 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 24/10/2014 22:01, StephenS wrote:
> Try "StereoPhoto Maker", it has many types of output.
I had already looked at that and it's profusion of outputs and no
documentation, for the latest version, put me off.
After your post I downloaded an earlier version with a help file and
found that I can run the NVidia viewer from inside StereoPhoto Maker.
This is just what I wanted. Thanks Stephen.
Also thanks for your "parsing the camera" code. I used that along with a
modified "screen.inc" to automate setting up the stereo cameras.
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 24.10.2014 22:16, schrieb Stephen:
> On 24/10/2014 20:37, Le_Forgeron wrote:
>> I have a patch for a stereoscopic camera in povray.
>>
> Interesting?
>
>> You have to consider a few things:
>> * what is the distance between the centre of each "eye" ?
>
> Indeed! and what effect does increasing the distance have?* Some of my
> reading suggests that for distance shots an extended baseline is
> recommended.
I consider that recommendation nonsense: Yes, long-distance shots with
standard eye distance are problematic for 3D because they provide little
to no 3D effect at all.
However, increasing the eye distance has exactly one primary effect: It
makes the scene appear smaller. (Likewise, decreasing the eye distance
makes it appear larger.)
>> * where is the focus point where both centres of picture's rays meet.
>
> Again, is there a recognised behaviour when you change that point?*
My recommendation would be to put the focal point at a distance
equivalent to the expected viewing distance to the screen.
Changing that point won't mean a thing to the 3D effect, but may cause
discomfort.
There's one other very important thing to consider though: Camera angle.
For best results, the camera angle should be the same as the expected
"screen angle" in real life. Any other than that, and you'll introduce
distortions that may cause discomfort.
All in all, the camera parameters should match the expected viewing
situation as closely as possible, with the display corresponding to the
look_at plane.
Obviously, following this advice is not practical, because it severely
limits your choice of camera perspective :-P - but it should give you a
sound starting point.
>> * do you want parallel or crossed visualisation
>
> I don't know. I have lcd shutter glasses I guess that parallel would be
> more comfortable.
With LCD shutter glasses, the question is irrelevant, as it applies only
to side-by-side viewing.
> The articles I've read talk about a lens's near and far points depending
> on the focal length of the lens. Which in itself gives an estimate for
> the angle of view.
> That is not pinhole camera stuff. (I hope)
Stick to the pinhole stuff, it is actually a better fit to the viewing
situation (presuming you don't have one of those fancy new curved panels).
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 26-10-2014 8:05, clipka wrote:
> Am 24.10.2014 22:16, schrieb Stephen:
>> On 24/10/2014 20:37, Le_Forgeron wrote:
>>> I have a patch for a stereoscopic camera in povray.
>>>
>> Interesting?
>>
>>> You have to consider a few things:
>>> * what is the distance between the centre of each "eye" ?
>>
>> Indeed! and what effect does increasing the distance have?* Some of my
>> reading suggests that for distance shots an extended baseline is
>> recommended.
>
> I consider that recommendation nonsense: Yes, long-distance shots with
> standard eye distance are problematic for 3D because they provide little
> to no 3D effect at all.
>
But, it is the reason why on goggles, the baseline /is/ extended.
Consider also those military stereoscopic viewers. From my own
experience with stereoscopic viewers used for aerial photographs,
extending the baseline makes objects appear like billboards.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
You are quite right. It may cause discomfort. :-)
I had a bit of eye strain yesterday. So I will give it a rest today.
On 26/10/2014 07:05, clipka wrote:
> Am 24.10.2014 22:16, schrieb Stephen:
>> On 24/10/2014 20:37, Le_Forgeron wrote:
>> Indeed! and what effect does increasing the distance have?* Some of my
>> reading suggests that for distance shots an extended baseline is
>> recommended.
>
> I consider that recommendation nonsense: Yes, long-distance shots with
> standard eye distance are problematic for 3D because they provide little
> to no 3D effect at all.
>
> However, increasing the eye distance has exactly one primary effect: It
> makes the scene appear smaller. (Likewise, decreasing the eye distance
> makes it appear larger.)
>
>
I have two options.
Experiment with a simpler scene. Or make my mind up which scale to use.
