Genesis / Megadrive core ported to MiST

[Gehstock]

yes is supported, https://github.com/phoboz/fpgagen/tree/ ... Configs/TV Copy to SDCARD Root

[farvardin]

thanks a lot, I've tried a few games, and it's very promising!

[jotego]

We are still working on this and with many problems to solve, but I'd like to share how it sounds as of today. I got JT12 almost done, that covers the YM2612/YM3438 part. I got PSG sound from the Sega Master System core today. Phoboz and robinsonb5 are working on the video and system aspects, particularly on the flickering problem.

As I said, still issues but have a look:

EDIT: I see that my work versions of the .rbf files got leaked!
I have deleted those trial versions and uploaded the current one here. Please report sound issues to the github issue page or to this thread.

[crashman]

Increible.

Thanks for your work!

[NML32]

Thanks to everyone that's involved in this project. Great Job!!

[DrOG]

@ Phoboz, robinsonb5 & jotego (and everyone else involved):

!!!THIS IS AWESOME!!!
(Sorry for shouting.)

Thanks for your hard work, fantastic progress!

[DanyPPC]

Good work !

[Newsdee]

It's great to see a collaborative effort like this. Keep up the excellent work guys!

[ericgus]

Nice will have to try it out.. is it just me or is it one of your bugs that needs fixing but the sonic ring "ding" seems a bit.. different in your video..

[phoboz]

Nice will have to try it out.. is it just me or is it one of your bugs that needs fixing but the sonic ring "ding" seems a bit.. different in your video..

The LFO is still missing in the sound chip. So whenever that is used to modulate the signal, there will be a differance to the sound, once the LFO has been implemented.

[mahen]

Wow, super impressive ! Great work !!!

BTW does anyone know why most cores are off-centered on my old PAL TV (a part of the picture is missing on the left on the NES, PCE & MD cores) while I have the whole picture perfectly centered with Amiga & ST cores ?

[phoboz]

Wow, super impressive ! Great work !!!

BTW does anyone know why most cores are off-centered on my old PAL TV (a part of the picture is missing on the left on the NES, PCE & MD cores) while I have the whole picture perfectly centered with Amiga & ST cores ?

In order to exclude that this has to do with 50Hz(PAL) vs. 60Hz(NTSC), can you try the Sega Master System (mist sms) core with scandoubler_disable?
In that core you can switch beteween NTSC and PAL.

[seastalker]

Really nice to see this core being developed! Big thank you Jotego & Phoboz!!!!

[jotego]

It's great to see a collaborative effort like this. Keep up the excellent work guys!

Yes, I am very happy to be working in a team this time. It is much better than doing things alone

[mahen]

Wow, super impressive ! Great work !!!

BTW does anyone know why most cores are off-centered on my old PAL TV (a part of the picture is missing on the left on the NES, PCE & MD cores) while I have the whole picture perfectly centered with Amiga & ST cores ?

In order to exclude that this has to do with 50Hz(PAL) vs. 60Hz(NTSC), can you try the Sega Master System (mist sms) core with scandoubler_disable?
In that core you can switch beteween NTSC and PAL.

Hi ! I checked ! So with the SMS and Amiga core, when switching to NTSC, the picture is still perfectly horizontally centered. With the PCE, NES and MD cores, at 60 Hz, the picture is cropped on its left.

[mahen]

Heh, BTW, do you also see some strange "flickering" portions of the screen with some cores at 15 KHz that are not visible at 31 KHz ? I thought that was due to the beta genesis core, but also noticed that for instance in the NES core (in Mario 2). I'm 100% sure there was no such flickering in VGA.
I'm refering to some parts of the screen that are flickering. Not sprites. I have the feeling some gfx artifacts occur only in 15 KHz.

[jotego]

Hi ! I checked ! So with the SMS and Amiga core, when switching to NTSC, the picture is still perfectly horizontally centered. With the PCE, NES and MD cores, at 60 Hz, the picture is cropped on its left.

