#NEC PC Engine / TGFX16
1 messages Ā· Page 4 of 1
I need to check the sanity of my sd card š¢
Good idea
Srg320 has potentially found the cause of those freezes. See the Patlabor bug.
Maybe he can share a testbuild?
It sounds like he's not conifdent that the fix is the right fix
I've replied with sme more context
I will try to take logs of the SCSI bus as soon as I can
@stiff lotus I saw that last night
I am not familiar with the screen this person is referring to, so it would be good to get a screenshot. But here on the discord, not in the GitHub issues
This seems like a relatively straightforward user help issue rather than a core issue
@cerulean saddle can you please photo a screenshot of the issue youāre referring to?
So starting with the left, thatās the coreās/ Turbografx CDās startup screen that loads instead of the game. The middle shot is the screen that appears after I press run/start. And then the screen to the right is what happens if I select delete: it shows my save file for my current play through of Ys I&2. Before I properly enabled autosave, this āformatā screen never showed up, and when I loaded the game, it bypassed the coreās intro screen as well. Other games without save options load as expected, but other games with save data seem to go to this same set of screens, with the exception that ādeleteā doesnāt show any files that have been saved (only for Ys, since this is the only game Iāve saved at this point). I was thinking that maybe I can backup the gameās save data, and then select āformat,ā then copy the save file back into the right location?
Here's what is going on here:
- The Super CD-ROM screen is correct. It should wait there until you press RUN.
- The Backup RAM screen shows when:
a) you press "SELECT" instead of "RUN", and
b) That particular version tells me that you are running a US version of the Super CD-ROM card. - From there, the "DELETE" process looks normal
So, I can only conclude that you are pressing SELECT instead of RUN.
OR
Maybe you have the core setup for 6-button controller, when the CD-ROM (or Super CD-ROM) card is made for 2-button inputs
If it is advancing from the "Press RUN" screen to the "Backup RAM" screen without your input, then you have a problem with your controller configuration
A minor note - there are perhaps 4 or 5 games in the Japanese library which won't run properly on the US Super System Card (so they need the Japanese Super System Card). Ys 1&2 should run just fine though
Thank you so much! It was a controller issueā when I messed with the save settings for the system, I must have also changed the controls and altered the start buttonā since I never saw that format screen before, it really threw me! Thanks again!
Those apparently need to be used together - see the latest updates in the issue
Looking at the code, he's basically trying to associate the status that gets sent back, with the original command that initiated it, to prevent mismatches
Let's see if Tanjou Debut, Patlabor AND Chou Aniki work with that fix.
You'll need both the new MiSTer and the RBF
(Main Mister binary attached to the issue)
Previously, I had found Patlabor to be more susceptible than Tanjou Debut, but it's running strong here
I guess it should also be said that we want to verify that the head seek delays are still working as expected, so something like the opening cinema scene from Mugen Senshi Valis will be important for that
no go with Cho Aniki. the first audio track seems to be cut off early, difficulty selection screen is skipped, and it jumps straight into gameplay demo. Run during intro jumps to demo. Run during demo restarts intro.
if recordings for these tests help, I can get set up to capture but would take longer to post results
Yeah. Did the last set of updates fix or break Cho Aniki ? The ones from 4 days ago which fixed Tekipaki Working Love ?
no that one eventually froze after 45min (estimating, wasn't watching closely)
Cho Aniki is advertising that it ends that track early, when looking in the log output
None of my CD games are loading at all in the latest unstable. Nothing happens after ājust a moment..ā They work fine on the stable
did you grab the specific mister binary that is attached to the issue? https://github.com/MiSTer-devel/TurboGrafx16_MiSTer/issues/220#issuecomment-2769203896
I did not thank you
srg320 has just created another test version of Main MiSTer - in the most recent update to the issue here:
I won't be able to test until after office hours (3 or 4 hours from now), but I suspect Patlbor and Cho Aniki are now working. Other games should probably be tested as well
he rollbacks some changes on the core also
I'm not convinced that those test cases cover a spectrum of situations, so games are a better judge
Darn that 7MHz mode. Between Puyo Puyo and Outrun, it seems to be teetering on the brink
I tried to revert the commit because logically it should be the only one that could interfere.
And that 7MHz mode...
I've missed that when doing my tests some weeks ago.
srg320 needs to fix that
he overhauled the whole 7MHz mode, and that was the one tiny tweak which looked like it was need, but undoes OutRun
late feedback. using the latest unstable mister binary in the GitHub issue and unstable core... Cho Aniki intro is working again, but still freezes after about 45 minutes of demo loop
Interesting
fwiw I became interested in this because I was learning the game and it would eventually hang at seemingly random loading points. left it on as background noise one day and saw the hang during demo loop. decided to check this thread to see if it was a known issue
I have no real hardware experience with a TG16/PCE otherwise, but happy to help test where/when I can
srg320 seems to be busy working on 3DO for the time being; I'll get my board tomorrow to capture logic signals in order to start making progress again
@grand moss , great work at ripping the last mysteries from the soul of the tgfx Duo.
I'm sure there are still more...
At least, that Outrun/Puyo Puyo graphic issue is likely not trivial
hhahahah !
Pretty much my main thing
Patreon time? š
Nah, I don't want to feel obligated. I have considered Ko-fi from time to time (for "thank you" instead of Patreon's "please"), but I haven't gone to the point of making one.
Well we owe you some back pay, so if you ever get around to itā¦. š
I donāt understand most of this stuff under the hood, but it is so damn interesting. Video game archaeology for sure!
Man, that PC Engine disk drive is SOOOOO weird.
It's so cool that you're digging into it
No idea whether srg320 is looking at the results though
He certainly could be. I think there's several days' of work in there
Yeah, he seems interested in the problem (which I don't understand but that's nothing new). Hopefully he can help out
Wasnāt the drive the same used in the PC-98 computers?
Not that this bit of info helps anything lol
Thank you for the searching š
PC-88, and it doesn't help the PCE, but it may help the PC-88
Coolā¦someday I hope to check out PC-88/98.
Happy NecPCEngineday!
cookin'
Is that needed for the current unstable core to work correctly with CD games? If so, can you link the binary of that Main?
the binary is present on the link above
the new unstable RBF hangs after press run in CD mode.
The last main mister linked above with previous unstable no music.
Maybe something not right with the auto build
There is somethig broke on the last core, the CD-ROM² games can't load anymore.
I'm not having problems - needs both the new Main_Mister (need to boot into it), and the new core
I tried with my compilation too, it's something with the last changes.
Must be on the CHD side.
New Main and new TG16 core are a matched set; neither will work with another version of the other
I know.
I've changed the Main file too on all my MiSTer systems.
From the one uploaded by srg320 1 hour ago.
Bob Morton made a mistake. I am here to correct that mistake
You're fired!
thank you.
Without joking this time, I will try with a regular BIN+CUE.
OK, I'll let Patlabor intro run for an hour or two and see if it's also fixed. I haven't tested anything else yet, but it booted fine for me
BIN+CUE on your side ?
I can also test bin/cue if that is helpful
Yes, do it too.
kk
me too using bin/cue and it is stuck on loading screen of the bios
Yup, sitting on Just A Moment
Are we all using the ini trick to load the bespoke main?
Not me. I moved the previous MiSTer to a different filename, and created a new MiSTer, and waited a bit for it to be committed, and then rebooted
Put the special TG16 core in the root folder with a special name, and started that one up.
My games are all cue/iso/wav from my own collection
So far so good on Patlabor
ok, let me try that
I've changed my Main MiSTer file, tried with today's core and a collection of BIN+CUE from my collection + the same ones on CHD.
No dice.
If no issues in about an hour, I'll probably do a quick test on Puyo Puyo and OutRun before flipping over to Cho Aniki and Tanjou Debut
Very strange that we're getting two different results on the basics
Wait, you got Puyo Puyo to run - what CD is failing for you ?
