I have been doing some more development with my HDR lightdome system.  To speed
up parsing, the current system 'pixelizes' the HDR to do the calculations
according to the user setting, rather than sampling every pixel.  (e.g for a
setting of 200, it would pixelize it into 400x200).  However, this meant that
small, localized light sources could get missed, so the only choice was to
increase the pixelization level, which exponentially increases parse time.  I
have been experimenting with adding in a subsampling system with some good
results.
If first try to establish a general lighting profile of the base pixelization,
which is based on a power curve passing through 3 values: <0,0>, <0.5, A=total
average colour of the pixeled image>, <1, M=maximum sampled colour of the
image>.   From these I get the exponential power 'P' for the curve y=Mx^P, for a
general distribution curve.  Then I set a sub-sample threshold as a % of the
maximum value (e.g. x=0.9=90%).  For any pixel sample that is greater than the
corresponding colour (y), subsampling gets initiated.  Essentially the sampled
'pixel' is divided into 3x3 and the 8 new samples are taken in the immediate
vicinity (the 9th sample is the center which is already known).  If any of these
values is greater than the previously sampled value, then it is substituted in,
and that sub-pixel is then further sub-sampled in the same manner, and so on.  I
have added in two recursion limiters 1) variance value, if the previous sample
is within the variance of the new sample (eg. 0.99), thne recursion is
terminated. 2) number of recursions, if the recursion limit is reached, it is
terminated.

In the attached image, the HDR used for the lighting has a very bright but small
sun.    Based on manually checking, the maximum pixel value in the HDR is
~10,294.  For the top image, a pixel level of 200 is used, the maximum pixel
value found is 1,689 (~16% of the actual max).  In the bottom image, the
sub-sample threshold is set to 0.9, the susbsample variance is set to 0.99, and
the recursion limit is set to 5.  The parse (not including render) time for the
top image was 26.25s.  For the bottom image, the parse time was 28.75s.  To get
the equivalent pixelization level would mean setting the value at 48,600 instead
of 200, and parsing (if even possible) would be expected to take at least 60,000
times as long (3^5^2=59,049) or about 18 days!

Post a reply to this message

Attachments:
Download 'lightdome-sstest.jpg' (476 KB)

Preview of image 'lightdome-sstest.jpg'

The top image actually looks "better" because there is more contrast and  
the colours are more vivid.
The bottom image looks slightly washed out - there seems to be too much  
light. But it does look more like an actual photgraph compared to the top  
image.

May I ask where you got the hdr image from? I quite like it :)

-- 
-Nekar Xenos-

"The spoon is not real"

Post a reply to this message

"Nekar Xenos" <nek### [at] gmailcom> wrote:
> The top image actually looks "better" because there is more contrast and
> the colours are more vivid.
> The bottom image looks slightly washed out - there seems to be too much
> light. But it does look more like an actual photgraph compared to the top
> image.
>
> May I ask where you got the hdr image from? I quite like it :)
>
> --
> -Nekar Xenos-
>
> "The spoon is not real"

I may need to tweak the light levels, but the point is there is direct sunlight
in the HDR which should capture a noticeable shadow present in the second, but
not in the first.

This HDR is one of the few I have found that have a full brightness sun in it vs
one that crops out at a lower value.  This particular HDR is from Bernhardt Vogl
(http://dativ.at/lightprobes/index.html) Direct Evening Sun - Schoenbrunn
Vienna.

-tgq

Post a reply to this message

On 02/19/2011 03:03 AM, Trevor G Quayle wrote:
> I may need to tweak the light levels, but the point is there is direct sunlight
> in the HDR which should capture a noticeable shadow present in the second, but
> not in the first.

ah ... thanks for pointing that out as I didn't pick that out right
away. maybe a difference image (like that used in radiosity tutorial)
might be helpful when trying to illustrate the effect of your technique.

nice subject material btw ... that's the GT40's (ferrari killer) younger
cousin correct, and at 150K+ per copy I'll just have to keep dreaming!

Post a reply to this message

"Trevor G Quayle" <Tin### [at] hotmailcom> wrote:

> This HDR is one of the few I have found that have a full brightness sun in it vs
> one that crops out at a lower value.  This particular HDR is from Bernhardt Vogl
> (http://dativ.at/lightprobes/index.html) Direct Evening Sun - Schoenbrunn
> Vienna.

Feel free to give it a test with one of mine:
http://s3.edouard.info/te-atatu-sunset-1.hdr

Overcoming the Bright Point Lightsource problem is one of the biggest issues
I've had with light probes, so it's fantastic to see your work on it!

(My hack was to delete the sun from the image, generate a light dome, and then
manually added a light for the sun in by hand. Worked well, but I only ever did
it for that particular HDR...)

Cheers,
Edouard.

