Falcosoft wrote on 2026-08-10, 11:08:

Would you be so kind to post here the exact model of your 2 laptops that worked perfectly? Then I could add them as tested exact models to the Readme.

Yes, of course. The Lenovo laptop is a Lenovo IdeaPad Yoga 13, model number 20175, with i5-3427u processor, 4gb RAM (DDR3) and Intel HD 4000 graphics onboard GPU and 13.3" touchpad display.

The HP laptop is an HP EliteBook Folio 9470m with i5-3427u processor, 14" internal display, Intel HD 4000 graphics, 4gb DDR 3.

zyzzle wrote on 2026-08-11, 03:27:

Falcosoft wrote on 2026-08-10, 11:08:

Would you be so kind to post here the exact model of your 2 laptops that worked perfectly? Then I could add them as tested exact models to the Readme.

Yes, of course. The Lenovo laptop is a Lenovo IdeaPad Yoga 13, model number 20175, with i5-3427u processor, 4gb RAM (DDR3) and Intel HD 4000 graphics onboard GPU and 13.3" touchpad display.

The HP laptop is an HP EliteBook Folio 9470m with i5-3427u processor, 14" internal display, Intel HD 4000 graphics, 4gb DDR 3.

Thanks, one more question that is relevant in the context of ASPECT/ASPECTD. When you run ASPECT/ASPECTD what values are printed to the DOS console as detetcted native and corrected display window? The screens of both models are 900p or at least one has a different resolution?

The Lenovo laptop is 1600x900, but the HP laptop is only 1366x768:

zyzzle wrote on 2026-08-11, 20:38:

The Lenovo laptop is 1600x900, but the HP laptop is only 1366x768:

Thanks, it seems it gives a perfect 1024x768 aspect ratio corrected window so no rounding errors at all. And the 1024 pixel wide virtual screen can be perfectly centered on 1366 with an offset of 171.

I have another happy coincidence. The ASPECT.COM also works for me with a Broadwell chip and Intel HD-5500 graphics. Is it meant to be so? However, the backlight brightness program BLCSET does not work (which makes sense since I do not have the DSDT data). Incidently, the X-VESA program appears to be able to read this data in DOS via pressing F9 in that program, but there seems no way to write it to disk.

This is the Broadwell laptop which works with 4:3 aspect ratio. It is a Dell Inspiron E5550 with 1920x1080 native display:

zyzzle wrote on Yesterday, 07:11:

I have another happy coincidence. The ASPECT.COM also works for me with a Broadwell chip and Intel HD-5500 graphics. Is it meant to be so? However, the backlight brightness program BLCSET does not work (which makes sense since I do not have the DSDT data). Incidently, the X-VESA program appears to be able to read this data in DOS via pressing F9 in that program, but there seems no way to write it to disk.

This is the Broadwell laptop which works with 4:3 aspect ratio. It is a Dell Inspiron E5550 with 1920x1080 native display:

Thanks for your report!
Hmm. Interesting. No, it's not meant to be. This result is somewhat unexpected and may imply that ASPECT/ASPECTD can also work on Haswell generation laptops. But it's hard to tell without explicit test results.

In the meantime a new version of ASPECT/ASPECTD has been made. In order to be in feature parity with the Windows drivers it is now capable to produce pixel perfect centered images.
Two new parameters added:
/C -> display pixel perfect centered image
/A -> display aspect ratio corrected image (this is the working mode of the previous versions).

These are also run-time arguments so the display mode can be changed while the TSR is running. Without parameters it starts in aspect ratio corrected mode the same way as the previous versions.

Duke Nukem 3D in pixel perfect 1024x768 mode (1600x900 HD+ LCD):

zyzzle wrote on Yesterday, 07:11:

Incidently, the X-VESA program appears to be able to read this data in DOS via pressing F9 in that program, but there seems no way to write it to disk.

According X-VESA.TXT it's possible to press S to save EDID file. Could this be used for my Intel 945GM?

No, the EDID contains information about the monitor's firmware. The DSDT is about the laptop's firmware.

How much control over the aspect ratio settings is possible? Can fine grained control over how certain resolutions get scaled be allowed in a configuration file or something? Is there a way to control which scaling algorithm is used? (bilinear, nearest neighbor, etc.)