Sorry, OutRun only, I've mistyped and done two things at the same time. @grand moss
yeah, still sitting at "just a moment" even after swapping main
I'm checking some Puyo Puyo VS on the Mega Drive core too and read too fast.
What game are you trying to boot, and what BIOS (and is that BIOS in the same folder as the game, or a folder above, or where ?)
Patlabor
I've tried with Bonk III, Akumajou Dracula X and Palabor on my side.
Files inside the TGFX16-CD folder, on an external USB stick.
mine are on a NAS
You're not using a special BIOS, right ? Just plain old, stock Japanese CD 3.0 Super CDROM ?
Yes for me.
whatever update all grabs
(I stay away frm update all, so I can't comment, but I guess I'll assume that it's plain)
It's plain on the update_all system too, I can confirm.
Well, it sounds like a version mismatch, whatever is happening - he separated audio from data channels, so communication would be disrupted if they aren't matching versions - which appears to be what you're expereincing. Although I can't explain why
With the new main mister (build by srg320) I can run cd from previous update without sound but with the new rbf unstable it is stuck
so i'm trying to build the core on my side let you know
ok
Same result on my side.
You guys are both on stock Terasic DE10-Nano based systems ?
yes
yes
I've already tried, no change.
Standard seed and seed 10 too.
Tried on DE-10 Nano and MiSTer Pi, same result.
Dual and Single.
you tested everything !
OK, you've got more configuration than I do.
I've only got 2020-era DE10-Nano based, single-RAM setup
I've 5 MiSTer at home at the moment (DE10 single, DE10 dual, QM Tech, Mister Pi and a MisterCade), I will equip a retro convention next week end.
Great !
I can cover all bases for a week.
So i will roll back on my side testing next one I think
I'm leaving everything we've done here in the last hour on the issue log, I gotta go (and that wasn't expected at all on my to-do list today.)
OK, I'll keep testing (since I can)
Are you guys overriding the core's MiSTer in ini ? And if so, are you sure you're pointing the correct version ?
Sometimes when you simply overwrite a file, it doesn't "take" (seen it in the past). In general, I need to move the existing file to a different name, and create a new file.
I've been hearing a lot about confusion around this from Slamy
I'm not overriding anything, as I'm changing the whole Main file and using only one for all cores.
We were refrained to use multiple Main, back then when testing the CD-i core.
And same thing about replacing an exsting file versus moving it to the side and making a new file
I've had overwrites "not take effect" in the past
Seems like uncommitted/unflushed buffers
(which is why I do it this way now)
I'm powering off and on my MiSTer systems 1 minute after replacing the file, just to be sure that I'm flushing everything correctly.
I understand - make sure that the MiSTer file is recreated new, rather than overwritten though
That's what I do too.
OK
I did what you did, @grand moss. I renamed the other main to MiSTer_bak and then added SRGās bespoke main
Running the cd_verificator tests, it looks like there's a change in behaviour on the init_cd section which hangs on the verificator test
No issues in Patlabor; I found something in the CD init on verificator; moving on to Cho Aniki
So far so good on Cho Aniki
Hey guys, you want to try something ?
put this at the tope of your mister.ini file (just below "[MiSTer]":
I just realized that I run with this on, and maybe you don't
yes debug=1 it is working now on my side
Cool, but weird
That's definitely some sort of timing thing
Works on all my MiSTer systems too.
OK, there's still a bug (it's actually the read-table-of-contents), and the timing needs to be fixed on that, but at least there's a workaround so you can test other things are working
Couple of new graphic issues with the new core, but other than the debug=1 requirement, the CD portion looks to be working well
Going to run Tanjou Debut for an hour or to now
@faint dagger Are you cooking on all three stack types? Or should I fire up on a de10/qmtech/pi?
I found - during the opening loop of Patlabor - a graphical artifact at one point. I have confirmed it doesn't exist on original hardware, but I haven't confirmed whether it's new to today's core:
Fixed on all types.
See the horizontal light-colored stream at the top of the scren ?
I don't think I saw that in the past, but I didn't want to raise it until I confirmed:
- That you guys were able to test - looking good now
- That it doesn't occur on original hardware - confirmed
- Whether it is a new issue, or an issue that we didn't previously report on older versions - maybe one of you guys could check that.... if not, I can check in a few hours
I'll look right meow
Either way, it's really minor
it is just for test purpose I think. You live on the edge š
debug is just for the MiSTer main half - doesn't affect graphics, just CD, gamepad, and so on
OK thanks
Yeah, he made changes to two distinct things today - one was graphics for PuyoPuyo/OutRun, and that's the one that has upset Sherlock Holmes and this smudge
The other was CD, and - aside from the debug = 1 thing - seems to have fixed everything
(as far as I can see so far)
Oh wow, he re-enabled the READSUBQ debug output, and it's only reporting once per sector
just catching up on the thread. is this one worth testing since I'm the jerk that keeps seeing issues with Cho Aniki?
David took a look and didn't find anything, but you should definitely check it out too
just make sure you have the bespoke main and debug enabled
Yeah. I ran it 75 minutes without issues.
Tanjou Debut also for an hour, and Patlabor for an hour and half
I've created issues for the two graphics things
okay cool. I've got it looping in the background and will report back
need to compare against older versions and find hardware recording, but it sounds like the demo's death sound is cut off instead of playing to completion
You talking about the "in game demo" where it shoots randimly until it gets killed ? The "getting killed" sound ?
yeah that one
my memory is kinda crap but I thought it played to completion previously instead of cut off abruptly
but I also don't know what it's supposed to do on real hardware š
Maybe you're thinking of the title screen explosion ?
...which seeme to be correct I think
yeah I'll check it against last stable core once the loop gets past 45min or so
I'm running it on real hardware now - you mean the "hu-aaaaeeeh" ?
yep, that one. I thought it played completely on stable core/binary combo
I'll run new core now to see what it does by comparison to real hardware
Sounds the same to me.
sweet ... it's been a while so I couldn't remember
Wow, I hope srg320 doesn't feel like I'm giving him a bunch of new work
(filing new bugs)
I mean, the goal of MiSTer is to be as accurate to original hardware as possible, so this trial and error just comes with the territory.
well it froze at around 90 minutes
We need to prevent him from wasting time on a 3DO core. TG16 rules š
That's right, gotta get the 6270 video working perfectly so it can be a building block in the greater PCFX scheme of things
Anybody find any freezes yet ?
I would have to check against the stable, but the crop seems off in Patlabor (the intro anyway). More blackspace on the left than the right
ehh, its on the tanjou debut intro as well - might be me
No, crop is definitely off (might be related to the Patlabor bug?)
New update fixes CD loading no need debug=1 anymore !
Wow, I just looked at the codebase and realized that the last official release of the core was one year ago from yesterday
Anyway, if new core doesn't require "debug =1", we need to check for any additional hangs when playing when it's removed
I can do some testing. I know my picture is bad, but the crop is definitely different in the unstable. I dunno if that is something we care about or not
If something is cut off, it's an issue. If it's just extra blank space, gather more information first
yeah, just extra blank space
it's shifted to the right (but nothing is missing as far as I can tell)
So, I guess that there have been no hangs found by anybody on this new version ?
(I didn't see any on yesterday's core with the setting "debug=1", and I haven't seen any on today's core without that line)
Ran Cho Aniki for about an hour this morning, Tanjou Debut running now...
I just booted Patlabor
There is a another small graphical glitch in the patlabor intro. I'll grab a screenshot of it in a sec
very minor
New unstable incoming fixes on VCD part
testing patlabor with the new unstable
Maybe put it together with the other one I reported yesterday
The other one is fixed by the new core
I am checking this new one against the stable right meow
oh cool, I hope to close a few issues today then
Ok, this is pedantic as shit. Stable on the left, new unstable on the right. See that single red pixel at the bottom of the screen on the right image? Think it's worth reporting?