Post a reply to this message

On Sat, 19 Feb 2011 09:03:42 +0200, Trevor G Quayle  
<Tin### [at] hotmailcom> wrote:


> This HDR is one of the few I have found that have a full brightness sun  
> in it vs
> one that crops out at a lower value.  This particular HDR is from  
> Bernhardt Vogl
> (http://dativ.at/lightprobes/index.html) Direct Evening Sun - Schoenbrunn
> Vienna.
>
> -tgq
>
>
Thanks :)

-- 
-Nekar Xenos-

"The spoon is not real"

Post a reply to this message

"Trevor G Quayle" <Tin### [at] hotmailcom> wrote:
> sun.    Based on manually checking, the maximum pixel value in the HDR is
> ~10,294.  For the top image, a pixel level of 200 is used, the maximum pixel
> value found is 1,689 (~16% of the actual max).  In the bottom image, the
> sub-sample threshold is set to 0.9, the susbsample variance is set to 0.99, and
> the recursion limit is set to 5.  The parse (not including render) time for the
> top image was 26.25s.  For the bottom image, the parse time was 28.75s.  To get
> the equivalent pixelization level would mean setting the value at 48,600 instead
> of 200, and parsing (if even possible) would be expected to take at least 60,000
> times as long (3^5^2=59,049) or about 18 days!

I realized that I was a bit over zealous here with the recursion level as the
HDR image is only 1024pixels high.  A recursion level of 2 should suffice
(200*3*3=1800).  This would have still resulted in an equivalent increase in
parse time of 81X using the original method.

-tgq

Post a reply to this message

"Edouard" <pov### [at] edouardinfo> wrote:
> "Trevor G Quayle" <Tin### [at] hotmailcom> wrote:
>
> > This HDR is one of the few I have found that have a full brightness sun in it vs
> > one that crops out at a lower value.  This particular HDR is from Bernhardt Vogl
> > (http://dativ.at/lightprobes/index.html) Direct Evening Sun - Schoenbrunn
> > Vienna.
>
> Feel free to give it a test with one of mine:
> http://s3.edouard.info/te-atatu-sunset-1.hdr
>
> Overcoming the Bright Point Lightsource problem is one of the biggest issues
> I've had with light probes, so it's fantastic to see your work on it!
>
> (My hack was to delete the sun from the image, generate a light dome, and then
> manually added a light for the sun in by hand. Worked well, but I only ever did
> it for that particular HDR...)
>
> Cheers,
> Edouard.

Thanks.  Yes, this is something I've thought about for a while and finally
decided to give adaptive subsampling a try.  It's likely not perfect, but it
looks like it may be helpful in a lot of cases.

I'll have a look at your HDR. Looks, good, nice and clean.  How did you make it?
 Mirror ball or panoramic stitching?  It looks like you may not have quite the
dynamic range to capture the sun level as it flattens out at about 5 stops.
This is one of the most difficult things for creating HDR with direct sunlight,
and I have found very few that have the true full dynamic level of the sun.  One
thing I would suggest is to always take a photo at your fastest shutter speed as
part of the bracketing to capture as high as you can.  I will also try with my
dynamic range rebuilding option in my lighting which can recover or fake it
fairly well at times.  Do you have any other HDRs available?  I always liek to
find new ones that may have different interesting lighting effects.

-tgq

Post a reply to this message

Jim Holsenback <jho### [at] povrayorg> wrote:
> On 02/19/2011 03:03 AM, Trevor G Quayle wrote:
> > I may need to tweak the light levels, but the point is there is direct sunlight
> > in the HDR which should capture a noticeable shadow present in the second, but
> > not in the first.
>
> ah ... thanks for pointing that out as I didn't pick that out right
> away. maybe a difference image (like that used in radiosity tutorial)
> might be helpful when trying to illustrate the effect of your technique.
>
> nice subject material btw ... that's the GT40's (ferrari killer) younger
> cousin correct, and at 150K+ per copy I'll just have to keep dreaming!

Yes, 2005 Ford GT.  If I had $150k to throw around...

-tgq

Post a reply to this message

"Trevor G Quayle" <Tin### [at] hotmailcom> wrote:

> I'll have a look at your HDR. Looks, good, nice and clean.  How did you make it?
>  Mirror ball or panoramic stitching?  It looks like you may not have quite the
> dynamic range to capture the sun level as it flattens out at about 5 stops.
> This is one of the most difficult things for creating HDR with direct sunlight,
> and I have found very few that have the true full dynamic level of the sun.  One
> thing I would suggest is to always take a photo at your fastest shutter speed as
> part of the bracketing to capture as high as you can.  I will also try with my
> dynamic range rebuilding option in my lighting which can recover or fake it
> fairly well at times.  Do you have any other HDRs available?  I always liek to
> find new ones that may have different interesting lighting effects.

The source images should have captured the range OK; I used +2 to -7 as the
exposure range (so 10 stops in total). I think there are two causes for it not
being as higher dynamic range as the dativ.at one: 1) it's right at sunset, so
there is less light from the sun than normal, and secondly, the source image is
quite small, so there a single pixel probably covers a larger area than the
actual sun, thus "diluting" its brightness a bit.

I've got some more light probes I've done myself with a mirror ball. I'll get
around to organising then and putting then on my website just as soon as I get
around to getting my website :-)

Cheers,
Edouard.

Post a reply to this message