I'm also wondering how this thing actually works. It's a TSR and hooks Int 10h. Does that mean if you set up some nonstandard video mode without using 10h that it won't work? I'm assuming that's the case.

Maybe someone can look at the assembly and see what the bot actually does and where it got its information from. Presumably there is public documentation that explains how to do this, that nobody has yet delved into because it was too obscure of a problem and no one has yet tried to solve it. Would be nice to see that documentation to gain an actual understanding of what's going on here.

What if you try to run Windows 3.1? Or even 9x? Could this aspect ratio and scaling control functionality be built into the generic VESA drivers for Win31 and 9x?

ntalaec wrote on Yesterday, 09:53:

According X-VESA.TXT it's possible to press S to save EDID file. Could this be used for my Intel 945GM?

For intel 945GM onboard graphics, you may try the attached INTEL.EXE file. I compiled this over 15 years ago and it works on my Intel Atom netbook (Asus with 1.6 Mhz Intel Atom 230 CPU). This file will change aspect ratio to 4:3 in graphics modes (VESA, VGA), but not in text mode. It consumes a lot of low DOS memory, though. Try it and report if it worked on your system.

Oh here's another question. I wonder if there's a way to enable the overscan background color.

keenmaster486 wrote on Yesterday, 19:48:

How much control over the aspect ratio settings is possible? Can fine grained control over how certain resolutions get scaled be […]
Show full quote

How much control over the aspect ratio settings is possible? Can fine grained control over how certain resolutions get scaled be allowed in a configuration file or something? Is there a way to control which scaling algorithm is used? (bilinear, nearest neighbor, etc.)

I'm also wondering how this thing actually works. It's a TSR and hooks Int 10h. Does that mean if you set up some nonstandard video mode without using 10h that it won't work? I'm assuming that's the case.

Maybe someone can look at the assembly and see what the bot actually does and where it got its information from. Presumably there is public documentation that explains how to do this, that nobody has yet delved into because it was too obscure of a problem and no one has yet tried to solve it. Would be nice to see that documentation to gain an actual understanding of what's going on here.

What if you try to run Windows 3.1? Or even 9x? Could this aspect ratio and scaling control functionality be built into the generic VESA drivers for Win31 and 9x?

Hi,
1. No config file is needed since both algorithms are rather simple. Both the LCD native resolution and the source resolution are reported in MMIO registers. In case of the internal LCD panel In DOS by default all video modes are projected to the native resolution.
In case of aspect ratio correction mode the TSR instead calculates the 4:3 projection this way:

1TargetHeight  = NativeHeight
2TargetWidth  = NativeHeight × 4 / 3 
3HorizontalOffset = (NativeWidth - TargetWidth) / 2.

So in case of my 1600x900 LCD you get:

1TargetHeight = 900
2TargetWidth  = 900 × 4 / 3 = 1200
3HorizontalOffset= (1600 - 1200) / 2 = 200

Then all CGA/EGA/VGA and VESA resolutions are scaled and projected to this calculated window instead of the native panel resolution (regardless of the source size).

I n case of centered pixel perfect mode the calculated window size is equal to the source size and only the vertical and horizontal position/offset has to be calculated the same way as above.

2. Nothing has been found related to an explicit scaling mode. Most likely bilinear scaling is used in aspect ratio correction mode. In pixel perfect mode no scaling is applied at and above 640x480. For smaller resolutions 2x line doubling is applied implicitly.

3. In practice even in case of custom modes a basic mode is usually set before register level customization. And it seems it's enough since the real source size is not relevant (at least not in aspect ratio correction mode as can be seen above).
So the aspect ratio corrected projection can also work if the default native screen size projection works. In case of pixel perfect mode the final source size is relevant but I have not found problematic games/programs so far but of course they can exist anyway.

4. There was no clear documentation found for any of the problems besides the Linux driver sources. The solutions have been born through many trial and error steps. You can find the whole story here:
https://github.com/Falcosoft/T430LCD/blob/mai … eEngineering.md

Oh here's another question. I wonder if there's a way to enable the overscan background color.

Nothing related has been found so far and I do not think such thing exists since neither the Windows drivers nor the Linux drivers offer such option.

Neat. Would love to see this matured into a universal tool.

ASPECT utility confirmed working on Thinkpad x201. I'm pretty sure what the AI has hit on here is an undocumented Intel HD Graphics feature of some kind, not something Ivy Bridge specific.

keenmaster486 wrote on Yesterday, 22:40:

Neat. Would love to see this matured into a universal tool.