But the other one is fixed
gonna let patlabor run
I can look for it on real hardware
that would be amazing if it was there. Let me know and i will file it if not, I guess
and apologize to SRG for wasting his time with it :p
@grand moss do you want me to look at sherlock holmes? or just let patlabor run for 90 minutes or so?
It's there on real hardware !
I'll check Sherlock
Just let Patlabor run
Amazing. Finally the core is accurate!
ok. I will let it go while we go get some lunch
man, some old bugs are going to get closed today
Shit, I think the intro to patlabor just crashed :/. Booting it again
yeah, crashed again :/
didyou have debug=1 or not ?
no
While running Sherlock Holmes' introduction video, there was one run where the video became garbled (without "debug=1").
But I couldn't reproduce it
I feel like that might just have been timing - like the data wasn't available in time
I wonder if that might be what you saw here
Run it again for a while, and then run it with debug=1 and see if it stops the crashing
Cho Aniki hung around 90min in the previous build. I'll test again with today's build later
Please test.
I never saw it hang on my system, stopped it after 75 minutes
Ran it for another hour this morning without debug
looking back at my own notes in the channel, it was about 45min on stable releases and it's at 90min now. I'll report back when I get a chance to test it
Dang another new build dropped?
Wat, another ?
Oh, that's the one from 2 hours ago
Already did a bunch of testing and closed 3 issues - but opened one and reinforced one
Did you get patlabor to crash?
Not yet, just Cho Aniki
Ok, Iāll load it back up when I get home. With debug off it only took two or three attract cycles
You saw Patlabor crashes ?
If so, try setting debug=1 and try again, to see if that affects it
Yeah. It was quick. Saw it twice
Might be related to SDCard speed or something. If the frequency drops significantly with debug=1, it's probably a timing issue
All the roms are on a nas. Should I try an sd card?
yeah
queuing up a test for cho aniki loop, with debug=0
debug=0 ?
I've only used debug=1, or removed the line
yeah it was removed
So this core is perfect now?
Is that the last issue?
no
The CD verificator and SUBCODEQ are effectively no impact
Wow just four issues and three feature requests
CD+G, SaveStates, and TGX extension are requests
pretty amazing
It's been pretty much this state for a couple of years
The issues that remain are not serious
And I think the Cho Aniki and new Sherlock Holmes issues are possibly a small adjustment to what he's been touching the past couple of days
It's less than that now, but TheJesusFish was seeing it more often on his NAS-linked images
I should probably mention that I had seen real hardware lock up back in the day - extremely rarely, but on more than one occasion
Iām running it now while I rock the baby. Debug 1 and still on the nas. If I get a crash here, Iāll move it to the sd card and go back to debug 0
OK. I haven't seen it on Patlabor, only Cho Aniki
Amazing also that it hasn't had an official release update in over a year
But best to let the current situation calm down and stabilize a bit first
Its fun when a core gets to the āwell yeah, if you let the attract cycle play for 8 hours, it crashesā state of doneness
Yeah, at what point do you want to say that it's not really an issue
yeah still hanging after 90min. tbf, I only noticed the hang because I was learning Cho Aniki and it has infinite continues
Oh so it hangs while playing too?
Rayxanber 3 boots but does not progress past this screen on the latest unstable. This game works fine on the 4/4 unstable as well as the stable
yes. at this point I'm honestly starting to wonder if it's related to how many cd track changes there are
Do you have the bespoke mister main as well?
There was a new one as of yesterday
Are you doing chd or bin/cue?
bin/cue over network share
I donāt. Is it somewhere other than unstable nightlies?
Yeah, sorry, was attached to the issue https://github.com/MiSTer-devel/TurboGrafx16_MiSTer/issues/176#issuecomment-2815405884
Original statement of issue was here: https://misterfpga.org/viewtopic.php?p=50704#p50704 Two actual issues were reported: a) Choppy sound due to excessive logging of "PCECD: Command TESTUNIT,...
Huhā¦so weird that itās around 90 minutes
And do you have to be playing? Not just rolling through the attract cycle over and over
When it was hanging in the past, it would hang in the intro loop
Which game is hanging - there are two conversations. Are you talking about Cho Aniki ?
Are you using the special MiSTer file and the core from about 6 hours ago ?
Changed to this and it works fine. False alarm sorry!
OK
No worries, itās hidden. The unstable going into the unstable repository should have had a note š
So in summary, I'm seeing Cho Aniki hang during intro loop with no debug, and on SDCard; albertprime is seeing it on Patlabor on a NAS share. Anything else ?
Is the mister main not in unstable nightlies because it breaks other cores for the sake of the Turbo?
It's because it breaks the official TG16
and he's not in control of that core, so wants to make sure all the changes will take effect at the same time
Ah ok so just the current stable is broken
No, Iām seeing it on patlabor in a nas share. Albertprime is seeing Cho crashes after playing for 90 minutes
OK, is @turbid jetty also on NAS ?
Either of you using debug=1 ?
I just enabled it before coming up to rock the baby. I havenāt been down to see if patlabor crashed again
It was crashing under debug 0 for me
OK, I've been looking Patlabor for hours, and it's not having any issues here no debug
By "crashing", you mean "hanging", right ?
To me, a "hang" is to become unresponsive, whereas a "crash" shows some weird berzerk behaviour
nope...no debug
Ok, itās still playing down here with debug1
how long ?
An hour
It was hanging after like 10 minutes before
Iāll try again with debug 0
Ok, not hanging now - I dunno, I was able to reproduce it earlier, but now it seems fine. Transient issue maybe
That's no debug and on NAS ?
I think Patlabor and Cho Aniki are wo separate issues though
Just want to make sure you're not mixing them up
yeah, I was definitely seeing hanging earlier today on Patlabor intro cycle and debug 0, but further testing this afternoon, I haven't seen it in an hour on both 0 and 1
I have not loaded Cho Aniki, so not mixing my meats there
OK just confirming
I've been running Patlabor for about 5 hours here, and I'm starting to get tired of the music...
What is TGX?
Exactly.
It seems to be some random filename extension used by somebody somewhere, but I never heard of it
We should close that issue then, what a silly request
Thanks @summer lintel !
@grand moss why is Jinmu Denshou the best game on the PC Engine?
wait did TGX get added

Not really mocking per se, just pretend outrage
that was on behalf of a friend who says it was an extension they came up with with cowering or something and that software emulators support it
I just thought it was funny you are one of the open issues
tell your friend to play Jinmu Denshou and if he beats it then then the extension can be added
Edit the comment!
Anybody having problems with CD games on the latest unstable?
I'm not getting any CDDA audio during games intro and they won't load any further.
You need this main https://github.com/MiSTer-devel/TurboGrafx16_MiSTer/issues/176#issuecomment-2815405884
Original statement of issue was here: https://misterfpga.org/viewtopic.php?p=50704#p50704 Two actual issues were reported: a) Choppy sound due to excessive logging of "PCECD: Command TESTUNIT,...
I had another patlabor freeze letting it run over night. 2/3 freezes have been after the river hill animation, when it's showing the copyright information.
I moved it to the SD card and am now running it there
It solved my issue, many thanks.
So the solution is to change the main MiSter? Never done that before...
The test core has a companion test MiSTer. He changed fundamentally how they communicate
Obviously when this becomes part of the release, it will need both to be updated
Guess I'll stick with the stable release until everything is sorted out then...
Admittedly how srg is handling the TurboGrafx-16 core updates is kind of confusing.
All depends on the nature of the issue being fixed
That's when it has traditionally had the problem
I've been running Cho Aniki for a few hours on this morning's core, and no issues. SDCard, no debug
I am out right now, but I left it running back home
sd card, debug 0
it was fine after an hour - it may become like we talked about: "after 36 hours on the attract cycle, game freeze"
Came back to this @grand moss - thatās sd card no debug
That's today's, right ?
yup, when I took the pic
I didn't try the new RBF that he attached to the issue, though
I wasn't sure if that was for Cho
No, I mean the core - after srg320 said "Unfortunately, I don't have any of these issues. I need to localize the problem. I made a test build without the latest SCSI changes. Please try it and report back."
