Adamski_NZ

Generic PTA testing scenario - suggestions please!

Recommended Posts

Hi Dave,

 

Many thanks for your helpful input. Keeps us on our toes! I'll answer each of your points in turn.

 

Can you double check your time/day for your comparison Nelson shots? Look at the tree shadows at the north end of the field. they are different between your shots.

 

I am sure we have this covered. I used Adam's provided 'PTA Generic TEST 02' flight scenario files thus ensuring our initial viewpoint, weather, time and season setups were the same. However, your eagle-eyed observation regarding the tree shadow differences were due, I believe, to the position of our respective 'Scenery Object' sliders in P3D. Having brought this to my attention (thanks!) I find Adam's sliders are hard right (Extremely Dense). I will post another image in my next update for Adam below and you will see the change.

Are you running the same resolutions?

 

I am running 2560x1440x32. Adam's could be different.

 

With dynamic lighting, how will you ever know that you are both allowing the same amount of time for brightness and contrast and saturation to adjust? Or do you both run HDR off?

 

Screen captures throughout this exercise are always performed with the sim Paused so any spontaneous adjustments should be irrelevant. Anyway, 'Turn off HDR luminance adaptation effect' has been checked in Adam's PTA 20_10 Preset. HDR must be active in the sim for the Preset to work as intended when applied to the Shaders.

 

Edit: Been playing around with HDR settings and, yes, you are correct, the shadows are longer in Adam's image! Nice spot! Now how did that happen, I wonder? The settings in my Sim after loading Adam's flight file are Summer, Day, 7:00am. Therefore my image, presumably, was taken a little later. But then the image appears darker which appears to fly in the face of this theory, at least as far as it affects the lighting of the scene.

 

Regards,

Mike

Share this post


Link to post
Share on other sites
Help AVSIM continue to serve you!
Please donate today!

EDIT: It's also possible that your "Paint" program is changing the palette slightly. My scenario archive has the full .tif file in it - how does it compare with your uncompressed one?

 

I just checked my 378.49 install date: 5 days before taking the test shot.

Hi Adam,

 

Sorry, I should have been more specific. The 'Paint' Application is the one that comes with Windows7. I ran a few TIFF/JPEG image comparisons using Picasa Photo Viewer (makes it easy to flip back and forth) and my images are identical apart from being a little less sharp in the JPEG image.

 

Sadly my monitor calibration adventure turned out to be very disappointing :Hmmmph: Not sure why, but the resultant screen image turned out to be far too dark, in fact virtually unusable. Haven't used Colormunki Photo for a while, installed and updated the software and followed the instructions to the letter, including pre-test Calibration, and yet, to my eyes, the result is unacceptable. Tried upping the monitor brightness fom the default setting (80%) to 100% and still the image was too dark. I used the 'Easy' calibration mode which, I'm told, handles everything automatically, including making allowances for ambient lighting. perhaps if I knew what I was doing I could have experimented with the Advanced calibration route but, hey, life's too short. That will have to wait for another day.

 

So, clearly the ASUS ROG team really do know what they're doing. I've reverted back to the previous default .icm monitor profile, as shown in post #18 above, and all is well once again :smile: Also, looking at this test image seems to be providing further confirmation that the monitor is, in fact, set up correctly using this default profile.

 

http://web.ncf.ca/jim/photography/photofinishers/multitarget5.jpg

 

Here is your Image again, Adam. This time I have fiddled with my HDR settings producing another comparison image which now appears to be getting closer to yours in appearance. I am still using the REX TD Dawn (Set 22 Chill) / Day (Set 14 Fair) / Dusk (Set 03 Dust) Sky selection.

 

pta_generic_test_02_adam (Brightness: 0.70, Bloom: 0.20, Saturation: 0.85)

Bi4HNpj.jpg

 

My image with HDR adjustments (Brightness: 0.95, Bloom: 0.20, Saturation: 1.20

0dejiew.jpg

 

Unfortunately these HDR changes come with a cost in that the Sky is lighter in colour. The general lighting of the sim is brighter and I have noted some cirrus clouds exhibit clipping with loss of detail. Your settings, on the other hand, are so much better.

