Altirra 3.90 released

[Mazzspeed]

I also runs Under TwisterOS on the Pi400 using Wine. I can't test the ARM variant as I only have the 32bit kernel installed.

 

[gorgh]

Hi there,
my prod "Alive" seems to give different results in Altirra and my 130 XE PAL, it's the matter of different sounds (some notes play differently)
https://www.pouet.net/prod.php?which=89436
It can be fault of my monitor but I wanted to signalize that

Edited August 11, 2021 by gorgh

[phaeron]

12 hours ago, gorgh said:

Hi there,
my prod "Alive" seems to give different results in Altirra and my 130 XE PAL, it's the matter of different sounds (some notes play differently)
https://www.pouet.net/prod.php?which=89436
It can be fault of my monitor but I wanted to signalize that

Do you have an idea of which distortions sound different? Noise-based ones are the ones that are most likely to be sensitive to monitors, esp. if you have one that has bad aliasing. But that's fairly rare on modern monitors.

[gorgh]

I will check that and respond to your question phaeron

[baktra]

Hello @phaeron

 

I have a question on the emulated H: device.

It restricts file names to adhere to the 8.3 standard. Anything else gets the 165 error (bad file name).

Is it intended restriction (in ATHostDeviceParseFilename()) or an unfortunate omission. I am asking because e.g. atari800 has no such restriction.

 

Thank you

[Synthpopalooza]

This piece of music seems to work ok on an older Altirra, but on the latest, 4.00 test 37, it misses notes.  Can you confirm?

 

On first two channels we are using two tone mode at 1.79 khz, plus a 15khz clock.

 

 

exo-title4.xex

[VinsCool]

35 minutes ago, Synthpopalooza said:

This piece of music seems to work ok on an older Altirra, but on the latest, 4.00 test 37, it misses notes.  Can you confirm?

 

On first two channels we are using two tone mode at 1.79 khz, plus a 15khz clock.

 

 

exo-title4.xex 4.61 kB · 2 downloads

I would like to add that I found the last version of Altirra that did play this correctly was 3.20, after testing out some versions in decreasing order.
Altirra 3.90 seems to have introduced the anomaly.

Recordings in the .zip below were made from Altirra 3.20, 4.0 TEST37, and finally my NTSC 800xl  

Exo Title4 Comparison.zip

[oo7]

@phaeron

 

feature request. Allow option for fullscreen to be exclusive with a user selected resolution & refresh rate.

 

This can help on odd systems like my media pc or my portable 2 in one where lets say user runs very high or odd resolution and would like altirra to run in a diff or lower resolution full screen while also being in exclusive mode rather than borderless window.

 

 

this can even free up some horse power.

 

I know I might sound crazy but it helps to free that horse power when running on battery on a 2 in 1 convertable and the power gain on exclusive full screen is nice.

 

 

can also be nice on htpc on 75" screen where upscaling to 4k is not really required.

 

 

 

[pseudografx]

That is already implemented:

[oo7]

1 hour ago, pseudografx said:

That is already implemented:

Oh wow. Im blind and stupid lol. Use this thing daily and somehow missed that. Im losing it

[oo7]

@phaeron

 

setting fullscreen to 640 x480 or any other 4:3 mode results in 16:9 stretch.

 

other emulators such as nestopia ue 1.51 do not suffer from this

[ggn]

2 hours ago, oo7 said:

@phaeron

 

setting fullscreen to 640 x480 or any other 4:3 mode results in 16:9 stretch.

 

other emulators such as nestopia ue 1.51 do not suffer from this

That sounds like a monitor issue, not a software issue. For example, my monitor has settings for "1:1" aspect ratio, as well as "stretch". Setting it to "1:1" gives black borders left and right of the image but no stretch.

 

Are you sure that the other emulators you mention don't use "fullscreen borderless" (check the Altirra screenshot above) and perform some clever bordering, thus avoiding the issue?

Edited August 18, 2021 by ggn

[Keatah]

It could be both. But on my setup:

I can make it "fail" by doing, View> Stretch Mode> Fit to Window

Or "fix" it back to 4:3 by doing, View> Stretch Mode> Preserve Aspect Ratio

 

Additionally, between View> Stretch Mode and View Overscan Mode, you should be able to handle just about any situation.

 

NB, if you're still having a problem.. I have Tools> Options> Display> set to Direct3D 11 and Borderless mode ("windowed fullscreen"). Not using a custom mode. So start like that and see what happens!

Edited August 18, 2021 by Keatah

[oo7]

17 minutes ago, Keatah said:

It could be both. But on my setup:

I can make it "fail" by doing, View> Stretch Mode> Fit to Window

Or "fix" it back to 4:3 by doing, View> Stretch Mode> Preserve Aspect Ratio

 

Additionally, between View> Stretch Mode and View Overscan Mode, you should be able to handle just about any situation.

 

NB, if you're still having a problem.. I have Tools> Options> Display> set to Direct3D 11 and Borderless mode ("windowed fullscreen"). Not using a custom mode. So start like that and see what happens!

Will try a bit and report back. Same system using snes9x and nestopia ui and custome exclusive full screen modes no issue so i dont blame the setup. Before I say more I will do more testing. Will test on my laptop with nvidia, my gaming pc with nvidia and my 2 in 1 with Radeon. Also I would like to note I understand drivers and aspect ratios. That being said i noticed this only on my 2 in 1 with the radeon which isnt the best card when it comes to scaling etc.