Yeah, try the new one
I haven't had a freeze in Cho Aniki, and it's been running maybe 4 hours here
No music soundtrack for me when playing Rondo of Blood (CHD) with yesterdayās latest core
Testing now and the last build it was working was 25.04.04
and you have the bespoke mister main from the issue?
My English isnāt that good but I think you are asking about which main version am I using , right? If that is the question, I am using the latest unstable Mister main
ahh, yes, then you need this version of main https://github.com/MiSTer-devel/TurboGrafx16_MiSTer/issues/176#issuecomment-2815405884
Original statement of issue was here: https://misterfpga.org/viewtopic.php?p=50704#p50704 Two actual issues were reported: a) Choppy sound due to excessive logging of "PCECD: Command TESTUNIT,...
Thank you!
@stiff lotus can you pin a message that says something like "If you are using the latest unstable, you must also download the Mister Main from here. NOTE: The version of Mister main will BREAK the stable core's CD reading - use only if you want to test"
@grand moss had another crash on the test core, but realized the nas was still mounted - doing the sd card now. Also, looking at this video, I think the sound is slightly desynced https://www.youtube.com/watch?v=Qar3Olik3YU (looking at the beginning where the robot mans appear against the red background)
Nine hours of Cho Aniki, still going
entering hour 3 of patlabor, still going
No, that's pretty much how it's supposed to look
the timing of the music is slightly off (at least on my setup). In that video, the robot slides in and appears on beat. On my game, itās like 1/3rd a second early
I'm looking at real hardware versus MiSTer, and I'm not seeing a difference
Games with lip movements would be more obvious though
ok cool, if it matches real hardware on your end, then itās good
what did we lose by SRG rolling back the SCSI changes?
@turbid jetty If you have time in the near future to go back to the Cho Aniki mines with the latest test build, that would be awesome
Theoretically, no gain no loss.
The difference is whether the SCSI signal timing matches the original machine
We were looking at that as a potential source of the hang instability
so far so good
I just left an update
Came back down stairs and patlabor had hung again. This time at the title.
- Latest Core
- SD Card
- Ran for 5ish hours
no, it was SD card and no debug
but then I left it going over night and it's fine this morning
so it's either related to my setup, or not easily reproducible. I'd vote to close the issue since no one else has hit it
Well, he would still need to commit the latest update (this core was temporary and isn't committed)
oh right - but yeah, if I am the only one having issues, then itās probably not worth keeping open
I'm happy for him to commit as-is, I just wonder if anybody else is testing at the moment
Either way, this is lower hang frequency than before the changes (of the past 2 weeks or so), right ?
Oh yes, much lower. In probably 15+ hours of letting it run via SD Card, I have had it crash once. And it went all last night without a crash
I can run some tests on this when I get home @grand moss (if that is helpful). I noticed in the (neverending loop) of the Patlabor intro music, there would occasionally be a hiccup. Never back to back and rarely in the same spot (that I noticed). I wonder if this is what happened to you with sherlock holmes
First of all, what brand of SDCard are you using ?
Samsung has some pretty long gaps at times
It is a samsung card: https://www.amazon.com/dp/B09WB3D5GQ
Samsung PRO Endurance New Portable SSD for monitoring devices is built to last up to 140K hours of recording (256GB model).** This solid state drive stands up against magnets¹, X-rays², water³, dropsⓠand more. Record and play in FHD and 4K. The PRO Endurance is engineered to deliver long lasting...
I'm using SanDisk Ultra (the Silver one), and never get that problem (that I can see). I'm pretty sure that the Sherlock issue I saw was related to a read request having some deformed feature, and missing its schedule between cores
Never saw it again, and I will close it soon if this core gets committed
sounds good
missed this with all the drama kids play fighting at Wrestle Mania. taking a look sometime today
david ran it for 9 hours or something without a crash, so hopeful the new core fixes it for you too
sounds good. is there a new unstable mister main binary?
okay weird....this froze in about 15min
NAS
maybe try SD card, if it's not too inconvenient
giving that a go now
froze around 45min
I've been testing on a qmtech with soldered ram with a second memory module. i have a real de10 with single module that I can try later with a clean memory card. starting to question my setup/methodology if everyone else is reporting success
No, i'm still having crashes in Patlabor's intro
I think dshadoff is just God's favorite or something :p
new unstable still needs bespoke mister main, yeah @grand moss?
New unstable uses the same special MiSTer from a week ago
You say it freezes on NAS, and it also freezes on SDCard - what brand ?
Is it Samsung ?
srg320 has the same results as me
But if you guys are still seeing problems, you need to pop in to the issue log. It seems that he wanted to try to figure out what's going wrong by running more tests, and I'm now disqualified, because I don't see any issues anymore.
yeah, 256GB Samsung Evo select
Yeah, I know that Samsung is known to have severe pauses in processing at times, on amny differnet types of devices. Fast when available, but pauses.
I think that might be what's happening here
Any other CD cores have similar issues ?
(including gaps in music playback)
I've always loaded from NAS so this is really the first time I've had issues. I have three other brands I can try later
I use NAS for everything - I think this is the first time I have loaded a game on the SD
Yeah I'm wondering if it's just sensitive
I commented on the patlabor issue
With the current SCSI rollback: is that back to the state of March 7th? Or did that roll back just some specific changes?
Just some specific changes
Is the current "unofficial-unofficial build" of MiSTer Main still needed?
I don't understand your question
Of mister Main
You mean "new MiSTer executable needed" ? yes. And the latest of latest of unofficial RBF files which was only pasted in the issue log
Thank you, David
It's getting hard to keep people coordinated with proper versions
The last Tgfx-16 secrets are hard to unveil obviously
yeah
I know. Thus asking in advance.
Appreciated. Better to catch it early
just to be sure, the mister binary to test with is the one in the pinned message?
It's near the pinned message.
Here:
https://github.com/MiSTer-devel/TurboGrafx16_MiSTer/issues/176#issuecomment-2815405884
Original statement of issue was here: https://misterfpga.org/viewtopic.php?p=50704#p50704 Two actual issues were reported: a) Choppy sound due to excessive logging of "PCECD: Command TESTUNIT,...
https://github.com/MiSTer-devel/Main_MiSTer/pull/983 our bespoke main days are numbered
Ah OK
Is this the one thatāll break the stable PCE core?
Sure. Just wondering when I need to move over to the unstable PCE core too š
On the bright side, you will be able to watch the Patlabor intro on repeat for HOURS without a crash
(you're welcome)
I appreciate everyoneās work in here for trying to make the core better for everyone
Even if weāre getting to super niche stuff now š
I was making a joke, but I barely did anything
You tested, which helps to narrow down the problem
š
More than Iāve done š
The changes for the unstable are now in unstable main: #unstable-nightlies message NOTE: Upgrading to this main will BREAK the stable core
Just to be clear, the most up to date and functioning Turbo core is the 4/21 unstable core + todayās unstable mister main right?
yep!
but using the new main will break the stable core's ability to read cds
Got it thank you
Does this mean going forward everytime a new stable mister main is pushed there will be a corresponding unstable main for the most recent turbo core? At least until the issue is fixed
no, these changes will make it upstream to the new stable main when that is released. I believe the plan is to publish the unstable tg16 core as stable then as well, but you would have to ask SRG and @grand moss to be sure
Got it thank you again
The goal at this time is to push the MiSTer and TG16 changes to proper release in the near future (next couple of days I expect), and then unstables can go ahead without needing this synchronization
At that point, new unstables probably won't add much
(until somebody adds CD+G)
CD+G is the only request that seems like it will need a MiSTer revision update, but even that shouldn't interfere with basic operations
The whole mister is a paperweight if it canāt play talking heads
If you want, you can play it on the Sega CD (MegaCD core ? Or was it another one ?)
