Am I oversharpening
in [Digital Cameras]
|
From: Ken Tough on 6 Jun 2005 02:55 george_preddy(a)yahoo.com wrote: >You understand wrong. An AA filter (a very simple blurring filter) is >required by Bayers, but not by any camera with full color photosites >like Foveon or 3-CCD/CMOS. Are you really such a whacko, or just a troll with an incredible amount of time on your hands? Anti-aliasing is required by the digitisation process. Even the odd-ball foveon sensor produces digital output. (I don't know of any analog electronic still cameras). Digitising an image requires sampling by an array of sensors, whether it's in a Bayer pattern or the foveon attempt. The sampling process requires anti-aliasing so that spatial frequency information which exceeds the spatial sampling frequency, is eliminated. If it isn't eliminated, it leads to aliasing artifacts like moire. So even your foveon-based camera has an AA filter. Every time you post, your credibility goes one more delta along the exponential towards 0. -- Ken Tough
From: David J Taylor on 6 Jun 2005 03:09 Ken Tough wrote: > george_preddy(a)yahoo.com wrote: > >> You understand wrong. An AA filter (a very simple blurring filter) >> is required by Bayers, but not by any camera with full color >> photosites like Foveon or 3-CCD/CMOS. > > Are you really such a whacko, or just a troll with an incredible > amount of time on your hands? Experience of his/her previous postings suggests the latter. Perhaps both, though! > Anti-aliasing is required by the digitisation process. Even the > odd-ball foveon sensor produces digital output. (I don't know of > any analog electronic still cameras). > > Digitising an image requires sampling by an array of sensors, whether > it's in a Bayer pattern or the foveon attempt. The sampling process > requires anti-aliasing so that spatial frequency information which > exceeds the spatial sampling frequency, is eliminated. If it isn't > eliminated, it leads to aliasing artifacts like moire. > > So even your foveon-based camera has an AA filter. Apparently, the earlier Sigma SD9 camera relied on poor lens MTF as the only anti-aliasing filter, which did no good for its image "quality", of course. David
From: Bart van der Wolf on 6 Jun 2005 07:11 <george_preddy(a)yahoo.com> wrote in message news:1118034754.368804.231090(a)g49g2000cwa.googlegroups.com... SNIP > You understand wrong. An AA filter (a very simple blurring filter) > is required by Bayers, but not by any camera with full color > photosites like Foveon or 3-CCD/CMOS. Which again demonstrates you ignorance (not that such reinforcement is really needed). ANY regular discrete sampling system will suffer from aliasing artifacts if presented with signals of higher (in the case of sensors, spatial) frequency than half the sampling density. To put it into terms even you could grasp, any regular detail smaller than 2 pixels will lead to aliasing artifacts (larger than actual detail aliases, and with the wrong luminance). It's a physical given, which can only be attenuated/eliminated by pre-filtering the signal to remove the too high frequency content *before* it reaches the sensor. The only 'benefit' of the Foveon sensor array is that the aliasing artifacts are achromatic in reconstruction of the Raw data so they may be a bit less noticable (it won't prevent stairstepped diagonal lines and things like line doubling artifacts). Color fringing with Bayer CFA reconstruction can result from less strict AA-filtering in the camera, combined with design choices in the Raw converter when balancing between color accuracy and detail enhancement. Different Raw converters will give different results, even with post-processing of the same Raw file. Bart
From: Bart van der Wolf on 6 Jun 2005 07:18 "David J Taylor" <david-taylor(a)blueyonder.co.not-this-bit.nor-this-part.uk.invalid> wrote in message news:vESoe.46039$G8.27494(a)text.news.blueyonder.co.uk... SNIP > Apparently, the earlier Sigma SD9 camera relied on poor lens MTF as > the only anti-aliasing filter, which did no good for its image > "quality", of course. And it's successor, the SD10, still didn't have an AA-filter. It did correct some other design shortcomings of the SD9. It improved the performance by adding microlenses to avoid the artifacts caused by the SD9's very small fill factor (it behaved more like a point sampler than an area sampler). Bart
From: Bart van der Wolf on 6 Jun 2005 07:32
<george_preddy(a)yahoo.com> wrote in message news:1118011376.102432.112740(a)g49g2000cwa.googlegroups.com... > > > Bart van der Wolf wrote: >> <george_preddy(a)yahoo.com> wrote in message >> news:1118002917.281821.236530(a)g44g2000cwa.googlegroups.com... >> SNIP of all your inaccuracies and lies >> >> Apparently leaves no text. > > What's the matter? All you contribute to this newsgroup, is negative intelligence. That's what's the matter. Boy, it must suck being you. Bart |