Hi Peony,
The four images that you uploaded have the following exposure values:
0.jpg ISO68 F2.8 0.0666667s 6.32EV
1.jpg ISO199 F2.8 0.0666667s 7.87EV
2.jpg ISO200 F2.8 0.125s 6.97EV
3.jpg ISO200 F2.8 0.125s 6.97EV
You need a wider range of exposures to cover the dynamic range of the
scene. Visually, the images appear to have almost the same brightness.
They need to noticeable different. See
http://www.jaloxa.eu/webhdr/demo.shtml
for an example exposure-bracketed sequence.
It is recommended to vary only the exposure time, although in principle,
playing with the ISO sensitivity should work, too. Just make sure the
aperture is fixed, which it is in your example shots.
The last two images (I don't have your original file names) have exactly
the same exposure. Visually, however, one is a little darker than the
other. This can mean that the EXIF information (aperture, ISO, shutter)
is unreliable.
You must also mount your camera/smart phone on a tripod. hdrgen can do
some magic trying to align the frames, but it's better to not rely on
it. For accurate results, a tripod is a must.
You will also need to get hold of a luminance meter against which you
need to calibrate the HDRs. Note that the calibration factor depends on
the white balance setting with which the images were taken.
I don't know how you can control the exposure times on a smartphone, but
my recommendation would be that you borrow a decent camera and get some
experience with HDR photography before attempting to do this on a smart
phone. Alternatively, invest in a second hand Canon PowerShot and put
CHDK on it: http://chdk.wikia.com/wiki/CHDK. Make sure your model is
supported. This will allow you to take exposure-bracketed sequences
without touching the camera. For most PowerShots, it will also give you
the option of saving in RAW format for when you need the results to be
VERY accurate.
Hope this gets you started. Good luck
Axel
> From: Axel Jacobs
Sent: Friday, August 24, 2012 11:16 AM
To: [email protected]
Subject: Re: [HDRI] Photosphere "Cannot solve for response function"
Hi Peony,
if you upload your sequence to
http://www.jaloxa.eu/webhdr/roll-your-own.shtml
I'll have a look over the weekend. However, as Greg pointed out--don't
hold your breath. You might have to use a proper camera if you're
interested in HDRs that are 'as accurate as possible'. What are you
trying to measure?
Cheers
Axel
On 08/23/2012 09:50 AM, Peony Au wrote:
Dear all,
I am trying to create HDRIs using an Android Smartphone. I have taken
six photos ranging from –3 to +3 and have tried to fuse these in
Photosphere, however I am getting a “cannot solve for response function”
message. Does anyone know how I can create a response curve or what I am
missing to create the response curve? I can create a HDR image using the
generic response curve, but for my thesis I would need all the HDRIs to
be as accurate as possible.
_______________________________________________
HDRI mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/hdri