 

 

[oo7]

So my goal was to run 640x480 60hz exclusive full screen. And this is where i was having a problem. I had set altirra to dx11… before trying other recommendations which would work but would take me out of the exclusive resolution i wanted i dropped to dx9 mode. This fixed it. Also explains why the other apps work, they are running in dx9

 

i now think the wrong behaviour i observed is either a issue directly with dx11 or with dx11 and my radeon or dx11 and altirra. 
 

will test other combos on other and report back.

 

 

[oo7]

17 hours ago, VinsCool said:

I would like to add that I found the last version of Altirra that did play this correctly was 3.20, after testing out some versions in decreasing order.
Altirra 3.90 seems to have introduced the anomaly.

Recordings in the .zip below were made from Altirra 3.20, 4.0 TEST37, and finally my NTSC 800xl  

Exo Title4 Comparison.zip 17.83 MB · 5 downloads

So is there an actual sound regression after 3.20?

[Keatah]

Sometimes learning where the scaling happens, in the graphics card, the monitor, or the emulator can be tedious. Especially if more than one of those doing something.

 

Personally I like it when the "emulator" commands the graphics card to operate in a native-to-monitor resolution and the emu itself does the scaling. So here we're rendering many modern & tiny pixels to make up one vintage pixel.

 

No mode switching takes place, the GPU is always outputting a monitor native signal (1920x1200, 3840x2160, or 7680x4320). And today's CPUs & GPUs can certainly handle this while falling asleep of boredom.

 

If I figured it right, it takes 540 8K-pixels to draw 1 single star in Star Raiders.

[Keatah]

8 minutes ago, oo7 said:

will test other combos on other and report back.

Yes. Don't forget the View> Stretch Mode> to fine-tune your aspect ratio and get the stretch just right.

[Keatah]

1 hour ago, oo7 said:

Also I would like to note I understand drivers and aspect ratios.

I don't doubt it. I'm simply so used to always explaining this for newcomers to the emulation scene. If it's one thing I demand of emulators it's getting the aspect ratios and sizings perfect to a "T". And colors too. But that's discussion for a different juncture.

 

1 hour ago, oo7 said:

That being said i noticed this only on my 2 in 1 with the radeon which isnt the best card when it comes to scaling etc.

Never had a Radeon card. I don't have many issues with Nvidia. Or at least none that I couldn't resolve.

 

I would assume there'll come a time (if not already) that all emulators will have to properly scale internally since 4K is pretty much standard and 8K awaiting price drops. Modern resolutions are huge compared to a 41-year old Atari.

[Mclaneinc]

My machine came with a Radeon, I swapped it out quite quickly for an Nvidia, nothing but hassle with it and its drivers. While the Geforce isn't fault free driver wise, it's a lot better than the other card.

[oo7]

So went back tested this on 4 machines, 2 with nvidia gpu, one with intel igpu and one with a radeon all the same results. It seems in dx11 mode with low resolutions they all have undesirable results but i think this is a dx11 issue having found the same results with other apps. While in dx9 all configs work as expected. So i think there is no bug.

 

on a side note the only reason i went radeon in my 2 in 1 is nvidia doesnt make a competitive option in that form factor. My radeon machine is a 11 inch laptop that converts into a tablet and i get approx 10 hours running altirra under windows 10 on battery with full screen brightness and wireless gamepads, this system uses emmc for storage is passive cooled and sealed with no mechanical parts. Basically its my atari xem lol. Only thing its missing is a skin which im having made to wrap it in. Ill take some pics at some point

 

anyway nothing wrong with altirra i should of just tested more

[oo7]

I am very curious if altirra after 3.20 is considered to have an audio emulation regression.

 

I currently run 3.90 stable daily.

 

I have actually encountered no issues whatsoever. (but am wondering is 3.20 is considered more stable or not)

 

 

I use it to game and to program in basic.

 

I save my basic programs to a folder using drive h:

 

 

 

 

[VinsCool]

8 hours ago, oo7 said:

I am very curious if altirra after 3.20 is considered to have an audio emulation regression.

I don't want to claim there really is a regression, because everything else has been working perfectly as far as I could tell.
That binary used some obscure tricks to produce these sounds, an improvement from an earlier Altirra version could possibly have introduced a bug with the Two-Tone filter emulation, and remained unnoticed until now  

[oo7]

1 hour ago, VinsCool said:

I don't want to claim there really is a regression, because everything else has been working perfectly as far as I could tell.
That binary used some obscure tricks to produce these sounds, an improvement from an earlier Altirra version could possibly have introduced a bug with the Two-Tone filter emulation, and remained unnoticed until now  

hmmm so its a polite regression lol.  

 

regressions are bound to happen. even to the best coders.

 

And not kissing ass here but this coder clearly knows what he is doing. this emulator requires far less power to do what it does than expected.

[Synthpopalooza]

Yes, to my knowledge, I don't think anyone has touched this particular setting before.  It uses 1.79 mhz Distortion A with two-tone mode, which can be paired atop either the standard 64khz clock or 15khz.  Two waveforms are available, the square wave is very ringy, and very close to the famous CSAVE tone (not the boop boop, but that high pitched whine).  The other is a saw wave which is transposed up an octave.  Certain frequencies must be placed in the first channel with each standard distortion A note in the second channel for it to play properly.  Using 15 khz mode roughly puts the output down by almost 2 octaves, basically the same as in standard A distortion.

 

The issue is that certain high notes in the range are not getting played in the new Altirra, or in PokeyMax, an ultrasonic pitched beep is being played instead.  In the EXO Area 3 music, which uses both the CSAVE bell and the saw wave, the CSAVE bell plays just fine, but the saw wave drops out on the high notes.