:

Now someone needs to make a batch of these 🤣

Holy cow... this just got even weirder. I don't know what to think about this now. After using the 36KB 94DINIT2, if I switch back to the ~70kb Hoontech 94DINIT.EXE - IT NOW WORKS FINE! I can do a cold boot, and run either of those and it works fine!

The original ~35KB one from the 1996 MaxiSound drivers still gives no MIDI music at all in DOS... but now I can use either of the newer ones and everything works as expected in DOS. 🤯

Is there a good technical explanation for this? Is it possible for a bit to get stuck on\off or something in a firmware, and until the right thing pokes it, it doesn't work? But after that it is fine?? I'm actually kind of sad to say that I don't think I can get it to make the insane distorted music it was making before if it has basically fixed itself now after loading a proper firmware. Nothing else has changed in the system that should have any impact on this issue... other than the fact that I had tried loading those newer firmware files from the Guillemot ISIS earlier, but the music was still borked with those loaded. I also installed the basic maxisound driver package, which did nothing but add some files I already had to a folder and put them into autoexec. I removed all of that before I even tested the working 94DINIT2 though, so that shouldn't have had any effect at all.

I'm kind of at a loss for words, but still so happy that it's working. 😀

Any ideas as to what would cause this are welcome.

... anyway, today I will make a batch of the files that work, as well as the couple of extra files that are nice to have to simplify changing sound banks (in DOS and Windows). Though I'd love to be able to reproduce the crazy distorted stuff again just to deduce what actually fixed it.

Now for some blitting from the back buffer.

Reply 57 of 72, by Ozzuneoj

User metadata
Rank l33t
Rank
l33t

OKAY... I think I've got it.

If I use the buggy Windows-based bank editor that came on the CD and load a bank that makes the system crash, it apparently borks something on the card. If I reboot the system and load the 70KB Hoontech 94DINIT.EXE with a 4MB GM bank, music will be very screwed up (the 1996 MaxiSound one does not work at all). And it apparently will do that forever after that point? Because I had that card in and out of the system many times, it sat for at least a day, was rebooted, firmware flashed dozens of times, banks loaded dozens of times, went in and out of different slots... and the problem was still there.

... however, as soon as I run the file 94DINIT2.EXE (I have renamed it now for convenience) from elianda's 2009 post, telling it to load the same bank with the same firmware, it fixes everything. 🤯

I will experiment more with this later. What a mind-bending situation. 🤣

Now for some blitting from the back buffer.

Reply 58 of 72, by stacker139

User metadata
Rank Newbie
Rank
Newbie

Congratulations Ozzuneoj !
Could we achieve the same results with the SAM9407, Terratec and Maxisound cards, disabling the codec chips, Crystal and ESS ?

Project: SB-XXXL, Extensive Sound Cards System

Reply 59 of 72, by mattw

User metadata
Rank Oldbie
Rank
Oldbie

@Ozzuneoj

indeed, very interesting result, according to this:

Re: SC8500 sound card - what does it do?

"94DINIT2" is Internet find and the source is unknown. I was not able to find the same file in my archives either. So, I guess real technical explanation could be given only if someone can modify the Linux open source SAM9407 drivers:

https://sam9407.sourceforge.net/

to work with 97PNP2 card - that can answer what other 94DINIT versions are missing/doing wrong. In any way, great findings from your side all the way through - from the card itself to the software solution.

[EDIT] I found another version of "94DINIT", I believe it's still untested by you. it comes from Hoontech driver update. I am attaching it and so you can try it - that one is closer in size to the currently known working one.

[EDIT2] I think it's good occasion we make a list of all "94DINIT" that are available, because otherwise could be very confusing. I used HashCalc (https://www.slavasoft.com/hashcalc/) to generate CRC32 hash for each of those EXE files I have and so we can easily ID them, @Ozzuneoj please edit the list if you have more or if I made a mistake. So, here it's "94DINIT from":

Hoontech ST97 --> CRC32 e09c1a7b : NOK with 97PNP2
Guillemot MaxiSound 64 Home Studio (SC8500) --> CRC32 e7a0a6dc: NOK with 97PNP2
Hoontech Driver Update (attached to this post) --> CRC32 19899967: not tested yet with 97PNP2
Unknown source version --> CRC32 0c157c13: OK with 97PNP2

is that the list or there are even more?

Attachments

  • Filename
    94DINIT_64ddma.rar
    File size
    16.5 KiB
    Downloads
    50 downloads
    File license
    Public domain