Altirra 2.80 released

[fujidude]

Hello Geoman!

[phaeron]

Update:
http://www.virtualdub.org/beta/Altirra-2.90-test22.zip
http://www.virtualdub.org/beta/Altirra-2.90-test22-src.zip

Significant rework on the NTSC color algorithms. After trying to do a vaguely representative color chart and a bunch of investigation I've come to the conclusion that it simply isn't possible to match classical Atari color palettes to an sRGB display with a standard YIQ matrix, because the color balance on the primaries is pretty far off and the standard adjustment procedures don't match to the current color interpretation of the colorburst. Trying to do it pretty much always gives at least one major color wrong, whether it be the colorburst, the blue OS screen, or color 15. The color settings dialog now has options to move the locations in YIQ space of the RGB primaries. In particular, this allows reducing green and boosting blue, which better matches what I'm seeing on a 1702 monitor. The defaults have been redone to a "worst of the modern era" setting, which is standard matrix and colorburst, but overrunning the hue circle a bit on the color adjustment, which seems to be pretty common. Also added is a new preset based on the Commodore 1702 monitor that adjusts hue 1 to be more orange -- the color adjustment has to be moved pretty far to hit standard colorburst line -- and has the color primaries adjusted. This particularly boosts dark blues, which seems to improve a lot of games that were previously dark.

The NTSC high algorithm has also been changed to better reject luma in artifacted areas, giving more solid colors and less striping. Artifact saturation now affects both NTSC and NTSC-Hi modes, and in the latter it controls how strong the chroma signal is (which varies between models). The artifact brightness setting has been removed as it only affected NTSC regular artifacting mode and was physically bogus for any setting other than 0. The "sharp bilinear" filter mode now also suppresses horizontal sharpening with the high-filters are enabled.

 

Choplifter 5200 compatibility entry has been fixed.

 

Fixed an issue in AltirraOS where it would do a type 3 poll before running a cartridge, which caused Pole Position and Missile Command to miss some sounds (SKCTL issue).

 

Fixed a couple of speed-switching issues in the SX212 emulation and handler.

 

BlackBox and MIO emulations now support multiple SCSI drives. The first drive is ID 0 LUN 0, the second is ID 1 LUN 0, etc. There is now also info in the help file about how to set them up.

 

I normally start Altirra in fullscreen with /f but i noticed that when i load a .cas file Altirra reverts to window mode. With .car and .atx files remains in fullscreen mode. Am i missing something?

 

This happens because of the progress dialog that pops up, which is a native-only dialog and thus forces the exit. One way around this is to use borderless mode instead (/borderless).

[Triads]

Much appreciated, as always

[Mclaneinc]

Thanks Avery, the colour issue is a tough one as you and other have mentioned many times that pretty much no two machines displayed the same colours but its appreciated that you are aiming to get what should be as near as to what was seen on a low res monitor like most of us used.

 

I know you are not in PAL land but could the same ideal be done for PAL or is there no reason to do it?

 

Paul.

[Keatah]

Great to see improvements in the NTSC color rendering. We'll get those Star Raiders colors accurate in no time!

[phaeron]

Thanks Avery, the colour issue is a tough one as you and other have mentioned many times that pretty much no two machines displayed the same colours but its appreciated that you are aiming to get what should be as near as to what was seen on a low res monitor like most of us used.

 

I know you are not in PAL land but could the same ideal be done for PAL or is there no reason to do it?

 

Well, the most basic problem there is that I don't have a PAL setup to check for any such issues. The PAL GTIA also has gaps in its color wheel and a much stronger preference for a "correct" setting; if it's misadjusted you'll see a drop in saturation or striping in the output due to excessive mismatch between the even and odd chroma lines.

 

The basic issue here is that there are many ways that a TV can deviate from "standard" and without some equipment it's hard to tell exactly how. The phosphors might be a bit off in color from ideal, the circuit might decode color in an interesting way (YIQ vs. YUV vs. direct RGB decoding). There are two main color standards for NTSC, and that's assuming the TV even follows it -- it's a common pastime for TV makers to deviate from standard to make the output look "better." That means short of getting a color calibration device like a Spyder to measure the color output on a reference picture, we're just eyeballing it or going off memory, which is really imprecise. And honestly, I'm not sure how a precision device like a Spyder is going to handle getting slapped onto a 30 year old CRT monitor hooked up over composite. The one thing I do know is that just trying to tweak colors based on peoples' anecdotes is getting us nowhere and just leads to breaking some games in order to fix others. Throw in games with different versions that have had their colors adjusted for PAL and it's just confusion all around.

