 |
|
| Science Forum Index » Optics Forum » Measuring reflectance spectra... |
|
Page 1 of 2 Goto page 1, 2 Next |
|
| Author |
Message |
| I.N. Galidakis... |
 Posted: Tue Oct 06, 2009 11:37 am |
|
|
|
Guest
|
I've been asked to record some reflectance spectra for use in a scene renderer.
I've recorded actual spectra before, but this one has me a little stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about gonioreflectometry, is
that I have this hunch that both source and reflection must be calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and reflected light
source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as the
source) its reflectance against a known standard with non-specular reflection
versus the sample for reflectance?
Many thanks,
--
Ioannis |
|
|
| Back to top |
|
|
|
| whit3rd... |
| Posted: Tue Oct 06, 2009 11:37 am |
|
|
|
Guest
|
On Oct 6, 10:37 am, "I.N. Galidakis" <morph... at (no spam) olympus.mons> wrote:
[quote:9e355e21e1]I've been asked to record some reflectance spectra for use in a scene renderer.
[/quote:9e355e21e1]
One of the things that happens in reflection (off any nonconductive
surface)
is polarization of the reflected light. Are you going to use
unpolarized
detectors and measure averages? The polarization effect is angle
dependent, too, of course. |
|
|
| Back to top |
|
|
|
| boxman... |
| Posted: Tue Oct 06, 2009 1:36 pm |
|
|
|
Guest
|
I.N. Galidakis wrote:
[quote:c546875036]I've been asked to record some reflectance spectra for use in a scene renderer.
I've recorded actual spectra before, but this one has me a little stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about gonioreflectometry, is
that I have this hunch that both source and reflection must be calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and reflected light
source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as the
source) its reflectance against a known standard with non-specular reflection
versus the sample for reflectance?
Many thanks,
[/quote:c546875036]
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there. |
|
|
| Back to top |
|
|
|
| I.N. Galidakis... |
| Posted: Tue Oct 06, 2009 1:51 pm |
|
|
|
Guest
|
boxman wrote:
[quote:83ca22196e]I.N. Galidakis wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
I've recorded actual spectra before, but this one has me a little stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about gonioreflectometry, is
that I have this hunch that both source and reflection must be calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and reflected
light source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as the
source) its reflectance against a known standard with non-specular reflection
versus the sample for reflectance?
Many thanks,
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there.
[/quote:83ca22196e]
Ah! Excellent! One reflection standard, many reflection samples!
Many thanks!
--
Ioannis |
|
|
| Back to top |
|
|
|
| I.N. Galidakis... |
| Posted: Tue Oct 06, 2009 2:28 pm |
|
|
|
Guest
|
whit3rd wrote:
[quote:9c42dfd12f]On Oct 6, 10:37 am, "I.N. Galidakis" <morph... at (no spam) olympus.mons> wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
One of the things that happens in reflection (off any nonconductive
surface)
is polarization of the reflected light. Are you going to use
unpolarized
detectors and measure averages? The polarization effect is angle
dependent, too, of course.
[/quote:9c42dfd12f]
I know. No. I simply plan on measuring at different angles, unless the designer
of the experiment specifies otherwise.
I will simply specify the angle in the end. One sample, at, say 5 degree angle
increments for a total of 90/5 = 18 spectra per sample for strongly "glossy"
samples.
I haven't gotten the specs yet for totally diffuse samples. I suspect I will
have to do only a couple of angles for such samples.
Thanks,
--
Ioannis |
|
|
| Back to top |
|
|
|
| JB... |
| Posted: Tue Oct 06, 2009 2:31 pm |
|
|
|
Guest
|
"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254858706.356536 at (no spam) athprx04...
[quote:213e225341]boxman wrote:
I.N. Galidakis wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
I've recorded actual spectra before, but this one has me a little
stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct
the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about
gonioreflectometry, is
that I have this hunch that both source and reflection must be
calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's
obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and
reflected
light source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source
to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as
the
source) its reflectance against a known standard with non-specular
reflection
versus the sample for reflectance?
