v0.4.6.4 MSFS EXE.xml Hotfix

[ResetXPDR]

MSFS Compatibility: 1.39.9.0, 1.6.34.0.

Release notes for 0.4.6.4

This update fixes EXE.xml formatting so all MSFS tools read it reliably and improves Exception 31 handling with clearer guidance.

Please review the updated readme here, which provides a detailed description of the app, installation, features and settings, before asking any questions or raising any issues about the app.

MSFS EXE.xml handling fixes:

• Fixed a legacy installer issue where older AutoFPS versions may have saved the MSFS EXE.xml in a format incompatible with some installers (FSUIPC in particular); AutoFPS now detects and rewrites it into a universally compatible one.
• Corrected the AutoStart with MSFS installer option so the EXE.xml is now saved in a universally compatible format for MSFS and all addons.

Exception 31 guidance:

• Updated the Exception 31 popup with clearer guidance and a link to the CLEAN reinstall instructions.
• Made the custom message box action button fully configurable for different message types.

---

This discussion was created from the release v0.4.6.4 MSFS EXE.xml Hotfix.

[jerryyyyy]

Hi the update I guess changed the test to false but I put it back to true. All is good now and here is the log for the record.
2D RENO MSFS_AutoFPS20260206.log

[ResetXPDR]

Can't say I've looked that closely. Cpt_Piett (where has he been lately anyway?), swears it made a difference with mega TLOD. Also, the fact that those super high TLOD could not always be achieve suggest there is extra processing load with increasing TLOD, regardless of whether or not it is visually descernable.

[jerryyyyy]

Well you know I have this computer to do astroimage processing so I decided to look at a couple shots and to get a basis for comparison. I use my takeoff point at RENO and took some screenshots with just MSFS at TLOD=300 and then with AUTOFPS at TLOD=800. Long and short of it when you blow them up in Photoshop they show almost exactly the same pixelation. Thinking it might be something about the MS screenshot that equated all images I took a screenshot of mt 4K monitor using Windows and then using another screen shot software and the results were the same.

I have to say I had the same impression as the Captain especially at TLOD=1000 at high altitude...

The images are too big to include here but you should be able to get them from a OneDrive Link

I note the 300 and 800 PNG TLOD screenshot images are almost exactly the same size, which would not be the case if one had twice the resolution as the other.

[jerryyyyy]

I have some good news. I used my Astronomy Software on your images at TLODs of 10 200 400 and 750. The bottom line is though it may be imperceptible to old eyes, there is more variance (differences) among the pixels in the high TLOD. MS did not internally cap the TLOD.

The Stats are: In PixInsight, AvgDev (Average Absolute Deviation) and MAD (Median Absolute Deviation) are robust statistical estimators used primarily in the Statistics process, ImageIntegration, and NoiseEvaluation to measure image dispersion (noise). They are preferred over standard deviation (StdDev) because they are less affected by outliers, such as bright stars or hot pixels.

The AvgDev of images TLOD of 750->010 are 11607 11150 11131 and 10928... so the one with your AUTOFPS is "scientifically/mathematically" "better".

I highlight maybe (the) one spot in the image you can tell the difference... so the Captain can trust his eyes.

[ResetXPDR]

I did take a look at your high detail images in your post before, then promptly forgot to reply 😄 I was going to say that the LOD increases CPT_Piett loved so much were at cruise altitudes, where it is much easier to see the increased draw radius. Nonethless, you subsequently went all sciency on me 😆 and proved it mathematically, so well done!

On another note, I have been watching the development of of a VR optimisation script by Shark, as showcased here, and found the bits on auto power profile switching, CPU core affinity optimisation and MSFS process priority interesting. I have actually beeen helping Shark fine tune the power profile part, then had a thought that I could possibly do all three in AutoFPS as a option, in a way that does not interfere if Shark's script is also running. I have floated the idea with Shark and received his endorsement to proceed, so I will probably have a test version out very soon.

[jerryyyyy]

Well I got very interested in the resolution of the screen and did one more picture in color showing what happens at altitude. Numbers are even better at TLOD 1000, and as the Captain mentioned it looks good. Getting the detail out of images is the name of the game in astrophotography.

Coming back to the new script I tried it and it is hard for me to tell if it worked but it looked good. I made several changes in the meanwhile especially invoking the ASUS GPU Tweak AI for the 5080, If you look at this log you will see the GPU is running 90% but it is stable and the image is very detailed for VR. I think it is TLOD 200 and ULTRA Clouds and Trees plus the aircraft injection at a custom airport... not much more I could ask.

Test MSFS_AutoFPS20260208.log

I think back in my 3070ti days we discussed background processes and I found Chrome did not hurt as much as Mozilla (which is not on his list). Whatever he did to turn on and off OneDrive is a good idea for sure since that is really a ghost in the machine. I was afraid it would not restart but it did. It has a pause function that may be safer.

The script run is the last on this log and a long RENO run earlier trying to invoke a crash. My system is not fully stable at -10 and have gone to -7 on advice of the OC gurus. It does not crash under load but just unexpectedly. Once a day maybe. My machine has never been fully stable and I get 1-2 crashes a week usually at night which I took to be an update thing. I have a series of again more numbers on my processor Cores speeds and loads and they vary substantially so it is hard to make a one stop shopping adjustment with a global -10 drop.

Oh, one more thing I have learned is that my system rate limiting step is not the processor itself but rather the memory. It is almost 4 years old now and was rated at 5200 which is all I can get from it. The under volting is not going to fix that as most seem to be running at 6000 or better. It is almost impossible to consider replacing my 64GB that I need for image processing.

Clear Skys.