 

That leads back to the question of what palette the emulator should actually ship with as default. The 1702 clearly deviates enough that there's a need for at least two settings, and the altered color matrix helps enough that I'm considering altering it in the default. That profile is adjusted a bit too far from ideal, though, so the color matrix adjustments would have to be toned down and the color range shifted back closer to ideal for a default. -57d start / 25.7d step showed some promise with an adjusted matrix, but I didn't have time to verify it. Ideally, I'd like to have two NTSC profiles in the emulator: how an average one looks now, and how it looked back then.

 

Another puzzle I'd like to eventually solve is how the 800 gets its weird cyan/green artifacting. The colors are so close together on the hue wheel that I can't emulate it using normal artifacting algorithms and haven't been able to come up with a plausible explanation for what kind of signal distortion could cause it.

[atx4us]

Hello all! My first post here and i'd like to wish you all a Happy New Year! I am an Atari 800 user (unfortunately i don't own it anymore), who spent countless hours in his youth with Fort Apocalypse, Shamus, One on One, Jumpman, Necromancer, Canyon Climber, Baja Buggies, ...

 

I am using Altirra 2.90-test21 with Launchbox (https://www.launchbox-app.com/) which allows me to pass different command line parameters to Altirra for each game, making running games with different requirements easy.

 

I normally start Altirra in fullscreen with /f but i noticed that when i load a .cas file Altirra reverts to window mode. With .car and .atx files remains in fullscreen mode. Am i missing something?

 

 

 

Welcome aboard GeoMan! You found the best Atari User Group on the internet and one of the best Forum threads (Altirra) for the Atari 800

[GeoMan]

This happens because of the progress dialog that pops up, which is a native-only dialog and thus forces the exit. One way around this is to use borderless mode instead (/borderless).

 

 

 

Welcome aboard GeoMan! You found the best Atari User Group on the internet and one of the best Forum threads (Altirra) for the Atari 800

 

Thanks! and Thanks!

[Mclaneinc]

Thanks for the explanation Avery, I was aware you did not have a pal setup hence the tentativeness of the question, what I've done on a few occasions is search out a few different screen shots from UK mags like Atari User and adjust until it sort of middles with the screens and leave it at that. Obviously print both fades over time and bleaches in sun so its a sort of best guess, also printing itself does not guarantee correct colours

 

The point is I'm happy with what I have and if there's no hardware to test via then any thing else will be a hack and I know you don't like hacks if avoidable.

 

As for the artifacting problem, wow, you stumped, can't be true

 

I'm sure at some point info or whatever will give the answer but its a credit to you that you want to affect emulation to such a level.....Thank you.

 

Paul.

[Roydea6]

I am wondering if Altirra_0s.rom has a place to inhibit the OPTION key needed to disable basic, like the Atari_Os.rom has... For us users that add Altirra OS to one of our Bios OS slots in the Ultimate 1Mb or the 4Mb CPU upgrade (Antonia). and also now the UNO_CART.

Edited January 6, 2017 by rdea6

[+Larry]

I've been using 2.70 release version successfully for quite awhile. And several times, I've tried the 2.80 versions and now the 2.90 versions. From 2.80 test 1 on, none of these that I've tried work with my XP computer. However, testing the same ones with Win7, they work fine. Both 32 and 64-bit versions work fine with my Win7-64 OS. When I launch Altirra with XP, it goes to the normal balck screen momentarily, then goes to a gray screen and freezes. Not doing anything tricky -- just trying to boot to a MyDos 4.55 disk image.

 

My Xp system is XP (32-bit) with SP3

Core2 Duo E6420

All-in-Wonder 9000 video card.

 

The Win7 system is

AMD Athlon II (dual core) 2.70 GHz

Intel onboard video (on a ASUS mobo)

 

I originally was thinking it must be an XP issue, but reading a few entries, maybe it's a video card issue?