Whether it is the internal scale of the scene or the scale of the
external viewer. (I think the answer is in the option)
>>> * where is the focus point where both centres of picture's rays meet.
>>
>> Again, is there a recognised behaviour when you change that point?*
>
> My recommendation would be to put the focal point at a distance
> equivalent to the expected viewing distance to the screen.
>
Bingo! Thanks. A flash of insight there.
> Changing that point won't mean a thing to the 3D effect, but may cause
> discomfort.
>
>
It does, on some of my tests I noticed eye strain when focusing on
different parts of the scene.
> There's one other very important thing to consider though: Camera angle.
> For best results, the camera angle should be the same as the expected
> "screen angle" in real life. Any other than that, and you'll introduce
> distortions that may cause discomfort.
>
>
> All in all, the camera parameters should match the expected viewing
> situation as closely as possible, with the display corresponding to the
> look_at plane.
>
>
I think I have it now. I have been using the wrong viewpoint.
> Obviously, following this advice is not practical, because it severely
> limits your choice of camera perspective :-P - but it should give you a
> sound starting point.
>
If only the "Mind's Eye" camera could be developed. ;-)
>
> With LCD shutter glasses, the question is irrelevant, as it applies only
> to side-by-side viewing.
>
I thought it might.
>
>> The articles I've read talk about a lens's near and far points depending
>> on the focal length of the lens. Which in itself gives an estimate for
>> the angle of view.
>> That is not pinhole camera stuff. (I hope)
>
> Stick to the pinhole stuff, it is actually a better fit to the viewing
> situation (presuming you don't have one of those fancy new curved panels).
>
I will and I don't.
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 26/10/2014 08:13, Thomas de Groot wrote:
> But, it is the reason why on goggles, the baseline /is/ extended.
> Consider also those military stereoscopic viewers. From my own
> experience with stereoscopic viewers used for aerial photographs,
> extending the baseline makes objects appear like billboards.
Would that not be more of the "telescopic" effect of the lens's used?
--
Regards
Stephen
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 26-10-2014 9:54, Stephen wrote:
> On 26/10/2014 08:13, Thomas de Groot wrote:
>> But, it is the reason why on goggles, the baseline /is/ extended.
>> Consider also those military stereoscopic viewers. From my own
>> experience with stereoscopic viewers used for aerial photographs,
>> extending the baseline makes objects appear like billboards.
>
> Would that not be more of the "telescopic" effect of the lens's used?
>
Hmm... I am not sure how much the lens influences the effect indeed. My
guts tell me it is the baseline mostly, but then who would rely on his
guts? :-)
Remember those stereoscopic viewers from our youth? Viewmaster
http://www.bbc.co.uk/ahistoryoftheworld/objects/zkFmx_c9QM6C23Ld3Qd6sg
they also had that exaggerated 3D effect, less billboard-like maybe but
still.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 26.10.2014 10:22, schrieb Thomas de Groot:
> On 26-10-2014 9:54, Stephen wrote:
>> On 26/10/2014 08:13, Thomas de Groot wrote:
>>> But, it is the reason why on goggles, the baseline /is/ extended.
>>> Consider also those military stereoscopic viewers. From my own
>>> experience with stereoscopic viewers used for aerial photographs,
>>> extending the baseline makes objects appear like billboards.
>>
>> Would that not be more of the "telescopic" effect of the lens's used?
>>
> Hmm... I am not sure how much the lens influences the effect indeed. My
> guts tell me it is the baseline mostly, but then who would rely on his
> guts? :-)
Actually it is a combination of the telescopic effect of the lens -
which enlarges the viewed objects in the horizontal and vertical - and
the /absence/ of (sufficient) baseline extension - which would emphasize
the depth cues accordingly.
For instance, in 20x binoculars I think you'd ideally need a baseline of
around 12m; obviously this would make such devices very impractical to
carry around, so that's not done.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 26-10-2014 14:19, clipka wrote:
> Am 26.10.2014 10:22, schrieb Thomas de Groot:
>> On 26-10-2014 9:54, Stephen wrote:
>>> On 26/10/2014 08:13, Thomas de Groot wrote:
>>>> But, it is the reason why on goggles, the baseline /is/ extended.