Zakk put CD+G in way back in 2021 or so
ARRRRGHHH the song is now stuck in my head!
What song is that ?
(It's probably not one that's on CD+G...)
You may find yourself in a beautiful game room, with a beautiful PC Engine.
And you may ask yourself, well, how did it not sell in the US.
š¶ Letting the shmups go by, let the turbo go to town.
š¶ Letting the shmups go by, turbo flowing shots down.
š¶ Into the Bonk again, after NECās moneyās gone.
š¶ Once in a lifetime, turbo flowing shots down.
Oh boy, Robby stopped making sense
When has he ever made sense!?
it's true, he is constantly speaking in tongues
Probably the lack of sleep from crunch, kids and pets
Have the outstanding issues been resolved with the latest Main and core or is there still more work needed? I've been holding off until it stabilized.
It's pretty stable, but at this point you're best to wait for the next official release of both MiSTer and TG16 which should be any time now.
Yeah, I've been holding off since I don't want the lastest unstable main to break the current stable TG16 core.
I do wish stable cores were, in general, more frequent
I believe Sorgelig wants to be in control of that (for the most part), and it also includes updating the framework to the current version. I believe that he enjoys working on improving the framework, so more releases is counterproductive to advancement... though he's always been accommodating when a new release is specifically requested.
dude, did you make up those lyrics?
Depends on whether you like it or not
I enjoyed them a tad too much I'd say š
Then yes lol
Did the core change recently? It won't show any of the .chd I've used before when trying to load a game
Doesnāt it say in the unstable nightlies that the unstable breaks CD loading - for testing only?
yeah, you need both the unstable main and the unstable core - having just one or the other will break cd loading
that must be it, thanks
So I haven't really kept up with the latest unstable builds of the core. Is it worth updating to the latest unstable main at this point? Been avoiding it since I didn't want to break the stable build of the core. But if the latest unstable build is in pretty good shape now, it might be worth it. I usually keep up to date with unstable main since other cores use and benefit from it!
New stable should be coming soon
Might as well wait I think
It requires new main mister which has other interdependencies too
Not sure why it hasn't come out already
very fun game. got to my 2nd car already
What type of game is it? Doesnāt seem like itās only racing.
I couldn't figure out how to change gears correctly in the drag race mode. It seemed like you needed to hold the break button and then press the d-pad to switch gears. But holding break slowed the car down, making it impossible to win?
Well, here's a bit of text describing the game (though you might have read it elsewhere already)
its basically a life simulator where instead of pursuing girls you spend all your money buyjng new gear to try to reduce your time in 400 meter drag races
its not the brake but the clutch for the gears (it helps if you drove stick before). Basically when you push that pedal the gears arent connected to the engine, so you slow down due to friction
you need to be fast to switch gears
its a simple game but I'm finding it very fun. You can feel it when your fast gear shifting is beating a more powerful car...
Anyone know if the Dailor Moon game for PCECD has multiple endings? (Per character)
You're talking about the one that got a translation ? I haven't done a lot of playing it, but my understanding was that each character had a different storyline, but I don't think there are multiple different endings for any given character
The ending basically said "too bad we never found out what was with Past Wiseman".
Maybe you need to go through with all the characters
That's either a clue to a better ending, or lampshading a dropped plot.
I thought that but they don't ask you to save when you finish the game. Does it save automatically?
I don't know offhand. It's a late enough release that it could
The description from the translations says:
All five of the original Sailor Senshi are available as playable characters, each with their own story; their five separate plots overlap to form a larger whole. It features extensive voice acting from the cast of the TV show, as well as two original vocal songs
That implies that playing one character won't give you the whole story
I have a question. Ive got a cue/wav (even has a .iso in it??) version of Castlevania Rondo of Blood with the arranged soundtrack, etc etc. Works just fine on my MiSTer, but Im wanting to convert the thing to chd using CHDMAN but every time I try the file it creates, it doesn't boot. Ive tried the resulting chd in my MiSTer. Im lost as to how to convert the thing. I even tried using imgburn to create an image from the .cue and the resulting ISO is also not usable in something like MagicEngine to test. Not sure what I'm doing wrong.
Might just not be the right format for CHD conversion? Either way, at least it works without CHD, so I'd honestly just call it a day on that front. Unless you're working with an extremely small micro SD card, one CD game not in CHD format should be fine.
I thought the same thing, but what Im trying to get at is getting the cheats to work. I can get it to work with the standard english patch and even the original, just takes renaming the file appropriately, but even naming the file how the folder is "Castlevania - Rondo of Blood - Arranged Perfect Audio + English Text (Japan)" doesnt work. And if I try to simplify (renaming to like "DracX Perfect" or something) it still doesnt work because the .cue file has the original long version of the name there so the game wont boot. I COULD go into the .cue and rename all of those, but I was hoping a quick chdman convert would avoid that :/
Some rom hacks don't work with cheats, regardless of how they're named. Might just be that.
yeah could be
Romhacks should work fine, as long as the CD image/cue names (and the image name inside the cue itself) and the cheat zip are the same. Never tried the arranged soundtrack though.
Use a simpler name, maybe without + in it.
I'm also using the normal T-Eng patched version. Cheats are fine as you said.
I ended up trying that, opening the cue in notepad and editing it down to just "DracX", and then it wouldnt boot on the MiSTer either lol. I guess Im just gonna leave it as is and not have cheats on it š
Just out of curiosity, why are you using a "remix" soundtrack version ? Other people have complained that the sound levels become imbalanced (i.e. CD-DA versus ADPCM versus PSG) - I've noticed that some "remix" soundtracks have improper levels, and others can cause problems if the table-of-contents is not exactly the same as the original
I would imagine that in this case, the cheats could also activate based on a CRC check of some sort (which could be different on such a rip)
Just to try it, honestly. Im fine with the original translation but much like the SNES MSU-1 ROMs Id like to just fart around with the remix version. I hadnt heard about imbalances in the sound, though. Thats a bummer.
Well, I understand that people want to mess around with various parts of the game that they are able to manipulate, but often they don't consider all of the things that were necessary in the original release. I recall hearing about a Lords of Thunder remix where the CDDA was simply made louder, and people seemed to like it... until the PSG sounds couldn't be heard.
Noticed some graphical issues on Rainbow Islands on the latest unstable that is not present on the stable. The score numbers float through the āgoal inā text and cause sprites to go wacky.
Happens on both the 4-22 and 4-30 unstable mains as well
Those aren't textures, they're sprites
Edited
Shouldn't be related to Main; that's related to the the whole sprite fetch item affecting Puyo Puyo and Outrun
How easy is it to get to that part of the game ? Whoever is looking at fixing it will need to reproduce it
Itās the very first level. Takes about 15 seconds from start
Does a person need to complete the level for example ?
Yes the āgoal inā screen is the level complete screen
You just need to make a few very easy jumps
A bunch of CD games are broken on the stable if you are using a recent unstable main
oh yeah, right
OK, I created an issue
Thanks
Thank you to whoever fixed this in the latest unstable!
srg320, a God among us.
Just tested both Puyo Puyo and Outrun and everything looks as it should with the latest unstable build
No regression found so far.
Cool. I'll check it for myself soon, adn close the bug
@grand moss do we know if SRG is going to push out a new stable? Or is cd loading on with the latest stable main and stable core?
The latest stable main and stable core are in sync; the latest unstable main and unstable core are also in sync.
wickerwaka had requested Sorgelig to make a new Main release, which he said he would try to do today - Sorgelig knows that both Saturn and PC Engine (at least) also need to be released at the same time
So it's up to Sorgelig at this point, and should be very soon
New main is out, so yeah š¤ #news message
This is exciting
Real PC Engine fans know what this is
I'm not a PC Engine fan...