 

Regards,

Mike

Share this post


Link to post
Share on other sites

 

Hi Mike - two other thoughts - both long shots:

 

On my system, on one of the previous versions of P3D, when I first loaded the sim and for quite a while after I had a problem with dim and dull looking skies. I even complained about it at the LM forum. After running the sim for a while the problem seemed to go away. And then one day I decided to delete my shaders and the dullness problem returned. It appeared that my shader cache had to build up for the sky to look better. Maybe that's an issue here?

 

The latest versions of P3D seem to be good from the start for me but I cant help but think that the P3D setup program may not include all possibly useful Visual C++ Runtimes and if I install a game like Train Simulator or some driving sim then it may install something that makes P3D look better. It may install a runtime that has a function that P3D can use (or an updated version of the same function).

 

I really have no idea if this may be the case but it couldn't hurt for you and Adam to compare installed Visual C++ Runtimes. I've got 8 or 10 on my system and I try to avoid deleting my shader cache too often as I'm convinced its not a good idea.

Share this post


Link to post
Share on other sites

Dave - the shader cache is possibly a red herring. Just waiting a while or just panning around usually does the trick. I sometimes do that, then simply reset the scenario to take the test shot. In any case, when I'm tweaking the tweaks, I *have* to rebuild the shader cache.

 

I'm intrigued by the apparent co-relation with C++ runtimes. Coincidentally, I also run Trainsim - so you may be onto something. It's really confusing when you have to have so many versions on a system (not to mention the same thing with .net). If the libraries are affecting colour output, then that's really bad news - as the combinations of libraries different people may be running will be limitless :-(

 

Here are mine:

c_libraries.png

 

Mike - tweaking HDR is a really dodgy business. I find that even 1% or 2% changes in brightness can cause the clipping you've been getting. A better workaround for you would be to play with the PTA settings (Post-processing contrat, maybe), but that still isn't addressing the problem of *why* our results are different.

 

I'm wondering whether it's simply hardware related (GFX card). When reading reviews of similar chipset cards (but different manufacturers) I often find comments that imply that there are colour differences between (say) EVGA and MSI for the same card (eg GTX 970). If that's the case, then producing identical results will be nigh-on impossible unless we all own decent calibration hardware as well.

 

Shame about your Munji. I've heard from people in the publishing industry that many of the cheaper calibration devices are totally useless - and only the top end ones work correctly. A professional outfit can recoup the costs fairly easily, but it's just too expensive for the average home user.

 

Adam.

Share this post


Link to post
Share on other sites

Hi Adam,

 

...and here are mine:

 

71IkXOp.jpg

 

I do see a few differencies.

 

Missing from my installation:

Microsoft Visual C++ 2005 Redistributable (x64)  (The other 2 are present)

Microsoft Visual C++ 2008 Redistributable
(x64) 9.0.21022    
(x64) 9.0.30729.17    
(x64) 9.0.30729.4148

Microsoft Visual C++ 2012 Redistributable
(x64) 11.0.60610
(x86) 11.0.60610

Missing from your Installation:    

Microsoft Visual C++ 2015 Redistributable
(x64) 14.0.24210
(x86) 14.0.24210

 

I confess I am finding it hard to know how much importance should be attached to these findings in relation to the subject of this thread. To date no adverse impact has been experienced by any of my installed Applications. Presumably each App will either install the appropriate prerequisite, if not  present, or flag its absence for user intervention.

 

Edit: I've submitted a Support Case with x-rite regarding my ColorMunki issues. Notwithstanding its relevance here I felt it would be daft not to pursue this further. It will be interesting to see what they have to say.