Many thanks,
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there.
Ah! Excellent! One reflection standard, many reflection samples!
Use a target coated with 'Spectraflect' (from Labsphere). This is a suitable[/quote:213e225341]
spectral reflectance material standard for this type of measurement.
see: http://www.labsphere.com/data/userFiles/Spectraflect%20Datasheet_1.pdf
You may even be able to find someone who can coat your target with this
material.
A solid piece of Spectralon is even better but $$$$$$$$$$$$$$!
regards,
JB |
|
|
| Back to top |
|
|
|
| I.N. Galidakis... |
| Posted: Tue Oct 06, 2009 2:41 pm |
|
|
|
Guest
|
JB wrote:
[quote:64b3944436]"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254858706.356536 at (no spam) athprx04...
boxman wrote:
I.N. Galidakis wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
I've recorded actual spectra before, but this one has me a little
stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct
the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about
gonioreflectometry, is
that I have this hunch that both source and reflection must be
calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's
obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and
reflected
light source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source
to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as
the
source) its reflectance against a known standard with non-specular
reflection
versus the sample for reflectance?
Many thanks,
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there.
Ah! Excellent! One reflection standard, many reflection samples!
Use a target coated with 'Spectraflect' (from Labsphere). This is a suitable
spectral reflectance material standard for this type of measurement.
see: http://www.labsphere.com/data/userFiles/Spectraflect%20Datasheet_1.pdf
You may even be able to find someone who can coat your target with this
material.
A solid piece of Spectralon is even better but $$$$$$$$$$$$$$!
[/quote:64b3944436]
Many thanks, John. I will send the spec to the designer. Hopefully he can get
hold of some.
[quote:64b3944436]regards,
JB
--[/quote:64b3944436]
Ioannis |
|
|
| Back to top |
|
|
|
| anorton... |
| Posted: Tue Oct 06, 2009 4:26 pm |
|
|
|
Guest
|
"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254860923.782672 at (no spam) athprx04...
[quote:c82c0f4305]whit3rd wrote:
On Oct 6, 10:37 am, "I.N. Galidakis" <morph... at (no spam) olympus.mons> wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
One of the things that happens in reflection (off any nonconductive
surface)
is polarization of the reflected light. Are you going to use
unpolarized
detectors and measure averages? The polarization effect is angle
dependent, too, of course.
I know. No. I simply plan on measuring at different angles, unless the
designer
of the experiment specifies otherwise.
I will simply specify the angle in the end. One sample, at, say 5 degree
angle
increments for a total of 90/5 = 18 spectra per sample for strongly
"glossy"
samples.
I haven't gotten the specs yet for totally diffuse samples. I suspect I
will
have to do only a couple of angles for such samples.
Thanks,
--
Ioannis
[/quote:c82c0f4305]
Trying to accurately measure reflectance versus multiple discrete angles on
a diffuse or partly diffuse sample can be deceptively difficult. Your
results will depend strongly on the NA of the illumination light, the NA of
the collection optics, and relatives sizes of the illuminated area on the
sample compared to the area "seen" by the collection optics. For example,
if you measure a perfect lambertian diffuser, and the area seen by the
collection optics is much smaller than the illuminated area, then you will
see no variation in signal versus angle. If it is the other way around, you
will see the classic cosine curve.
As the previous poster mentioned, polarization can be a big issue on even
slightly glossy surfaces, but mainly if the illumination is at a high angle
and the reflected light is being sampled near the specular angle.
