[Coolhand101]

The scrolling has been around longer than when i first started using OR. I find that i use pan up to view gantry signals at close range or at night when looking at the stars and panning down to see the buffer stops at dead end tracks. I believe i even requested that the scroll wheel on a mouse can also use this feature. And i like the way OR now remembers the scroll angle when switching views. I have had my angle from the video option page at 37 degrees for some time now.

The nose down view is greater in them pictures is because the cab angle in the CVF is defaulted to 14 degrees down. In my cabs, i try and use a angle that the cab scrolling angle is equal for up and down.

85% of my cabs are at 16x10, but when i got my new monitor(16x9) i amended a few of these cabs to 16x9, so this equals 100% 2D cab stretch. After much testing, i found that 16x10 cab images showing correctly on a 16x9 monitor is 63% 2D cab stretch which still allows the angle scrolling( to a lesser degree though ).

I for one would like to keep this scrolling angle of the cab. Maybe the up and down arrows keys can be use for translation of the cab image and shift with the up and down arrow keys for scrolling the angle of the cab! The best of both ?

Thanks

[Csantucci]

The incoherent world view from the cab in widescreen windows or monitors has been solved with x.3917, using the patch present in the files of post #31.
The rule of modification of the world view when vertically scrolling the cab is another issue. At the moment that hasn't been modified.

[James Ross]

Csantucci, on 14 August 2017 - 12:38 PM, said:

The incoherent world view from the cab in widescreen windows or monitors has been solved with x.3917, using the patch present in the files of post #31.

If you'd included a .patch file in post #31, I could have warned you that this change would not be agreeable to me. It is a Band-Aid specifically applied when two settings are set to certain values (admittedly both the default I believe) and does not do anything for the general case. :(

[Csantucci]

Hi James,
the question is: what should the patch solve?
At the moment my answer has been: the most frequent problem, that is a 4:3 cabview seen on a widescreen monitor, with no stretch (like Lindsay, I believe stretching a cabview and therefore having circular gauges becoming ellyptic is not a solution). The patch in fact solves this case and leaves the other cases untouched, so I think it is a step forward.
How to generalise it? What other cases should be considered and how should they be solved? I'm open to generalise the patch if I have precise and feasible inputs, having however in mind that, like Lindsay, I'd think some freedom should be left to users to define what FOV they prefer for their cabview. There is also quite a bunch of possible cases, taking into account preset vertical FOV, cabview aspect ratio, monitor aspect ratio (there are also various widescreen aspect ratios) and, if really important, cabview stretch. There could also be the case of a widescreen cabview used on a 4:3 monitor, or in general with a monitor with a lower aspect ratio. This is a case which is uncovered at all by the code: see here the extreme case of a 4:3 cabview seen on a 1:1 window (however the same effect would occur in the case of a widescreen cabview seen on a monitor with a lower aspect ratio)

[James Ross]

Csantucci, on 15 August 2017 - 12:35 AM, said:

How to generalise it? What other cases should be considered and how should they be solved? I'm open to generalise the patch if I have precise and feasible inputs, having however in mind that, like Lindsay, I'd think some freedom should be left to users to define what FOV they prefer for their cabview. There is also quite a bunch of possible cases, taking into account preset vertical FOV, cabview aspect ratio, monitor aspect ratio (there are also various widescreen aspect ratios) and, if really important, cabview stretch. There could also be the case of a widescreen cabview used on a 4:3 monitor, or in general with a monitor with a lower aspect ratio. This is a case which is uncovered at all by the code: see here the extreme case of a 4:3 cabview seen on a 1:1 window (however the same effect would occur in the case of a widescreen cabview seen on a monitor with a lower aspect ratio)
1to1.jpg

One way of generalising it would be perform the "correction" for all FOV values, adjusting them to match what would have been seen with that same FOV on the cab view's original aspect ratio. In other words, no matter what display aspect resolution and FOV the user selects, the cab view will have the same outside view as if it was being loaded on the cab view's own aspect ratio.

Though I will say that the stretch option is hard to link up to this. Maybe the stretch should just be a toggle instead, on or off? Then you apply the above correction for any FOV on any display aspect ration for all cab views if stretch is off. If stretch is on, it's stretched and that's all we need to do.

[James Ross]

James Ross, on 15 August 2017 - 11:36 AM, said:

One way of generalising it would be perform the "correction" for all FOV values, adjusting them to match what would have been seen with that same FOV on the cab view's original aspect ratio. In other words, no matter what display aspect resolution and FOV the user selects, the cab view will have the same outside view as if it was being loaded on the cab view's own aspect ratio.

In particular, we should also be correcting 16:9 cab views being loaded in 4:3 aspect ratios - not because that's especially likely, but if we correct all mismatches then things should be better for everyone, on 4:3, 16:9, 16:10, or any other aspect ratio. That's what I meant about generalising; yes, it's probably more complicated to calculate the correction, but it's not conditional.

[James Ross]

Coolhand101, on 13 August 2017 - 02:07 PM, said:

I for one would like to keep this scrolling angle of the cab. Maybe the up and down arrows keys can be use for translation of the cab image and shift with the up and down arrow keys for scrolling the angle of the cab! The best of both ?

I'm certainly quite happy to remove the view rotation that occurs when you scroll up and down. It's never felt right to me (and probably never can, due to the 2D/3D mismatch).

[cjakeman]

I wonder how many people still use "stretch" values > 0.

Perhaps we could remove it or move it to the Options > Experimental.

Does anyone here make use of it?

[James Ross]

cjakeman, on 15 August 2017 - 12:00 PM, said:

I wonder how many people still use "stretch" values > 0.

Perhaps we could remove it or move it to the Options > Experimental.

Does anyone here make use of it?

In the survey I did in 2014, we got:

Values used in setting Cab2DStretch:

    0    27 (Default)
   65     1          
  100     7          

We should do another survey and/or add opt-in instrumentation to OR. :)

[Csantucci]

With x.3918 a first generalization has been implemented: now the patch works for any preset vertical FOV.
Missing generalizations:
- intrinsic cabview aspect ratio != 4:3
- viewing window aspect ratio < intrinsic cabview aspect ratio.
Both generalizations require also generalizations of the actual cabview resize logic that, as can be seen within code and within code comments, considers cabview aspect ratio = 4/3 and window aspect ratio >= than 4/3. Therefore implementation time is unpredictable at the moment.

Horizontal stretch is not foreseen to be considered (I would agree to remove it).