 

Regards,

Mike

Share this post


Link to post
Share on other sites

here are mine for what its worth:

cppruntimes170211.png

 

You both have more than I but it appears to me that both Adam and I have the x64 2008 9.0.30729.17 and , Mike, you don't. double check me on that. EDIT: also the 2015 runtimes. ackk!

 

The only reason I brought this up is because of a video editor I have as freeware that worked on one computer but then wouldn't work on a new computer and I tracked it down to a 2008 C++ runtime that had been installed by Test Drive Unlimited2. I installed that and the editor worked again. things like that do happen.

 

This may all be a distraction and if it is then I apologize.

 

I just have no idea why the same textures with same settings would produce such different results.

Unless the saved settings file doesn't actually save all needed settings - like when you save graphics settings in P3D, it doesn't save everything in the P3D.cfg related to graphics.

Share this post


Link to post
Share on other sites

Hi Dave,

 

It's a shame that nobody else has felt inclined to pitch in by submitting images using Adam's testing scenarios, especially 'pta_generic_test_02_adam'. It would certainly help to clarify the extent of the 'problem' if, indeed, a problem, in the strict sense, truly exists. We only have one submission so far which is hardly significant as I may be the only one affected. Alternatively, and as Adam freely admits, he could be the only soldier out of step. However, I do have to say that all his posted fabulous images to date strongly suggest that this is very unlikely to be the true position. Without more data there is just no way of knowing who occupies the high ground.

 

No one doubts the dramatically positive results achieved in P3D by using Adam's Presets, least of all me. Indeed, I was, and still am more than happy with what I am seeing on my monitor screen. I suspect others feel the same way. In the end it may be that such differences can be safely regarded as being largely irrelevant since true beauty, as always, will lie in the eyes of the beholder.

 

Still, as an intellectual exercise, it would be good to know what is responsible for these differences.

 

Regards,

Mike

Share this post


Link to post
Share on other sites

 

Mike - I'm in this solely for the intellectual exercise as I don't use PTA. I enjoy the challenge of making my own sky and clouds to suit my needs and I like to keep things as simple as possible which means as close to default as possible.

 

I really thought I was onto something with the tree shadows and time of day discrepancy but I guess not. I don't know what else it could be.

Share this post


Link to post
Share on other sites

it appears to me that both Adam and I have the x64 2008 9.0.30729.17 and , Mike, you don't. double check me on that. EDIT: also the 2015 runtimes. ackk!

Hi Dave,

 

Okay, I should follow this up by installing 2008 - x64 9.0.30729.17 and try again. Likely it won't make any difference, but I should eliminate this possibility nevertheless. As for the 2015 runtimes, these may not be relevant as they are missing from Adam's installation, not mine.

 

Edit: Surely these installed components are a reflection of what dependent Applications are installed at a particular point in time on our respective systems?

 

Regards,

Mike

Share this post


Link to post
Share on other sites

Hi Readers!

Okay, I believe we can now rule out the Visual C++ runtime libraries as being the causative factor behind this puzzle.

I ran the appropriate installers and ended up with:
pISFtt1.jpg

Summary of what actually installed:
Microsoft Visual C++ 2005 Redistributable (x64) -----> Installed
Microsoft Visual C++ 2008 Redistributable
(x64) 9.0.21022 ----------------------------------Did not install from the package
(x64) 9.0.30729.17 ----------------------------------------> Installed
(x64) 9.0.30729.4148 -------------------------- Did not install from the package
Microsoft Visual C++ 2012 Redistributable
(x64) 11.0.60610 --------------------------------Did not install from the package (Installer asked to Repair/Modify/Uninstall 11.0.61030)
(x86) 11.0.60610 --------------------------------Did not install from the package (Installer asked to Repair/Modify/Uninstall 11.0.61030)

My computer was rebooted and PTA v2.10 reloaded.
Original Shaders were Restored and I Set inital parameters for all tweaks.
Adam's PTA 20_10 Preset was then Opened again and Applied using PTA v2.10.