See the guide on the Labsphere website for more common techniques of
measuring diffuse reflectance.
http://www.labsphere.com/data/userFiles/Reflectance%20Spectroscopy%20Techguide.pdf
NIST publishes a copiously detailed guide to radiometry, but unfortunately
it does not cover any specifics of reflectometry.
http://physics.nist.gov/Divisions/Div844/manual/studymanual.html
--
Adam Norton
Norton Engineered Optics
www.nortonoptics.com
(Remove antispam feature before replying) |
|
|
| Back to top |
|
|
|
| I.N. Galidakis... |
| Posted: Tue Oct 06, 2009 6:20 pm |
|
|
|
Guest
|
anorton wrote:
[quote:932f683e19]"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254860923.782672 at (no spam) athprx04...
whit3rd wrote:
On Oct 6, 10:37 am, "I.N. Galidakis" <morph... at (no spam) olympus.mons> wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
One of the things that happens in reflection (off any nonconductive
surface)
is polarization of the reflected light. Are you going to use
unpolarized
detectors and measure averages? The polarization effect is angle
dependent, too, of course.
I know. No. I simply plan on measuring at different angles, unless the
designer
of the experiment specifies otherwise.
I will simply specify the angle in the end. One sample, at, say 5 degree
angle
increments for a total of 90/5 = 18 spectra per sample for strongly
"glossy"
samples.
I haven't gotten the specs yet for totally diffuse samples. I suspect I
will
have to do only a couple of angles for such samples.
Thanks,
--
Ioannis
Trying to accurately measure reflectance versus multiple discrete angles on
a diffuse or partly diffuse sample can be deceptively difficult. Your
results will depend strongly on the NA of the illumination light, the NA of
the collection optics, and relatives sizes of the illuminated area on the
sample compared to the area "seen" by the collection optics. For example,
if you measure a perfect lambertian diffuser, and the area seen by the
collection optics is much smaller than the illuminated area, then you will
see no variation in signal versus angle. If it is the other way around, you
will see the classic cosine curve.
As the previous poster mentioned, polarization can be a big issue on even
slightly glossy surfaces, but mainly if the illumination is at a high angle
and the reflected light is being sampled near the specular angle.
[/quote:932f683e19]
It'd estimate it's probably worse than what you say. According to Murphy's Law,
I expect to detect no variation at all, without any polarizers on very diffuse
samples. I am not even sure my equipment is sensitive enough to detect ANY
variation at ANY angle.
I will probably tackle two typical angles first on glossy and diffuse, see what
kind and how intense a variation I get and then decide what to do based on what
I get.
I understand that polarization can be a problem. I do have the liberty of
placing a polarizer in front of my optics and maybe then - if I understand the
situation at all - get a slightly bigger variation than without the polarizer(?)
But in this case, I am not sure these qualify! The renderer program that needs
the samples doesn't render scenes based on the assumption that the viewer is
wearing polaroids :-)
[quote:932f683e19]See the guide on the Labsphere website for more common techniques of
measuring diffuse reflectance.
http://www.labsphere.com/data/userFiles/Reflectance%20Spectroscopy%20Techguide.pdf
NIST publishes a copiously detailed guide to radiometry, but unfortunately
it does not cover any specifics of reflectometry.
http://physics.nist.gov/Divisions/Div844/manual/studymanual.html
[/quote:932f683e19]
Many thanks. I will take a look at these and try to extract any pertinent info.
But if the variation is a simple cosine, why can't one get just ONE sample (say
the normal) and then multiply that with the cosine of the angle?
PS: what do you mean by "NA" in your post above? Normal Angle?
Thanks,
--
Ioannis |
|
|
| Back to top |
|
|
|
| Phil Hobbs... |
| Posted: Tue Oct 06, 2009 7:31 pm |
|
|
|
Guest
|
JB wrote:
[quote:68ee1d66ff]"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254858706.356536 at (no spam) athprx04...
boxman wrote:
I.N. Galidakis wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
I've recorded actual spectra before, but this one has me a little
stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct
the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about
gonioreflectometry, is
that I have this hunch that both source and reflection must be
calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's
obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and
reflected
light source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source
to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as
the
source) its reflectance against a known standard with non-specular
reflection
versus the sample for reflectance?
Many thanks,
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there.
Ah! Excellent! One reflection standard, many reflection samples!