Real pc engine fans would know the pc engine only used heat sinks
Does the stable core work with the new stable main?
still waiting for a new release of TG16
unstable TG16 will work with new stable Main
yeah, I wasn't sure if I missed something because I would have thought that they would have pushed the core release by now
since it's incompatible
Me too, but I'm not in control
Bikkuriman?
Yup!
Yeah I liked Wonder Boy in Monster Land too.
I've asked sorgelig to make a release of TG16, so I think it should come soon
He thought there was still work being done on it, but it's stable enough to release
Do you know if a Saturn core is coming too?
I didn't ask, and I don't know if Saturn is also co-dependent with Main
No problem. Iām not sure either
Hmm... I thought he would have released by now...
Did SRG wave him off ?
I don't see any conversation with srg
only that sorgelig said "sure" a few hours ago
Maybe overnight then
Oh there it is
Great! Now I need to update my MiSTer and play PC Engine. Thanks guys
Radio NEC'stalgie, a PC Engine-dedicated radio, is online.
https://t.co/YHjvJVsOxe
I left a note and a thanks to Seb a moment ago.
I know him from the PCE Bible creation back then.
Oh, OK, looks like it came out basically the moment I said I didn't know when it was coming out.
And you can play the RPG of the quarter! Ys 4! 
Please donāt tell Quirrin
Hi. Does anyone know why my chd file for Rondo sounds much worse than using a bin and cue file?
Maybe some kind of bad compression or curruption when the CHD was made?
Did you convert it to CHD yourself, or did you obtain it from somewhere already in CHD format?
Maybe better to just say I did not convert it myself
I'd see if you can convert it yourself. The one I'm using that I converted myself sounds great.
Just a heads-up, the CHD conversions floating out there are of terrible quality and using a really old version of CHDMan. Weāve been telling people to avoid downloading those and to make your own.
Is there no good CHD pack for PC Engine CD?
pretty sure htgdpās pack is good
Who wants to play with CD+G ? #test-builds message
What contents used cd+g ? KaraokƩ ?
That's one
Thank you š
Neat, I had a cd or two with cd+g back in the day but never tried them on anything. This must be the first core to support it right? Looks like only Sega did with segacd and saturn not PlayStation (and PCEngine obviously)
@stiff lotus do a voice chat karaoke stream
No, MegaCD supported it in 2021 or so, but it had some issues. Not sure if it's been maintained since then.
Great work David!
I don't know if anyone has tried to put together a pack of CD+G games, for people who haven't swam in these waters before, Redump doesn't support CD+G because of the subdata
My two CD+G discs:
That site I linked above has a pretty comprehensive list of what was available in the West. Japan still uses it for some karaoke music.
The standard way to get *.sub dumps is with either CloneCD, or I saw a Japanese program called CD Manipulator
The output is generally:
*.cue
*.img
*.ccd
*.sub
cue/img is basically just like cue/bin
.sub is the data being used for the CD+G
and ccd is similar to a cue file, but is able to make references to the sub file. (cue files don't make reference to sub files)
There is no parsing of ccd files in Main. Instead, I just take the base filename and switch out the extension to '.sub' to see if such a file exists, and use it if it does. The assumption is that there is only one sub file for a given disc
Ah, are there discs that could have more than one .sub file? If so, why would a disc have more than one?
Got you
The reason is that subcode-Q information contains both the "current M:S:F location within the track" and "current M:S:F location from start of disc"
Would there be any benefits of getting Main to parse ccd files?
It wouldn't make sense to do it by track under those circumstances
As far as I'm concerned, even if there were some sort of benefit, I wouldn't think it's worth the trouble
However, Mednafen handles them
Nice.
This work may be of interest to @young lion and @grand eagle if they haven't seen it
Did PCE CD support any of the even more niche flavours of the format? Off hand there was CD+EG and CD+G Midi
I've never heard of CD+G MIDI. I barely heard of CD+EG, and the one person I heard it from, was checking whether it was available on PC-FX, and I believe that it wasn't properly supported
Remember that the CD BIOS supporting all this was around 1989
And version 3.0 - the final version - was in 1991
It might be kind of fun (but an incredible amount of work) to take the "subcode" output at the back of some CD players (especially older ones), and feed that into a RP2040 microcontroller in order to visualize the CD+G data.
Never even heard of a subcode output, what sort of port was that? Just an RCA style?
I saw something somewhere today about CD+G Midi today, been trying to see where, and it was on CDi Discord
Near the end of the article, there is a section "Jack on CD players"
I know I had this jack on my Technics player
It is a shame it doesn't seem DVD/BD players play them. Maybe a licensing cost involved so nobody bothered.
It seems you've not only caught the Photo CD virus, you also catched the CD+G. CD+G is probably also interesting for CD-i. Before starting the DVC I've actually considered it but haven't found a single image online for testing. Maybe I should have invested more into the search.
Maybe someone can quietly point you in the right direction in a way that doesn't break server rules. I will say you can get some cheap, the Lou Read one I have is cool and can be got cheap on discogs and sellers in Germany
The Jimi Hendrix one had a few different CD+G pressings, if you do a search for CD+G on this page
That is postage to the UK, so would be much cheaper for you
It would be nice if a pack of them did appear in the usual place, I never got my hands on any images either, but I never dug too deep and didn't have much use for them in image form at the time
They exist
Agreed. You're best off to buy a few before they go extinct
They're not super-expensive like videogames
Oh wow⦠I wonder what CD+G stuff Valis III had?
Presumably these two are in no intro packs somewhere? If so could be found and tried on the core now
Only if the subchannel data was properly extracted?
Someone will need to dig into this one
Valis 3 cue file contains ISRC on some tracks, and ISRC goes into subchannel data
it's probably just that ?
That's referring to an audio CD, not the game CD. Perhaps a soundtrack or a "drama" CD
So uhhhh, this enables karaoke discs?
Yup, that's one way of putting it
Are you still in Japan ?
If so, you should be able to locate some CD+G here and there.
There seems to be some graphical flickering issues on āCadashā(USA) on the latest stable. A white bar will occasionally flash on the bottom left of the screen just above the health bar along with a few other random flickers as you play through the first cavern of the game
Not present on OG hardware
See if you can trace back through the past unstables, and try to find out when this started
Can confirm itās still there as far back as the 12/2/24 unstable
Are you able to go back further ?
The previous release was 2024/04, but the first updates to correct some sprite issues were around October as I recall
Iāll check
Still there as of the 4/24 stable
Let me know. I checked it against a YouTube video of real hardware
US or Japan ROM ?
And are you showing or hidig overscan ?
You say Cadash ?
I'm not seeing anything like that on the Japan ROM
Maybe you can take a video ?
US rom. You can see it start to happen just as you initially exit the castle and a few more times as I go through the cavern. (Forgive poor gameplay Iām filming with one hand and playing with the other)
Hiding overscan
I'm sorry, I'm staring at the health bar and I'm not seeing what you're describing....
Sorry itās very subtle I know. Look at the top of the skull head on on the left side. It stutters a few times and a thin white box appears a few times within the video across the status box
Let me try to screenshot the frame
Also, does it happen when you choose any character, or a specific one ?
I see it. Need to zoom-in. Itās at the ~16s mark.
With glitch
Without glitch
Hmm. OK I saw a flash, and I thought I might have seen a stutter where the box shows on the line below for a single frame... but I dunno... I don't think I can reproduce it
Thin white line above the status bar only visible on the left
Even if this is a difference between the core and a real machine, I don't know if it will be possible to understand the reason. It could be that the scroll register is hit one cycle too late on one frame, but the reason for that would be tricky to deteremine.
I think this is one where we might need to see the VDC decapped and decoded first.
If this was a more easily-reproducible test case, ti woudl be better
I was able to reproduce it a couple of times from game start. Picked the fighter and named them "Fish" if that makes any difference. But I can reproduce it literally standing still
Yes leaving the game paused for awhile it continues to occasionally flash the glitch
Hell yeah officially not crazy
SloooooowMoooooo
Woo hoo ! Fixed the remaining issues on CD+G.