ASPECT utility confirmed working on Thinkpad x201. I'm pretty sure what the AI has hit on here is an undocumented Intel HD Graphics feature of some kind, not something Ivy Bridge specific.

If you are right then it's possible that the tool is already universal 😀 Unfortunately it's not likely that I will be the one who make this tool more universal since I'm not a big hardware collector. Besides the Thinkpad T430 I only have a HP-Compaq NX6325 (AMD Turion X2) and a Samsung Atom netbook. That's all.

I’ll test it on my other widescreen Thinkpads and see what happens. I have an X1 Extreme that is much more modern.

keenmaster486 wrote on Today, 00:06:

I’ll test it on my other widescreen Thinkpads and see what happens. I have an X1 Extreme that is much more modern.

OK, thanks in advance . If some of them work then would you be so kind to report the exact models together with CPU model and native LCD resolution?
I do not know if the Thinkpad x201 has more variations or not so I suppose that your model has Intel Core i5-560M and 1280x800 LCD panel. Is this true?

keenmaster486 wrote on Yesterday, 22:40:

Neat. Would love to see this matured into a universal tool.

ASPECT utility confirmed working on Thinkpad x201. I'm pretty sure what the AI has hit on here is an undocumented Intel HD Graphics feature of some kind, not something Ivy Bridge specific.

You may be right. I'm stil testing. ASPECT does not work on my laptop with Intel UHD-620 onboard graphics (Kaby Lake), a Dell Inspiron 5570 laptop, native resolution: 1920x1080.

I've got other systems to test: A Dell Latitude E5470 laptop, with Intel HD-520 graphics (1366x768 resolution) and an i5-6600 CPU, and a desktop with a Sandy Bridge system (i7-2600k). The desktop currently has a 16:9 monitor with a 1920x1080 resolution. If ASPECT works on the desktop with its an attached 16:9 display to force a 4:3 AR, that will again have provided additional pupose and use for the tool, beyond just changing AR to 4:3 in fixed built-in laptop screens.

Falcosoft wrote on Today, 00:10:

keenmaster486 wrote on Today, 00:06:

I’ll test it on my other widescreen Thinkpads and see what happens. I have an X1 Extreme that is much more modern.

I do not know if the Thinkpad x201 has more variations or not so I suppose that your model has Intel Core i5-560M and 1280x800 LCD panel. Is this true?

There are a few CPU models it came with, but all first gen Core i5/i7. Mine has an i5-540M. The panel is 1280x800, yes.

I also have a T430 that I have upgraded with a quad core i7-3840QM CPU and a 1080p IPS panel. Will test with that as well to see what happens. I'm guessing it will probably work just fine.

keenmaster486 wrote on Today, 00:35:

Falcosoft wrote on Today, 00:10:

keenmaster486 wrote on Today, 00:06:

I’ll test it on my other widescreen Thinkpads and see what happens. I have an X1 Extreme that is much more modern.

I do not know if the Thinkpad x201 has more variations or not so I suppose that your model has Intel Core i5-560M and 1280x800 LCD panel. Is this true?

There are a few CPU models it came with, but all first gen Core i5/i7. Mine has an i5-540M. The panel is 1280x800, yes.

I also have a T430 that I have upgraded with a quad core i7-3840QM CPU and a 1080p IPS panel. Will test with that as well to see what happens. I'm guessing it will probably work just fine.

When you run ASPECT on your Thinkpad x201 what calculated window size and offset is printed to output? Is it 1066x800 and offset 107,0 ?

zyzzle wrote on Today, 00:24:

... The desktop currently has a 16:9 monitor with a 1920x1080 resolution. If ASPECT works on the desktop with its an attached 16:9 display to force a 4:3 AR, that will again have provided additional pupose and use for the tool, beyond just changing AR to 4:3 in fixed built-in laptop screens.

At least on the T430 an external monitor (even through a digital connection) behaves completely differently to the internal LCD. Unlike in case of NVIDIA cards video modes are NOT projected to the native panel size of the external monitor. This is obvious right from the beginning since e.g.normal text modes are projected to 640x480 fixed resolution that makes text modes especially ugly on Intel HD graphics through a digital connection.
Currently ASPECT's logic relies on the inner LCD behavior.