Use a target coated with 'Spectraflect' (from Labsphere). This is a suitable
spectral reflectance material standard for this type of measurement.
see: http://www.labsphere.com/data/userFiles/Spectraflect%20Datasheet_1.pdf
You may even be able to find someone who can coat your target with this
material.
A solid piece of Spectralon is even better but $$$$$$$$$$$$$$!
regards,
JB
[/quote:68ee1d66ff]
Packed MgO powder is significantly better than Spectralon, and quite a
bit cheaper. Of course it doesn't work as well on the top of the
integrating sphere....
Cheers
Phil Hobbs
--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net |
|
|
| Back to top |
|
|
|
| anorton... |
| Posted: Tue Oct 06, 2009 9:41 pm |
|
|
|
Guest
|
"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254874821.964423 at (no spam) athprx04...
[quote:1a9becac5a]anorton wrote:
"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254860923.782672 at (no spam) athprx04...
whit3rd wrote:
On Oct 6, 10:37 am, "I.N. Galidakis" <morph... at (no spam) olympus.mons> wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
One of the things that happens in reflection (off any nonconductive
surface)
is polarization of the reflected light. Are you going to use
unpolarized
detectors and measure averages? The polarization effect is angle
dependent, too, of course.
I know. No. I simply plan on measuring at different angles, unless the
designer
of the experiment specifies otherwise.
I will simply specify the angle in the end. One sample, at, say 5 degree
angle
increments for a total of 90/5 = 18 spectra per sample for strongly
"glossy"
samples.
I haven't gotten the specs yet for totally diffuse samples. I suspect I
will
have to do only a couple of angles for such samples.
Thanks,
--
Ioannis
Trying to accurately measure reflectance versus multiple discrete angles
on
a diffuse or partly diffuse sample can be deceptively difficult. Your
results will depend strongly on the NA of the illumination light, the NA
of
the collection optics, and relatives sizes of the illuminated area on the
sample compared to the area "seen" by the collection optics. For
example,
if you measure a perfect lambertian diffuser, and the area seen by the
collection optics is much smaller than the illuminated area, then you
will
see no variation in signal versus angle. If it is the other way around,
you
will see the classic cosine curve.
As the previous poster mentioned, polarization can be a big issue on even
slightly glossy surfaces, but mainly if the illumination is at a high
angle
and the reflected light is being sampled near the specular angle.
It'd estimate it's probably worse than what you say. According to Murphy's
Law,
I expect to detect no variation at all, without any polarizers on very
diffuse
samples. I am not even sure my equipment is sensitive enough to detect ANY
variation at ANY angle.
I will probably tackle two typical angles first on glossy and diffuse, see
what
kind and how intense a variation I get and then decide what to do based on
what
I get.
I understand that polarization can be a problem. I do have the liberty of
placing a polarizer in front of my optics and maybe then - if I understand
the
situation at all - get a slightly bigger variation than without the
polarizer(?)
But in this case, I am not sure these qualify! The renderer program that
needs
the samples doesn't render scenes based on the assumption that the viewer
is
wearing polaroids :-)
See the guide on the Labsphere website for more common techniques of
measuring diffuse reflectance.
http://www.labsphere.com/data/userFiles/Reflectance%20Spectroscopy%20Techguide.pdf
NIST publishes a copiously detailed guide to radiometry, but
unfortunately
it does not cover any specifics of reflectometry.
http://physics.nist.gov/Divisions/Div844/manual/studymanual.html
Many thanks. I will take a look at these and try to extract any pertinent
info.
But if the variation is a simple cosine, why can't one get just ONE sample
(say
the normal) and then multiply that with the cosine of the angle?
PS: what do you mean by "NA" in your post above? Normal Angle?
Thanks,
--
Ioannis
[/quote:1a9becac5a]
If you want to measure randomly polarized light, you need to take pains to
do that. Most spectrometers are partial polarizers.