Here are the results. First, Adam's image again for ease of comparison:

pta_generic_test_02_adam (Preset 20_10 Brightness 0.70, Bloom 0.20, Saturation 0.85)
Bi4HNpj.jpg

My Image using the same HDR settings (Brightness 0.70, Bloom 0.20, Saturation 0.85)
jEUbNFR.jpg

My Image using adjusted HDR settings (Brightness 0.95, Bloom 0.20, Saturation 1.20)
mSCdC9T.jpg

Lastly, and simply out of curiosity, I reinstalled the ENVTEX Sky textures and captured another image again using Adam's 20_10 Preset recommended HDR settings (Brightness 0.70, Bloom 0.20, Saturation 0.85)
QErVlMw.jpg

This image appears marginally lighter when compared with the second image above.

Regards,
Mike

Share this post


Link to post
Share on other sites

Since commenting about the appearance of that last image I've now compared the two on my iPad with the brightness slider setting at maximum. Contrary to my earlier remark I think they are, in fact, very close to being identical.

 

REX TD Dawn (Set 22 Chill) / Day (Set 14 Fair) / Dusk (Set 03 Dust) Sky texture selection were in use for the top 3 images in post #35.

 

Mike

Share this post


Link to post
Share on other sites

Thanks for your perseverance, Mike! I'm beginning to go for the idea that video cards themselves introduce differences in contrast/colouring.

 

BTW - I found a good way of comparing the images: place them all in one location, then view them in Windows Photo Viewer (or Preveiw, whatever it is). It has really handy forwards/backwards controls (also via cursor keys). You can quickly flip back and forth.

 

Adam.

Share this post


Link to post
Share on other sites

'm beginning to go for the idea that video cards themselves introduce differences in contrast/colouring.

 

Hi Adam,

 

I think you might be right - mine are 2xGTX 980ti's, what's yours? You did refer to the GTX 970 in an earlier post although it was not clear whether that was your card. If manufacturers have to adhere to reference designs for the circuitry and board components then one would expect results to be similar if not exactly the same at stock.

 

However, if nobody else contributes to this exercise including details of their hardware we will remain forever stumbling in the dark ☹️

 

Concerning comparison of images, yes, I made the same discovery, only I have been using Picasa Photo Viewer which does the same. Very handy!

 

Regards,

Mike

Share this post


Link to post
Share on other sites

I think you might be right - mine are 2xGTX 980ti's, what's yours? You did refer to the GTX 970 in an earlier post although it was not clear whether that was your card. If manufacturers have to adhere to reference designs for the circuitry and board components then one would expect results to be similar if not exactly the same at stock.

 

 

It is, in fact, a 4Gb nVidia GTX 970 (sub-vendor: Gigabyte) - fairly middle range, I suspect.

 

I was going through my prepar3d.cfg and noticed a few things that might make a difference:

 

1) HDR_BLOOM_THRESHOLD=3.000000

2) HDR_BLOOM_BLUR_SIGMA=0.800000

3) DAWN_DUSK_SMOOTHING=1

 

I've no idea how the values for 1] and 2] are arrived at. I certainly haven't fiddled with them.

 

Adam.

Share this post


Link to post
Share on other sites

Thanks, Yuri - I looked at that page (advanced settings) ... I hope you understand some of it, because I don't LOL!

 

It's a shame page doesn't tell you what the *default* values are (though it gives the range).

 

Adam.

Share this post


Link to post
Share on other sites

1) HDR_BLOOM_THRESHOLD=3.000000

2) HDR_BLOOM_BLUR_SIGMA=0.800000

3) DAWN_DUSK_SMOOTHING=1

Hi Adam,

 

Mine are the same. Even if adjustment of these values produced worthwhile improvements they would only be seen on the end user's monitor, just as would any Calibration modifications made using NVIDIA's utilities (see post #18). Also, the latter would produce changes impacting all Applications and Games which would probably be undesirable. Fine if your rig is for the exclusive running of P3D, but I suspect most of us cannot afford that luxury.

