What's new
Thank YOU! the game sounds ok now with your fixed vroma0 on the multiMVS.
I think maybe it's a firmware issue, because the same rom without your fixed vroma0 works and sounds ok when i play it on the mister fpga Neo Geo core.
I'd say it's rather a problem when merging the files to create a romset.

Was working before well on mister or did it also have bad sound?
 
It’s not advisable to use the Neo Geo MiSTer core to test or report bugs against the cart as there’s issues with accuracy for sure, especially sound.
 
It’s not advisable to use the Neo Geo MiSTer core to test or report bugs against the cart as there’s issues with accuracy for sure, especially sound.
Yeah i know but i just wanna mention it because mister use the same multimvs romset.

@Darksoft the rom was working ok on the mister exept the sound pitch issue who was fixed in the last core update.
 
IMO, you shouldn't really report bugs using a non-100% core as a sample against real hardware.

Bugs reported should only be tested against original carts IMO. I love the MiSTer, but this is just opening up a can of worms.
 
Can you tell me who did the fix and what was exactly changed?

I followed the directions that Mitsu posted about concatenating the sound roms to form vroma0, but the example he showed only had adpcma roms so it wasn't clear how to handle the adpcmb roms. I inspected the existing rom and noticed that there was 1Mb of zero-padded data between the concatenated "a" roms and the "b" rom. I started to inspect some of the other games that had "a" and "b" sound roms, and noticed that sometimes the filesize was bigger than the total of straight-concatenation, so there must be some padding going on. Then I noticed Cyber Lip looked like it was straight concatenation of a+b, so I tried that with Ghost Pilots. The sounds were off, but the melody was more complete. It seemed like it was accessing all of the sound data, but the wrong data. I recognized at that point that it was likely that the sound driver was accessing the zero-pad region for part of the music where we noticed the missing instruments in the original rom. So I tried it with 512 zero padding before the "b" rom and it worked!

The structure of the old (non-working) vroma0 stacked top to bottom:
v11 (1Mb, pcma)
v12 (512b, pcma)
zero pad (1M)
v21 (512b, pcmb)
---------------------
total size 3Mb

The structure of the new (working) vroma0:
v11 (1Mb, pcma)
v12 (512b, pcma)
zero pad (512b)
v21 (512b, pcmb)
---------------------
total size 2.5Mb
 
I tried out SamShow III, and experienced the reboot after two matches. Original bios and latest firmware/menu.
I tried samurai showdown III AES or MVS rom.
I have some random freeze too. I use the latest firmware and menu and MV-1FZ motherboard and unibios 3.3.
 

Attachments

  • 20210924_181710.jpg
    20210924_181710.jpg
    144.3 KB · Views: 130
Yeah i know but i just wanna mention it because mister use the same multimvs romset.

@Darksoft the rom was working ok on the mister exept the sound pitch issue who was fixed in the last core update.
Then it was an error when making the romset. No need for a new firmware.
 
I followed the directions that Mitsu posted about concatenating the sound roms to form vroma0, but the example he showed only had adpcma roms so it wasn't clear how to handle the adpcmb roms. I inspected the existing rom and noticed that there was 1Mb of zero-padded data between the concatenated "a" roms and the "b" rom. I started to inspect some of the other games that had "a" and "b" sound roms, and noticed that sometimes the filesize was bigger than the total of straight-concatenation, so there must be some padding going on. Then I noticed Cyber Lip looked like it was straight concatenation of a+b, so I tried that with Ghost Pilots. The sounds were off, but the melody was more complete. It seemed like it was accessing all of the sound data, but the wrong data. I recognized at that point that it was likely that the sound driver was accessing the zero-pad region for part of the music where we noticed the missing instruments in the original rom. So I tried it with 512 zero padding before the "b" rom and it worked!

The structure of the old (non-working) vroma0 stacked top to bottom:
v11 (1Mb, pcma)
v12 (512b, pcma)
zero pad (1M)
v21 (512b, pcmb)
---------------------
total size 3Mb

The structure of the new (working) vroma0:
v11 (1Mb, pcma)
v12 (512b, pcma)
zero pad (512b)
v21 (512b, pcmb)
---------------------
total size 2.5Mb
Hats off. You saved me a lot of time ;)
 
Now the question is....which other games have that same A+B configuration, because games are usually converted using a script...
 
I tried samurai showdown III AES or MVS rom.
I have some random freeze too. I use the latest firmware and menu and MV-1FZ motherboard and unibios 3.3.
I roll back to previous firmware/Top/CPLD and no random freeze appears on samurai showdown 3. But still the same sounds issues on KOF 2003.
 
Now the question is....which other games have that same A+B configuration, because games are usually converted using a script...
Super Spy seems to have the same arrangement (1M "A" + 512 "A" rom + 512 "B" rom).

Edit: And missing sounds!! I checked against mame and with the unibios jukebox 0x0725 it's clear.
 
Last edited:
I tested my KOF2003 and I have good sound.
Thank you for the test. Other people who tested have good sound too 😭😭 but I read on another french forum about neogeo that some people have same sounds issues with original playmore cart like kof 2003 and samurai showdown 5 and some motherboard. One people test 3 original samurai showdown 5 on his motherboard but any of them worked on his motherboard, only a bootleg work. He advised to play on PCB KOF 2003 or with a bootleg one.
 
Last edited:
That would be great! I compared against mame and it sounds good to me, but an original cart is even better. The music in Super Spy is actually pretty cool.
 
Back
Top