>>>> Consider also those military stereoscopic viewers. From my own
>>>> experience with stereoscopic viewers used for aerial photographs,
>>>> extending the baseline makes objects appear like billboards.
>>>
>>> Would that not be more of the "telescopic" effect of the lens's used?
>>>
>> Hmm... I am not sure how much the lens influences the effect indeed. My
>> guts tell me it is the baseline mostly, but then who would rely on his
>> guts? :-)
>
> Actually it is a combination of the telescopic effect of the lens -
> which enlarges the viewed objects in the horizontal and vertical - and
> the /absence/ of (sufficient) baseline extension - which would emphasize
> the depth cues accordingly.
>
> For instance, in 20x binoculars I think you'd ideally need a baseline of
> around 12m; obviously this would make such devices very impractical to
> carry around, so that's not done.
>
...but which is achieved for instance with aerial photographs used for
cartographic photogrammetry, with baselines of hundreds of metres.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Am 26.10.2014 16:12, schrieb Thomas de Groot:
> On 26-10-2014 14:19, clipka wrote:
>> Am 26.10.2014 10:22, schrieb Thomas de Groot:
>>> On 26-10-2014 9:54, Stephen wrote:
>>>> On 26/10/2014 08:13, Thomas de Groot wrote:
>>>>> But, it is the reason why on goggles, the baseline /is/ extended.
>>>>> Consider also those military stereoscopic viewers. From my own
>>>>> experience with stereoscopic viewers used for aerial photographs,
>>>>> extending the baseline makes objects appear like billboards.
>>>>
>>>> Would that not be more of the "telescopic" effect of the lens's used?
>>>>
>>> Hmm... I am not sure how much the lens influences the effect indeed. My
>>> guts tell me it is the baseline mostly, but then who would rely on his
>>> guts? :-)
>>
>> Actually it is a combination of the telescopic effect of the lens -
>> which enlarges the viewed objects in the horizontal and vertical - and
>> the /absence/ of (sufficient) baseline extension - which would emphasize
>> the depth cues accordingly.
>>
>> For instance, in 20x binoculars I think you'd ideally need a baseline of
>> around 12m; obviously this would make such devices very impractical to
>> carry around, so that's not done.
>>
> ....but which is achieved for instance with aerial photographs used for
> cartographic photogrammetry, with baselines of hundreds of metres.
... and probably also far more than 20x zoom, right?
A billboard effect in objects is an indicator that the baseline is still
a compromise.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On 27-10-2014 15:34, clipka wrote:
> Am 26.10.2014 16:12, schrieb Thomas de Groot:
>> On 26-10-2014 14:19, clipka wrote:
>>> Am 26.10.2014 10:22, schrieb Thomas de Groot:
>>>> On 26-10-2014 9:54, Stephen wrote:
>>>>> On 26/10/2014 08:13, Thomas de Groot wrote:
>>>>>> But, it is the reason why on goggles, the baseline /is/ extended.
>>>>>> Consider also those military stereoscopic viewers. From my own
>>>>>> experience with stereoscopic viewers used for aerial photographs,
>>>>>> extending the baseline makes objects appear like billboards.
>>>>>
>>>>> Would that not be more of the "telescopic" effect of the lens's used?
>>>>>
>>>> Hmm... I am not sure how much the lens influences the effect indeed. My
>>>> guts tell me it is the baseline mostly, but then who would rely on his
>>>> guts? :-)
>>>
>>> Actually it is a combination of the telescopic effect of the lens -
>>> which enlarges the viewed objects in the horizontal and vertical - and
>>> the /absence/ of (sufficient) baseline extension - which would emphasize
>>> the depth cues accordingly.
>>>
>>> For instance, in 20x binoculars I think you'd ideally need a baseline of
>>> around 12m; obviously this would make such devices very impractical to
>>> carry around, so that's not done.
>>>
>> ....but which is achieved for instance with aerial photographs used for
>> cartographic photogrammetry, with baselines of hundreds of metres.
>
> ... and probably also far more than 20x zoom, right?
Yes, and I found this:
http://www.nrcan.gc.ca/earth-sciences/geomatics/satellite-imagery-air-photos/air-photos/about-aerial-photography/9687
which explains it pretty well.
Thomas
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