The difficulty here lies in that we cannot reproduce the problem. Maybe we can add some workaround as having slider controls that affect the horizontal position of the image...

[phoboz]

Wow, super impressive ! Great work !!!

BTW does anyone know why most cores are off-centered on my old PAL TV (a part of the picture is missing on the left on the NES, PCE & MD cores) while I have the whole picture perfectly centered with Amiga & ST cores ?

In order to exclude that this has to do with 50Hz(PAL) vs. 60Hz(NTSC), can you try the Sega Master System (mist sms) core with scandoubler_disable?
In that core you can switch beteween NTSC and PAL.

Hi ! I checked ! So with the SMS and Amiga core, when switching to NTSC, the picture is still perfectly horizontally centered. With the PCE, NES and MD cores, at 60 Hz, the picture is cropped on its left.

The genesis core generates simpified sync signals, while the sync signals are more complex for the sms core. The simplified sync signal works fine on my flatscreen TV, but maybe it doesn't on a CRT TV. There are also differances to when the first sync pulse is sent, so this might shift the image starting position? I will create an issue for this in the fpgagen repository, but a change to this will require a lot of work due to how this was already implemented in fpgagen, and it hasn't the highest prio for now.

[DrOG]

I can confirm the bug report, not only on CRT, but on flatscreen TV too, and this affects the VGA picture also.

[vebxenon]

Thanks again for the team working on this core. That's outstanding!

[Newsdee]

One strange thing I noticed is that the music in Dragon's Fury is slowed down... even though the whole game seems to be at full speed. Not a big deal, but thought I'd mentioned in case you're looking for bug reports / noticeable differences with the original.

[jotego]

One strange thing I noticed is that the music in Dragon's Fury is slowed down... even though the whole game seems to be at full speed. Not a big deal, but thought I'd mentioned in case you're looking for bug reports / noticeable differences with the original.

Thank you. I will check it out. I have heard about games whose music speed depends on the timing of a signal called busy, which the sound chip provides. Because this timing was different in YM2612 (Megadrive) and YM3438 (Megadrive 2) that caused severe speed differences on Megadrive2. The timing of my busy signal might be closer to the Megadrive 2 model.

As I said, thanks and I will check it out.

[jotego]

This has been a hectic week of development. Last sunday I went through all the posts in the spritesmind forum trying to collect any missing aspects of the Yamaha chip. I took notes in the documentation section of JT12 github site. While reading the posts, I realized how the keyon/keyoff (enable/disable sound of each operator) must have been done originally so I went on to modify my code. While doing it, I also changed the MMR unit to something that must be much closer to the original implementation as well. It was a huge change. And when I tried it out, I found that some sounds were wrong, but some others weren't.

It was the most confussing thing. Channel 6 seemed wrong most of the time and channel 1 always right. Eventually it looked like the upper channels were wrong (4,5,6) but there was only one line specific to those channels in the code and that could not be wrong. I wrote a comprehensive random test for the MMR, that would write random information with random timing and then compare the stored information in JT12 with the expected values. After +10000 random writes, not a single miss. That was good but also bad, as I still have the bug there.

I spent a whole day looking at clocks in the MiST system and adding constraints just in case it was yet another bloody timing issue. Now almost every single clock has a constraint! Nonetheless, the problem was not there either

From Wednesday to Saturday I was trying to figure it out until on a different simulation, with visual inspection of the signals I found that the generation of four signals marking the timing of each operator had gone wrong when I changed the counter scheme in the MMR unit. It was a minute to fix it and all the problems went away!

Now I'm back on track with my pending items: LFO and also check the SSG envelope generation as I am not getting the same sound than in the original chip.

[DanyPPC]

Very well, you are a great !

[mahen]

Something is really funny in Thunder Force IV : ingame music stalls until the player's ship fires. Then music seems to be playing at the shooting pace.
(I have no idea if what I say in understandable