Panasonic GX series

hexpoll wrote:

WhiteBeard wrote:

Adielle wrote:

WhiteBeard wrote:

Adielle wrote:

WhiteBeard wrote:

Adielle wrote:

WhiteBeard wrote:

Adielle wrote:

Keithpictures wrote:

In its favor, it improves upon the GX8 by losing the anti-alias filter

That's not "in its favor" at all, it's in its detriment. I despise this attitude. I'll take "10% less resolving power" over aliasing artifacts / "moire" any day. Cameras that lack an anti-aliasing filter are absolutely useless for me. I run into those artifacts all the time, especially because of the inability to properly handle the detail from the fabrics of people's clothes.

Your fake "extra-detail" in the form of disgusting aliasing artifacts trumps any slight "resolving power" that your AA-filter-less camera MAY have.

I despise this idiotic trend. Omitting the anti-aliasing filter is one of the most idiotic things anyone can do in an analog to digital sampling product, and it's not even disputable, it's a plain fact.

Cue the BS comments against anti-aliasing.

Just curious if you're basing your rant on 3 year old theory or real experience with the latest cams without AA filtre? My understanding is that the extra processing power of the latest gear somewhat corrects some/most/all aliasing artefacts.

There is no way to "correct" aliasing artifacts. Once aliasing happens, it's there to stay. Any POST-processing does not "correct" it, it only MAY somewhat mask the artifacts, and it will ALWAYS result in a MUCH worse result than you would ever have gotten had there not been aliasing artifacts in the first place. You and most people here don't seem to understand a single thing about the absolute basics of digital sampling, and that's how the camera companies are able to get away with crap like this. The only way to actually reduce aliasing artifacts is to include a low-pass filter, oversampling, or ideally (as has been done in audio for many many years) BOTH. There is no magical "state of the art" "computational photography" BS that can get around the basics of digital processing. I'm sick of this stuff.

...And you didn't answer my question: Is your rant comment based on actual experience with recent cameras or on theory which you might not completely master? Just asking...

Actual experience and more than just that.

Any comparison testing examples?

No, I've never been masochistic enough to do that, and the idea sickens me, so you'll have to do it yourself, or at least look at the countless complaints around forums about artifacts by cameras that lack anti-aliasing. Anyway, I have to go, long day ahead. Do yourself a favor and don't listen to people who claim they've "never seen it", and you definitely don't need to "trust me" either (or anyone else who has at least a basic grasp of digital sampling), you should try stuff yourself, especially because forums are full of an insane amount of misinformation regarding this subject. Later.

Oh, I've seen it all right; I'm a video engineer... It's just that it's still not clear to me if you've seen it on recent (2017/2018) cameras or older copies. Your anger seems to overwhelm your logic. Give us examples, pictures, camera models, etc. Otherwise, you're just ranting and not contributiing.

There is a good current thread on the Photographic Science forum that explains this in detail.

https://www.dpreview.com/forums/post/61997568

Adielle is correct in terms of aliasing. The issue is that, without bringing outside knowledge of the scene, it is impossible to actually fix aliasing artifacts. How do you know that your Moire is not actually a photograph of a Moire pattern? How do you know that leaves shouldn't look jagged and visually bright? It is mathematically impossible to reconstruct the original image after aliasing occurs. If you find a true way around it you can collect your Nobel prize.

Modern algorithms degrade image quality to remove the obvious signs of aliasing, but are not able to correct the subtle signs.

Part way there but the link seems to explain the theory of sampling and aliasing but I have not seen any pragmatic (read engineering) workarounds discussed or examples shown, though I admit to being too lazy to read all the posts. As for your statement that it is "impossible to fix aliasing artefact", I beg to disagree but only partly. I fully admit that it would be quite a feat to fit the processing power to recognize all scenes requiring correction in a present-day compact camera but I seem to recall that some primitive examples of scene recognition were present in high-end Sony TVs of ten years ago. With recent AI development, I propose that "some" amount of AA post-processing should be possible depending on cost limitations.  The proof being in the pudding, what I would like to see are real-life pics taken in typical moiré inducing situations between cameras with/without AA filtres at comparable resolution, such as GX8 vs GX9 for example.