 

Regrettably, I suspect we are on a hiding to nothing with this one. I believe we should start by acknowledging that Developers of Games or Simulators will likely be working their creations on accurately calibrated monitors and this may not apply here. That way they can be assured that their customers will, in the majority of instances, get a good, if not perfect experience with their default setups. Those that do not will have some 'free' tools (apart from those provided on the monitor) at their disposal to improve matters. These will come in the form of limited adjustments like Brightness and Contrast embedded in the Application or, if you are feeling more adventurous, additional Calibration tools and controls as are provided by Windows and NVIDIA.

 

I think the only way to ensure uniformity of results is for each of us to bite the bullet and endeavour to achieve accurate Calibration of our monitors. That way we can be reasonably sure that we do see stuff as the Developer intended. Provision of a generic test image could be helpful, but unless we can match what we see by making changes solely from within the simulator and without adversely affecting the lighting of other scenery elements like clouds then, like monitor Calibration, the effects will be witnessed system-wide.

 

Regards,

Mike

Share this post


Link to post
Share on other sites

Have been following the thread and as I haven't had FTX nz on for awhile so I decided to give it ago.

Monitor iiyama 27" 2560x1440.

Nvidia driver 378.49 and  colour settings same as yours Mike,

GTX 980Ti

u7mTxpS.png

Share this post


Link to post
Share on other sites

Hi Bob,

 

Thanks, that could be very helpful. At first glance your image appears much closer in appearance to Adam's than it is to mine.

 

Can I ask you to clarify a few points:

1. Did you use Adam's flight situation flight files as provided in his TEST scenario 02 (see post #14)?

2. Could you confirm the sim is Paused and your Scenery sliders are all far right and all Cast and Receive shadow boxes are checked?

3. Are you using EZDok or ChasePlane? I find it is easier to set up the image before capture by disabling ChasePlane and then selecting Cockpit view - the Cub's cockpit should then disappear.

4. Please check Season and Time of day - should be Summer and 7.00am respectively.

5. Are you applying Adam's 20_10 Preset with PTA 2.10 and using his recommended HDR settings (not my 'adjusted' settings) - they should be applied automatically when the Preset is applied, but might be worth checking just in case.

6. Is Wideview checked in the sim?

 

Would it be possible for you to recapture your image by recomposing to match Adam's image and Scenery content?

 

Edit: Also, capture your images as TIFF files (I think, if I recall correctly you will find you can configure that under P3D 'General Settings') by hitting the <V> key. Load the file into Windows 'Paint' and resize to 1280x720. Then save as a JPEG image for uploading to the forum.

 

Thanks again.

Regards,

Mike

Share this post


Link to post
Share on other sites

Hi Bob! Thanks for test shot. Great to have another one to compare with!! However ... it does seem to look like you have a different time of day <?>.

 

I find the easiest way of doing it is to set SLEW *and* PAUSE then use the P3D menu to "Reset the Scenario". That should take you back to where the shots should be taken and in pause/slew mode so you can take your time taking the shot.

 

You have the same card as Mike, so I'm really keen to see what yours looks like!

 

Adam.


3. Are you using EZDok or ChasePlane? I find it is easier to set up the image before capture by disabling ChasePlane and then selecting Cockpit view - the Cub's cockpit should then disappear.

 

Mike/Bob: I use OPUS Live Camera Views, but took care to NOT to use them for the test shots. It should be a standard P3D forward view (but, as Mike says,with the cockpit turned off) and at standard zoom (hit Backspace).

 

Adam.

Share this post


Link to post
Share on other sites

Hi Mike

 

1. sorry missed that bit , don't have Rex

2. yes

3. EZdoc v2

4. now altered from 1300 hrs t0 0700 hrs

5 yes

6. yes

 

new shot as requested with file type.

