We have moved to a new Sailfish OS Forum. Please start new discussions there.
![]() | 1 | initial version | posted 2015-03-14 01:05:47 +0200 |
It would be great to have a high resolution camera for jolla phone.If it's possible to upgrade the present one even better.
It would be great to have a high resolution camera for jolla phone.If it's possible to upgrade the present one even better.
![]() | 3 | No.3 Revision |
It would be great to have a high resolution camera for jolla phone.If it's possible to upgrade the present one even better.
Edit: 16.03.2015 (by MikErk)
I have done some measurements with the camera to determine the real resolution of the imaging system. I used the slanted edge method which is described in the next link. As test chart I used the ISO12233 printed on A4. I took a photo of this A4-chart with the Jolla and cropped into a slanted edge (pretty in the middle of the image):
http://carlesmitja.net/2011/02/06/image-quality-of-photographic-cameras/
Some results from that:
The Edge spread function shows pretty strong over and undershooting (see https://en.wikipedia.org./wiki/Gibbs_phenomenon). This is probably due to an over-sharpening in post-processing. The slope of the edge itself seems pretty flat. This can be a artefact of a low resolution test chart or some other thing which I will have look on.
The line spread function looks like a triangle function with a spread of about 5pixels. So by the camera a perfect line (f.e. only one black pixel line) is smeared out over some 2-3 pixels.
Due to the over-sharpenening the MTF shows an overshooting for low frequencies. The 50%-MTF lies only at around 0.25 Cycles per Pixel. Note that an ideal system would be at 0.5 Cycles per Pixel. So the camera shows only a resolution of about half its nominal one.
Further things to do:
Make a better measurement with averaging many frames and eventually using a better target
Make measurements with other smartphones or even DSLRs and compare the results
![]() | 4 | No.4 Revision |
It would be great to have a high resolution camera for jolla phone.If it's possible to upgrade the present one even better.
Edit: 16.03.2015 (by MikErk)
I have done some measurements with the camera to determine the real resolution of the imaging system. I used the slanted edge method which is described in the next link. As test chart I used the ISO12233 printed on A4. I took a photo of this A4-chart with the Jolla and cropped into a slanted edge (pretty in the middle of the image):
http://carlesmitja.net/2011/02/06/image-quality-of-photographic-cameras/
Some results from that:
The Edge spread function shows pretty strong over and undershooting (see https://en.wikipedia.org./wiki/Gibbs_phenomenon). This is probably due to an over-sharpening in post-processing. The slope of the edge itself seems pretty flat. This can be a artefact of a low resolution test chart or some other thing which I will have look on.
The line spread function looks like a triangle function with a spread of about 5pixels. So by the camera a perfect line (f.e. only one black pixel line) is smeared out over some 2-3 pixels.
Due to the over-sharpenening the MTF shows an overshooting for low frequencies. The 50%-MTF lies only at around 0.25 Cycles per Pixel. Note that an ideal system would be at 0.5 Cycles per Pixel. So the camera shows only a resolution of about half its nominal one.
Further things to do:
Make a better measurement with averaging many frames and eventually using a better target
Make measurements with other smartphones or even DSLRs and compare the results
Some conclusions:
The optical system can't resolve the 8MP from the sensor. Effectively we have way less resolution.
The image processing software is already trying to compensate for the bad resolution and makes it worse by over-sharpening the image. It little less might give some better results.
![]() | 5 | No.5 Revision |
It would be great to have a high resolution camera for jolla phone.If it's possible to upgrade the present one even better.
Edit: 16.03.2015 (by MikErk)
I have done some measurements with the camera to determine the real resolution of the imaging system. I used the slanted edge method which is described in the next link. As test chart I used the ISO12233 printed on A4. I took a photo of this A4-chart with the Jolla and cropped into a slanted edge (pretty in the middle of the image):
http://carlesmitja.net/2011/02/06/image-quality-of-photographic-cameras/
Some results from that:
The Edge spread function shows pretty strong over and undershooting (see https://en.wikipedia.org./wiki/Gibbs_phenomenon). This is probably due to an over-sharpening in post-processing. The slope of the edge itself seems pretty flat. This can be a artefact of a low resolution test chart or some other thing which I will have look on.
The line spread function looks like a triangle function with a spread of about 5pixels. So by the camera a perfect line (f.e. only one black pixel line) is smeared out over some 2-3 pixels.
Due to the over-sharpenening the MTF shows an overshooting for low frequencies. The 50%-MTF lies only at around 0.25 Cycles per Pixel. Note that an ideal system would be at 0.5 Cycles per Pixel. So the camera shows only a resolution of about half its nominal one.
Further things to do:
Make a better measurement with averaging many frames and eventually using a better target
Make measurements with other smartphones or even DSLRs and compare the results
Some conclusions:
The optical system can't resolve the 8MP from the sensor. Effectively we have way less resolution.
The image processing software is already trying to compensate for the bad resolution and makes it worse by over-sharpening the image. It A little less might give some better results.