I presume your rendering program wants data in terms of BRDF (bidirectional
reflectance distribution function), true? Read up on how to make these
measurments and make sure you are measuring the correct quantities needed to
calculate BRDF. You need to know the difference between radiance,
irradiance and radiant intensity. If you want a realistic BRDF for
rendering, it will take a lot of single angle measurements since you have to
vary the incident angle as well as the collection asimuth and altitude
angles. If you search online you will see people have developed a lot of
cute solutions to measure multiple angles at once.
What I mentioned before about the cosine curve applies only to a perfect
diffuser where the reflected intensity follows a cosine curve but the BRDF
is constant. By "NA" I meant "numerical aperture" which is the sine of the
half angle of the cone of light focused on the sample or collected from the
sample.
May I suggest something else. I get the feeling that you are more
interested in measuring the color and reproducing that accurately in the
rendering program rather than reproducing the exact gloss and scatter. If
that is the case you could take an existing BRDF for a white surface of
similar gloss and multiply it by the reflectance of your sample as measured
using an integrating sphere.
--
Adam Norton
Norton Engineered Optics
www.nortonoptics.com
(Remove antispam feature before replying) |
|
|
| Back to top |
|
|
|
| JB... |
| Posted: Wed Oct 07, 2009 1:03 am |
|
|
|
Guest
|
"Phil Hobbs" <pcdhSpamMeSenseless at (no spam) electrooptical.net> wrote in message
news:LOudnfey7PjxclbXnZ2dnUVZ_hZi4p2d at (no spam) supernews.com...
[quote:627406beda]JB wrote:
"I.N. Galidakis" <morpheus at (no spam) olympus.mons> wrote in message
news:1254858706.356536 at (no spam) athprx04...
boxman wrote:
I.N. Galidakis wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
I've recorded actual spectra before, but this one has me a little
stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct
the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about
gonioreflectometry, is
that I have this hunch that both source and reflection must be
calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in
the
thousands and the counts of the reflection are in the tens, there's
obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and
reflected
light source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the
source to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting
as the
source) its reflectance against a known standard with non-specular
reflection
versus the sample for reflectance?
Many thanks,
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there.
Ah! Excellent! One reflection standard, many reflection samples!
Use a target coated with 'Spectraflect' (from Labsphere). This is a
suitable spectral reflectance material standard for this type of
measurement.
see:
http://www.labsphere.com/data/userFiles/Spectraflect%20Datasheet_1.pdf
You may even be able to find someone who can coat your target with this
material.
A solid piece of Spectralon is even better but $$$$$$$$$$$$$$!
regards,
JB
Packed MgO powder is significantly better than Spectralon, and quite a bit
cheaper. Of course it doesn't work as well on the top of the integrating
sphere....
Noted. Thanks,[/quote:627406beda]
JB |
|
|
| Back to top |
|
|
|
| boxman... |
| Posted: Wed Oct 07, 2009 7:41 am |
|
|
|
Guest
|
I.N. Galidakis wrote:
[quote:fb691e4c06]boxman wrote:
I.N. Galidakis wrote:
I've been asked to record some reflectance spectra for use in a scene
renderer.
I've recorded actual spectra before, but this one has me a little stumped:
I will either use my double-entrance amici spectroscope or a USB mini
spectroscope.
The setup to do that for my amici spectroscope is pretty simple: The
spectroscope has two light entry points, so all I need to do is direct the
source into one and the reflected light source to the other.
The thing that bothers me and I know very little about gonioreflectometry, is
that I have this hunch that both source and reflection must be calibrated for
intensity BEFORE any measurement.
I mean, if the source intensity counts of the source spectrum are in the
thousands and the counts of the reflection are in the tens, there's obviously
going to be a calibration problem with the final data.
Does anyone know how I can calibrate the two sources (source and reflected
light source) for intensity before measurement?
I was thinking about adding a Lambertian diffuser in front of the source to
bring down the source intensity a bit, but I am not sure.
Maybe measure not the source directly, rather measure (still counting as the
source) its reflectance against a known standard with non-specular reflection
versus the sample for reflectance?