 

Thanks,

Larry

[+playsoft]

I noticed a minor issue when running Snow Plow - little flickering lines to the right of the display. The program reads the P0PL and P1PL collision detection registers quite frequently and I found if I removed them the flickering stopped.

 

I put together the following test program. It displays lines of ANTIC 2/4/5/F/E/D, waits for you to press a key and goes into a tight loop reading those registers. ANTIC 2 and F are fine but the other modes appear to be showing data from the next byte.

 

test.xex

test.s

[Mclaneinc]

Yes see it, its in the overscan area, set to OS screen only its gone

[+MrFish]

I'm not seeing anything different when hitting a key in any overscan mode. I get this no matter what:

 

 

@ Mclanince: "OS Screen" is not the the normal view mode; "Normal" is.

 

[+playsoft]

This is what I get:

 

 

And someone else confirmed the flickering lines on Snow Plow, so I don't think it's just me.

[+MrFish]

This is what I get:

 

altirra2.jpg

 

And someone else confirmed the flickering lines on Snow Plow, so I don't think it's just me.

 

I noticed you're using v2.90 t22. I'm using v2.81 release.

[serj]

I don't see any lines.

[+MrFish]

I'm not seeing anything different when hitting a key in any overscan mode. I get this no matter what:

 

modetest.png

 

To Clarify my config, I'm running:

 

Altirra v2.81 Release

 

Hardware - 65XE/130XE

Firmware - XL/XE v3

Memory - 1088K

Video - NTSC

Edited January 6, 2017 by MrFish

[+playsoft]

 

I noticed you're using v2.90 t22. I'm using v2.81 release.

 

I was running 2.80 when I first noticed the issue. I then updated to the latest 2.90 to see if it still occurred.

 

I am running under Windows 10.

 

I have been making my way back through the old versions and 1.6..1.9 are all fine, it is 2.0 where this starts happening.

 

I will dig out my old Windows 7 machine, see what happens there.

[+MrFish]

I am running under Windows 10.

 

I will dig out my old Windows 7 machine, see what happens there.

 

Yeah, that might be the difference. Serj and I are both running Windows 7.

 

I don't know what Mclaneinc is running.

Edited January 6, 2017 by MrFish

[+playsoft]

Running under Windows 10:

 

1.9 is 32-bit only and does not have the issue.

2.0 is 32-bit only and has the issue.

2.1 is 32-bit and 64-bit. The 32-bit version has the issue, the 64-bit version does not.

 

I imagine this will be true for all subsequent versions as with 2.90-test22 the 32-bit version has the problem, the 64-bit version does not.

[+MrFish]

Running under Windows 10:

 

1.9 is 32-bit only and does not have the issue.

2.0 is 32-bit only and has the issue.

2.1 is 32-bit and 64-bit. The 32-bit version has the issue, the 64-bit version does not.

 

I imagine this will be true for all subsequent versions as with 2.90-test22 the 32-bit version has the problem, the 64-bit version does not.

 

OK. I didn't notice you were running 32-bit there.

[+playsoft]

My Windows 7 machine has 2.70 installed and that gives the same results; the 32-bit version has the issue, the 64-bit version is fine.

[+MrFish]

Yeah, I just tried v2.81 32-bit on my Mac running Windows 7 32-bit and the problem shows up there.

 

Phaeron will soon speak.

Edited January 6, 2017 by MrFish

[Kyle22]

I've been using 2.70 release version successfully for quite awhile. And several times, I've tried the 2.80 versions and now the 2.90 versions. From 2.80 test 1 on, none of these that I've tried work with my XP computer. However, testing the same ones with Win7, they work fine. Both 32 and 64-bit versions work fine with my Win7-64 OS. When I launch Altirra with XP, it goes to the normal balck screen momentarily, then goes to a gray screen and freezes. Not doing anything tricky -- just trying to boot to a MyDos 4.55 disk image.

 

My Xp system is XP (32-bit) with SP3

Core2 Duo E6420

All-in-Wonder 9000 video card.

 

The Win7 system is

AMD Athlon II (dual core) 2.70 GHz

Intel onboard video (on a ASUS mobo)

 

I originally was thinking it must be an XP issue, but reading a few entries, maybe it's a video card issue?

 

Thanks,

Larry

The 32 bit version has always worked for me on all of my machines. I would agree that it may be a video card or driver issue.