Fetching CD+G rn!
needs unstable Main as well as unstable TG16 core
** ššš¬šš¢š§š šš”ššš **
Spreadsheet - [LINK]
(thanks to @grand moss for this invaluable community resource; reach out to him if youād like to contribute)
** šš¦š©šØš«ššš§š šš§ššØ **
You must source your own games and bios; DO NOT ask about them, DO NOT ask for them, and DO NOT ask where to get them.
This core also runs the following supported add-ons:
- TurboGrafx-CD / CD-ROM²
- Super CD-ROM²
- Arcade Card
PC Engine is the name of this console in Japan.
(please comment with your feedback on this core)
Does the latest TG16 core have your updates or its just the unstable one?
#news message
No, I committed stuff after that
** ššš¦š¢š§ššš« šØš§ šš®š š¬ **
If youāre seeing issues in games then do the following:
- Ensure you have the latest core by downloading it from the pins - [LINK]
- Double check you donāt have multiple versions of the core and delete older ones.
- Check your core config to make sure everything is set to default or simply delete your config and try again.
- Test against real hardware (if you can) to make sure itās not a bug with the original game/hardware.
- If youāre playing a fan-made patch, then check the original game first as patches may introduce new bugs.'
- Finally, look on GitHub to see if the issue was already reported - [LINK]
(the TurboGrafx-16 / PC Engine core is mature enough where most issues are already known and anything else is likely a bug with the original game/hardware or youāre doing something wrong)
I don't mind
Not a surprise as I believe the core was ported from the mister but I can confirm that the Cadash issue is also present on the analogue pocket core
Donāt know if itās on the analogue duo core
Even though you believe your Youtube video (showing that the issue isn't present on original hardware) was from original hardware, there is still a 50% chance that either:
a) It's from an emulator like Mednafen which doesn't simulate video scanline cycles accurately, or
b) even if it's real, YouTube might process the blip out
But I'll try to see if I can recreate it
All my hardware is Japanese though, and I'm not sure i fI have a cracked US image nearby
(Assuming I can recreate the issue here in the first place)
OK, I can see it - rarely, and not reproducibly - on the MiSTer core at my desk.
What do you mean when you say not reproducibly? Because itās pretty consistent for me throughout the opening area every time Iāve played it
To be reproducible does it need to be the result of a specific button sequence or action?
@slender bluff do we have a testing sheet for PC Engine/TG16?
David has one
To be reproducible, it shouldn't just "happen", it should be possible to "make it happen"
Gotcha
But with the core being live for years it isn't something we need for testing more widely
OK, I played to the end of the cavern, and I saw it occur maybe 3 hiccups on MiSTer, one had visible white, but the other two were a visible movement of the box
On real hardware, I saw one hiccup
I definitely donāt mean to come off pushy at all about this issue as itās obviously extremely minor. Iām just being anal about reporting everything I see
So I beleive that there could be a difference, but I'm not convinced that it's something material, or even that it can be acted on
Oh good so it is there in some form on og
The video I watched was completely without it
@grand moss can you link your testing sheet please
My guess is that the raster register is "missing the update window" by one cycle in some cases.
This could for one of two reasons:
- Maybe the window is off by one cycle, and shoudl be adjusted. -> This has a greater likelihood of braking other things. But if we can get a logic breakdown of the decapped 6270, it could be justified.
or - Memory access induce indeterminate delays, which might cause it to be one cycle late occasionally
(or a combination of both)
Either way, nothing is actionable right now
I appreciate you looking into it either way
I'm not sure if I still have a "testing" sheet. I have a full list of all games though, and I can make a copy of that
When I get home Iām going to test it on the TG-16 mini just for fun
Basically, each new revision to VDC invalidates the previous testing
What do we need a testing sheet for? Does the whole library need a retest?
I don't think so
I will personally retest the entire library in my cave
In that case, OK. But you will need real hardware
There are several title that "just do that crap"
@grand moss will get a DM every single line I see out of place we will be best buds
Is that what "best buds" means to you ?
Canāt think of a greater connection than testing 38 year old game hardware we are basically married
Eek
OK, I found the list. Nobody has touched it in over 2 years.
https://docs.google.com/spreadsheets/d/1EhBL4tBfoU8blGWYKpvbMrOUfXQBNP1ukyLfDlqyZiM/edit?usp=drive_link
Most of the testing was done on versions from 2021
If you would like to go through and retest and update the tested version, be my guest
Especially when the overscan is involved.
Yeah, people logged multiple issues against Loom. But the game really does those things.
@grand moss thanks again. I have pins up and ready. I made sure to credit you under Core Developer too.
I'm not "the" core developer, just "a" core developer. srg320 and soregelig have done more than me
Yeah well, if you're cool with it, I'd like to celebrate your continued support and effort.
Because you kick ass man
As long as it doesn't diminish srg320 or sorgelig
thanks for making me and countless others PC Engine fans too
Gotta find a Genesis/Megadrive/MegaCD person to add that XE-1AP support in. It should be easy, but those controller port bindings are too convoluted
X68000 could also benefit
(That one might actually be easier)
So featurewise we are done now? Congratulations to all the srgs, sorgs and Davids involved.
the ones and onlys
Srg320 might tackle one day the missing "cdd DMA timings" test on the everdrive testsuite which, hopefully, might make the attract screen of Patlabor last longer than 9 hours.
And save state support which might or might not be contributed by Paul NL
Other than that it seems pretty done now.
I guess that this is bonus. It's important that the core behaves just like the real thing.
Of course. I never lament lack of save states on FPGA cores but it would be cool if it did happen
Sure. It will come eventually.
I personally try to avoid to use save states. I rather get stuck in a game and will try again in a few weeks, months or years.
Using save states is like using cheats. It changes how the game is conceived.
I just use it to pick up where I left off in games without built in save systems. If I want to cheat I just use the cheats menu
Save states are also super useful for testing purposes
(thus I didn't KO Mike Tyson to this very day. And I am fine with that)
And practicing in Shmups which would be amazing to have on the PCE
Oh, there's always more - Print Booster, Illust Booster, Photo Reader, Tsuushin Booster, and so on. But yeah, for anything anyone ever heard of, I think it's done
Has anyone gone back to see how stable/unstable it is now ?
Nope
Probably a good idea to look at
Legendary Axe is such a joy to play. Probably the tightest controlling platformer on the system. Such a mistake not bundling it with the Turbo in the US
Do you expect that accuracy can still be improved by having some chips reverse engineered?
To bring it on furrtek levels?
At this point, I think the only ambiguities are in the 6270 video controller, and the CDROM controller chip
But I guess that nobody is currently looking at those.
There is a big bottleneck in the FPGA scene. With just a handful of people being able to decap and RE chips from die shots.
The common CPUs mostly done and many graphics chips still missing.
The die shots are there. And I see that a certain dshadoff is mentioned š
Yeah but nobody know how to read the standard cells
(different manufacturer than usual)
That's shame. It's like the IGS023 chip which is the greatest obstacle in having a PGM core (at all).
yeah. I bet that the 6260 is a much simpler die, and probably some of the standard cells can be reverse-engineered from that
Then those models could likely be applied to the 6270
I played through the beginning of Cadash on my CoreGrafx a couple of times to compare with MiSTer and the graphical glitches are exactly the same
So I think that behavior is accurate
nice
Still missing LD-ROM² support š
But yes, grats to the amazing devs!
That's a good question, David probably best placed to answer. Maybe some unreleased peripherals, but can't think of anything offhand the core can't do now and it must be about bug free now.
Thank you so much!
Uh, we're missing images....
There will always be very-small-run peripherals and consoles to add. For example, Print Booster, Illust Booster, Photo Reader, Tsushin Booster, GPS navi, LD-ROM... there are probably more, but not much
Completed Dracula X today....