 

bob

 

3G1Iisu.png

  • Upvote 1

Share this post


Link to post
Share on other sites

Well, its morning here and Im still waking up so let me just say thanks, Mike, for conducting the experiment and Im sorry I put you on a wild goose chase but its good to know that's not an issue.

 

I'll continue following the thread as I love a mystery. Good luck, guys.

 

P.S.: the HDR Day Exposure value in the config can greatly alter the scenes appearance. I used it in an earlier 3.x version - not sure which one now - to make the sky more alive. Those values are default. day exposure has been and should still be .28.

 

also - note that all three of you have different colored tarmac and runway in your Nelson shots. But , in Bobs case, that could be due to the sky texture used.

Share this post


Link to post
Share on other sites

It's starting to get harder to compare images on different web pages. So, I thought it would be helpful to repost our 3 'default' images to allow easier comparison.

 

First, Adam's image from Scenario 02

Bi4HNpj.jpg

 

Mine

V0laZj5.jpg

 

Bob's

8BdaNyM.jpg

 

Concerning the differences in Tarmac/Rwy: in my case this is almost certainly from REX-Worldwide Airports HD (My Ground Environments selection). I don't think it matters other than, perhaps, as an aesthetic distraction :smile:

 

Well, I'm not sure that this get's us any closer to an explanation as Bob's appears to me to lie somewhere between Adam's and Mine. The thick plottens! Seeing more sky in Bob's - would this have an influence? Also, he is not using REX Sky textures which, unfortunately, is likely to be an issue. If we are striving for meaningful results we do need to be on the same page with respect to Sky textures. Perhaps one of Gennadiy's compilations would be an appropriate choice for these tests?

 

Mike

Share this post


Link to post
Share on other sites

Just to add to the ingredients in the recipe my iiyama monitor has some presetts built in, Standard, scenery, cinema and game. it was set on scenery for this test shot.

There were slight differences on the lightness to dark and the game preset being more vivid.

 

Just to note I set my ShadersHLSL file back to default.

 

bob

Share this post


Link to post
Share on other sites

Hi Adam,

 

Don't hold your breath, but I think I may have made some progress!

 

It looks like the ColorMunki Calibration LUT loader didn't do its job properly first time around. I decided to try the recently created ROG PG278Q_Easy_D65_201702101603.icm colour calibration profile again today. When the system rebooted I thought I had done something wrong as the desktop image appeared very little different to what it had been before. Everything was checked again and verified and the system rebooted twice again and each time with the same result!

 

I reinstalled the REX4 TD Sky Textures, ran P3D, loaded the Test Scenario 02 and captured another image.

 

In each case the HDR settings remained unchanged (Brightness 0.70, Bloom 0.20, Saturation 0.85)

 

Judge the result for yourselves.....

 

Adam's (pta_generic_test_02_adam) image

Bi4HNpj.jpg

 

My image following initial monitor colour recalibration - ROG PG278Q_Easy_D65_201702101603.icm set as default

B0N8xLF.jpg

 

My image using the original ASUS PG278Q Color Profile, D6500 set as default

V0laZj5.jpg

 

Now, while my image following recalibration still does not match Adam's, it is getting closer and is certainly an improvement on my precalibration image. Of the two, Adam's is the warmer image but which, I wonder, is providing the most natural look? Tough call!

 

So, after all, I would suggest that this confirms monitor calibration to be one important variable in what is, quite evidently, a complex equation in our search for optimal imagery as displayed by Prepar3D following enhancements provided by PTA. Also, while its importance cannot be overestimated I think we can say with confidence that this is only one of many variables contributing potentially to any differences observed on our displays.

 

Regards,

Mike


 

 


Just to note I set my ShadersHLSL file back to default.

 

Hi Bob,

 

Again, just to clarify, after 'Restoring the original P3D shaders' (or did you mean 'Set initial parameters for all tweaks'?) did you 'Open' and 'Apply' Adam's 20_10 Preset?

 

Regards,

Mike

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now