Many thanks,
From your description, it seems as if a reflectance standard to
calibrate against would help you get the data you are looking for. For
example if you get a diffuse white reflectance standard that measures
80% diffuse reflection, then when you use your source and detector set
up you will know how many intensity counts equate to 80% reflectivity
and can then determine the reflectivity of your test sample from there.
Ah! Excellent! One reflection standard, many reflection samples!
Many thanks!
[/quote:fb691e4c06]
Here is a link to a reference paper that talks about the same kind of
measurements you have described above. This may be a bit more complex
than what you were looking to do, but should give you some ideas on how
others have done it and a small amount of the theory behind the
measurements. Section 4 briefly describes how they use the relative
sample method for calibration.
http://www.graphics.cornell.edu/~westin/SPIE_5878-29_Gonio.pdf
(Dr. Westin used to post to the sci.engr.lighting group in the past,
perhaps if he still checks in, he could chime in on your problem as well.) |
|
|
| Back to top |
|
|
|
| AES... |
| Posted: Wed Oct 07, 2009 11:10 am |
|
|
|
Guest
|
[quote:a42c7e0b1f]
Packed MgO powder is significantly better than Spectralon, and quite a bit
cheaper. Of course it doesn't work as well on the top of the integrating
sphere....
[/quote:a42c7e0b1f]
You're bringing back half-century-old memories of packing MgO into the
annular space between two concentric lengths of glass tubing, then
standing that structure on end and pounding the lower end on a stone
optical table, or leaving it for while on a vertical shake table, to
densify the powder, leading eventually to a diffuse pump cavity for a
spiral flashlamp pumped ruby laser.
(My intuition was all against this approach -- I was sure that no matter
how tightly the powder was packed, light would scatter down into the
interstices in the power, and ultimately get absorbed before it ever got
scattered back to the surface. Fortunately others knew better.) |
|
|
| Back to top |
|
|
|
| I.N. Galidakis... |
| Posted: Wed Oct 07, 2009 1:45 pm |
|
|
|
Guest
|
anorton wrote:
[snip]
[quote:5d4c5aee5a]If you want to measure randomly polarized light, you need to take pains to
do that. Most spectrometers are partial polarizers.
[/quote:5d4c5aee5a]
Yes.
[quote:5d4c5aee5a]I presume your rendering program wants data in terms of BRDF (bidirectional
reflectance distribution function), true?
[/quote:5d4c5aee5a]
Ideally, yes.
[quote:5d4c5aee5a]Read up on how to make these
measurments and make sure you are measuring the correct quantities needed to
calculate BRDF. You need to know the difference between radiance,
irradiance and radiant intensity. If you want a realistic BRDF for
rendering, it will take a lot of single angle measurements since you have to
vary the incident angle as well as the collection asimuth and altitude
angles. If you search online you will see people have developed a lot of
cute solutions to measure multiple angles at once.
What I mentioned before about the cosine curve applies only to a perfect
diffuser where the reflected intensity follows a cosine curve but the BRDF
is constant. By "NA" I meant "numerical aperture" which is the sine of the
half angle of the cone of light focused on the sample or collected from the
sample.
May I suggest something else. I get the feeling that you are more
interested in measuring the color and reproducing that accurately in the
rendering program rather than reproducing the exact gloss and scatter. If
that is the case you could take an existing BRDF for a white surface of
similar gloss and multiply it by the reflectance of your sample as measured
using an integrating sphere.
[/quote:5d4c5aee5a]
It's not my rendering program, so I wouldn't know if the designer of the program
wants what you say above. I don't write scene renderers. I write musical
renderers :-)
I just collect data.
I will now send the entire thread to the designer of the program and he will
decide which path I should take.
Many thanks to all.
--
Ioannis |
|
|
| Back to top |
|
|
|
|
|
All times are GMT - 5 Hours
The time now is Tue Nov 24, 2009 10:20 am
|
|