This Unlimited health cheat glitch fix works perfect. No more Whip glitches...
File needs just to be added in the cheat folder for this game and also in the
Gamefolder itself.
But the Boomerang Cross do have some weird green bars left and right ( hitbox )
while a cheat is activated.
Can this be fixed too please ?
I don't know who would do work on cheats; that's one thing I don't work on
I would suggest that you inquire from the person/source you got it from
yeah thats what i did a few minutes ago. I wrote the guy who fixed the health glitch fix.
Additionally i add a ticket on GitHub
The one against the TG16 core ? It's not a core issue, that should be closed.
Ah i see ok .. well than lets wait for some answers from the guy who did the other fix. Thx
but this is an issue .... Walterfall downwards with Maria. After the Boat Man....seems the Background Texture missing
what level is that? I can take a look
Is this one the same save file that you used the cheats on?
nope No cheats
there you Go...
PC Engine does not do "textures".
If there is an issue in an unmodified game, it will need to be confirmed
In order to reproduce and validate against real hardware, steps to reproduction need to be provided.
Developers are not expert players, and while developers know the core, they don't know the games themselves to the level that the players do
I need to unlock that path. give me a moment to look
But I realized this is the english patch. Let me move my save over for the normal version
Appears to be an issue on the japanese version too. Redump bin/cue
this save has the mountain level unlocked
I wonder how far back this issue goes as with the popularity of this game I would be surprised if it wasnāt caught by now if itās existed on the core for awhile.
yeah, also, I guess I cannot be 100% sure that this save didn't use cheats at some point because I do have a rondo cheat in my cheat directly, apparently
I know that several people played through the whole game way back when... but there have been (somewhat) recent updates to the VDP which might potentially have affected it. More affecting sprites though, so I'm hesitant
I donāt think this is a bug. It looks odd, but Iām fairly certain I saw this when I 100%ed Rondo a few years ago on my Coregrafx
Green background and all
Iāll double check though
Thank you
so it is what it is.... šš
Great thanx so much for research. The games so great and i just wanna check.
Thanx guys
Damn someone should fix the green screen. Immersion breaking smh.
I wonder if itās actually a bug š
As in, a bug in the game
might be! it definitely looks off. I only remembered it from years ago because when I played it I thought "that looks broken, if this weren't a real Coregrafx I would suspect an emulation error"
but I also checked youtube and other people's playthroughs also seem to have it
One of the issues with patches and cheats is that they can introduce unintended side effects that the original game did not exhibit. Usually when seeing those issues itās best to first check against the same patch/cheat on actual hardware or make sure the original unmodified game doesnāt exhibit those behaviors.
yeah, I wasn't running any cheats, but I wasn't sure if the save I was using was cheat free in the past
it may have spent time on the wrong side of town
Yeah thatās an interesting problem.
Some games have bugs, but more often they leave in "hiccups" as opposed to real bugs
Exactly. There are two reasons why I will reject issues based on cheats:
- A lot of cheats are built and tested on emulators which are inferior, and are dependent on running on that emulator, and won't work on a more accurate emulator.
- There is no "cheat" device on original hardware that is trying to be simulated - this is an artifact of emulation technology. Therefore the same behaviour can't be confirmed on original hardware (by definition). It's only there as a nicety for end users, it's not an integral part of the core.
Damn, well said
The same issue in 1) can exist for patches such as translation patches. If there's an issue, 99% of the time, it's a patch that won't run on real hardware - but at least this can be tested if somebody burns a disc.
I was under the impression that the cheats on mister were based on action replay/gameshark/game genie codes used on original hardware
There was no such hardware for PC Engine/TurboGrafx
And I'd be surprised if there was any such device for disc games on any system
3DO had a neat cheat device called Game Guru that let you edit save game files to insert cheats into
It's also true that a lot of the original translations and/or ROM hacks for NES and other systems in the 90's were built for emulators and didn't work on actual ROMs on the original hardware... when this was found, some people complained to the ROM device maker, not understanding that it was a bad assumption made during the development of the hack on emulator
This doesn't happen as much anymore, but it's genuinely a "THING"
Ps1 definitely had GameShark. It worked through the memory card slot
OK. But as I said, there was no such system on TG16/PCE, so it was based on assumptions on the emulator on which it was developed (and as for how well it was tested, I simply don't know)
In any case, consider this core to be "hardware" that we try to correct, not software.
It's pretty close as it is
I gotcha. Wasnāt arguing btw about the cheats being based on those devices. Just saying what my initial impression was that was obviously wrong
Ps1 definitely had GameShark though š
S'ok. I know the PCE hardware, and I don't know anything about the cheats driver that somebody added to the core, or the cheats that exist. Simply outside of my field of knowledge (and interest)
I was going to offer to remove the cheats engine. š
Weād no longer be best buds
what have i started š¤£
Cheaters never prosper, remove all cheats so the weak arenāt tempted
All kidding aside, I have a work around: watch a video of someone whoās good beat the game and fantasize about being that skillful and handsome
The green screen is also on any version even without using a cheat :/
And its on the converted Version for PS4 too... So yeah its seems like this part of the game
was not completed or whatever. Maybe someone will edit the game but its definitely not a
problem of the core.
https://youtu.be/3EfMTQEm_mU?t=2439
Ok, I've added text commands to give you a list of shortcuts to all the pinned posts in a channel. Type "#channel name pins" and you'll get it. I covered the most popular channels.
For Example: #1073797664195494028 pins
NEC PC Engine:
Robby has reached the pinnacle of his powers.
Or maybe this is the pinultimate step.
OK, somebody should be retesting the CD hang issue
This is the history:
https://github.com/MiSTer-devel/TurboGrafx16_MiSTer/issues/220
I believe that it has either been completely fixed, or almost completely fixed
We should come to a consensus though.
could be handy for CGD
whatās that
Ohhh ok, sorry Iām ignorant baby
I just showed up to this Discord and know nothing about the history
its a quieter server now, MiSTer had outgrown it
I stayed a mod there, so tweak it from time to time. Its still considered a place to get help with Analogue consoles and other accesories like FXPack
thereās this Carl bot thing, no real clue on what itās about but itās a bot that letās me run commands. One of them does replies when you specify a word. Which was the shenanigans I was doing last night and this morning.
Called triggers
The channel name + pins thing was me just coming up with a consistent method of making an easy to remember way to trigger a specific response
It can be any word you want, but I did it that way so itās very hard to accidentally trigger a response outside of purposefully doing it
yep good idea. does it have already a function to check pins or was that html with links?
For testing that CD function, either is fine. The unstable only contains CD+G code which is not related
Ok, I will let it run overnight on a couple of misters
Even if it does fail, it would be good to understand how much less frequently it happens compared to before
(if possible)
No that was just me copying and pasting a custom response per trigger lol. Extremely basic and low tech solution haha.
Ya basic
keep it stupid simple, good guideline š
makes sense
I am not sure what the previous frequency was, but last night 3 of the 6 misters running patlabor crashed at similar points in the intro. 2 De10s and 1 Mister Pi.
Running them again to see if I can get a better idea of timing.
Also not sure if this is helpful and if there is anything I could test that would be helpful
How long was your test period (for 3 of 6 to fail) ?
~8 hours
Previously, the average time to failure was less than 60 minutes on one unit
So it seems to be 4 hours now
yeah, I definitely didn't see any that fast (as I went down and looked at them before turning in...around an hour after I started them)
Would be nice to run another session, checking in on them every 30 to 60 minutes though
I have one open in front of me now
The PC Engine core has this issue with direct video where the picture is offset to the right.
No idea if it's the core itself, or if it's a bug with the Morph 4K I'm using. Does anyone else has this issue?
Has been a problem for a long time, so it's not related to any particular build.
What makes you think that it's an issue compared to the original hardware ?
yes, it's off center - but I believe that's what it is on real hardware?