The principle problem with "classical" digital imaging

Started 2 months ago | Discussions thread
OP FDecker Senior Member • Posts: 1,854
Re: The principle problem with "classical" digital imaging

xpatUSA wrote:

FDecker wrote:

The threads about image quality a pixel level, aliasing, AA filters, Foveon vs. Bayer, show the principle problem of digital photography, the representation of an analog picture in a discrete way, approximated by a regular grid of separated samples.

It doesn't matter how much the MP count is raised. At pixel level, the quality is not great and is getting even more disappointing the higher the pixel count (with a given sensor format). The Foveon, and this is why I really like that approach, is clearly better in this respect because the acuity at pixel level is higher for the reasons we all know here. But that's just a mitigation, not a solution.

The best solution to overcome the underlying problem is clearly oversampling a signal (image) which is limited in spatial resolution by an optical filter (or by other means as shaking the sensor while acquiring the image) so that no aliasing happens and using digital filters and spatial decimation (down-sampling) afterwards to a spatial frequency which is considerably below the optical filter bandwidth. <big snip>

Pardon the snip, Frank.

With reference to down-sampling and in the absence of any references or illustrations, may I point out that down-sampling can introduce aliasing, depending on the original detail frequency compared to the maximum detail frequency (Nyquist) of the final image.


If I understand my reference correctly, an over-sampled image, correctly anti-aliased for the degree of over-sampling, can still alias when down-sampled.

So, does that mean that the "limit in spatial resolution" applied during capture should be correct not for the captured image but instead for the final down-sampled image? If so, how does the camera AA filtering know what the final image size is?

Yes, you are right, down-sampling only works without aliasing is the frequency bandwidth in the original image is lower than the spatial frequency of the final down-sampled image.

My approach doesn't suffer from that because I would target for the output resolution of the camera (not the native raw resolution of the sensor). I would still buy a camera which gives me, let's say, 16 MP output files.  With no aliasing but still high acuity and sharpness at pixel level. The camera which gives me such output could be a 64MP Bayer or a 32 MP Foveon with a fixed AA filter which limits the spatial frequency such that the fixed down-sampling in the camera (for the so-called RAW files) results in no aliasing.

So, you are not buying a 64MP camera. You buy a 16MP camera which internally uses a higher resolution 64 MP sensor with a proper spatial frequency limitation and digital filtering and down-sampling.

If you would need or want to down-sample even more, you have to apply a proper filtering before down-sampling the native output file of 16 MP or use a down-sampling software which takes this into account automatically.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow