1. Reminder: Please use our affiliate links to get to your favorite stores for holiday shopping!

No Anti Alias Filtered Cameras MUCH Sharper

Discussion in 'This or That? (MFT only)' started by stripedrex, Jan 6, 2014.

  1. stripedrex

    stripedrex Do or do not. There is no try.

    373
    Jun 8, 2012
    Long Island, NY
    Alex
    So I don't consider myself a pixel peeper but I recently had a short stint with a Sony A7R which I didn't keep because I just found the EM5 way more usable from a day to day perspective and felt the gains in sharpness at this time isn't worth the extra $. I'm waiting for the Sony FF to mature a bit and I'm likely to switch over, probably later. Anyhow I posted up some pictures with taking snaps with the Sony 35mm F2.8 and my wife was immediately blown away. She said she's never seen my pix look so sharp and this is on a regular 1080p screen. I took a look and just felt maybe it's just the pop the system gives because of the better dynamic range and zeiss look but upon looking at examples of others I clearly notice the a7R samples around flickr and other sites MUCH sharper than the A7 even not blown up!? Knowing the lack of AA being a major selling point I looked at other cameras that don't have an AA filter and notice that similar CRISP pop from the images when focused well using good glass. I figured mathematically (16mp+ resolution down sized to 1080) you shouldn't notice it but man even 1/4 of the screen the sharpness IS noticeable. All of the cameras with AA removed look WAY better. Are you guys out there with good vision seeing the same thing?

    Here are some example pix of cameras with no AA filter:

    Olympus E-M1 - http://www.flickr.com/groups/em1/pool/
    Ricoh GR - http://www.flickr.com/groups/ricohgr/
    Sony a7R - http://www.flickr.com/groups/ricohgr/

    Compare to most of the images to AA filtered cameras below:
    Olympus E-M5 - http://www.flickr.com/groups/e-m5/
    Sony a7 - http://www.flickr.com/groups/sonya7/ (some mixed a7r in there, i bet you'll see the a7r when the images POP)
    Nikon D600 - http://www.flickr.com/groups/d600nikon/

    Here's my Sony a7R set in a dark home with mixed poor lighting and shooting the 35mm f2.8:
    http://www.flickr.com/photos/ageslani/sets/72157639382680723/

    Here's a different day when I brought my e-m5:
    http://www.flickr.com/photos/ageslani/sets/72157638751248986/

    After looking at some of the pix even I took with the 2 cameras it seems I see it also. I know this seems like a David and Goliath but I really didn't suspect noticing a lack of AA sharpness on a computer screen NOT 100% scaled. I take 1000s of pictures a year and starting to think paying a bit extra for a camera with no AA filter may be worth it. I believe my next camera will lack this AA filter.

    Am I crazy are you guys seeing this or knew this already?
     
  2. nickthetasmaniac

    nickthetasmaniac Mu-43 All-Pro

    Jan 11, 2011
    No.

    I always find it odd when people link to Flickr in an attempt to prove an image quality point. At most, it'll demonstrate drastic differences in noise performance and dynamic range (although to be honest I think even this is a stretch with current generation large sensor performance - the difference at 1000px just isn't there).

    All I see in the groups you linked is differences in the photographer's style regarding composition and, more significantly, processing technique.

    The 2-3% (perhaps) resolution difference between AA and no AA on a 36,000,000 pixel sensor? I call bull#@*^ :smile:
     
  3. stripedrex

    stripedrex Do or do not. There is no try.

    373
    Jun 8, 2012
    Long Island, NY
    Alex
    I blame my wife =p she's convinced so I had to look. After all she's the one that has to suffer through the 1000s of pictures I take =p. Looking at my own snaps on lightroom with the a7r they just seem super crisp but i guess 36mp will do that? I certainly notice the Olympus 17mm 1.8 is substantially softer. Even when you see hairs on thie 17mm there's just this soft blur on the edges (i shoot raw). I'm think I probably need to learn some sharpening techniques and avoid the gas this is going to cause.
     
  4. pdk42

    pdk42 One of the "Eh?" team

    Jan 11, 2013
    Leamington Spa, UK
    Nope, I don't see it either! I'm sure no-one's done it, but I'd wager a large bet that a controlled test of images shot with an E-M1 and and E-M5 would be indistinguishable under all but extreme pixel peeking. Flickr uploads (or even large prints) wouldn't show it.
     
  5. nickthetasmaniac

    nickthetasmaniac Mu-43 All-Pro

    Jan 11, 2011
    To be fair, you're comparing a lens criticised (fairly or not) for being a bit soft (the m.ZD), on a 16mp sensor, with one so far lauded as being remarkably sharp (the CZ), on a 36mp sensor. AA filter or not, of course there's going to be a difference...

    Not saying it doesn't make a difference - I've seen a few very critical real world investigations into non-AA sensors (Leica M9 and D800e come to mind) and there definitely does seem to be a difference, which you might even see in a large format print if you have perfect shooting and printing technique. But if it helps your GAS, I honestly think virtually all of the visible difference you'll see in real world samples on the net will be due to other, more significant factors.

    Something most of the pixel-peepers seem to miss is just how much difference individual processing technique makes. Gees, even loading your RAWS using default settings in Lightroom and Aperture is probably going to make more difference than the AA filter...
     
  6. flash

    flash Mu-43 Hall of Famer

    Apr 29, 2010
    1 hour from Sydney Australia.
    Gordon
    When I went from the EM5 to EM1 I immediately noticed a difference. I have yet to compare an A7r to my A7 but I would think there's some sharpness improvement from the lack of filter but more from the 50% more data. My M9 is very sharp at the pixel level compared to my partners 16mp Canons and better than my old 5D2s.

    Gordon
     
  7. stripedrex

    stripedrex Do or do not. There is no try.

    373
    Jun 8, 2012
    Long Island, NY
    Alex
    I went back and looked at the em1 pix, granted it's flickr and granted most are post processed there's definitely more sharpness on the em1 pix that I've ever seen on the any em5 pix. I just went to dpreview's studio comparison there's without a doubt substantially more sharpness on the em1 to em5 on their studio shot both in low light. The bright light studio shot is out of focus a hair on the em1 =(. Downloading the raws it seems noticeable to me. Going to do more homework and try to grab my buddy's em1 when he gets it to compare. If I notice the kind of difference I think I see with other's shots I need to put 'no aa filter' high on my list because if it's the cause of the pro looking pop with less post process work I think I'd pay extra for it. It's a subtle thing that I now believe makes a difference.
     
  8. scott2hot

    scott2hot Mu-43 Veteran

    318
    Aug 27, 2012
    west yorkshire
    scott
    I gotta say the difference between my pentax k5 and k5iis was night and day...even cheapo sigma / tamron super zoom lenses were far sharper with more contrast and more pop...maybe the K5 was just a little off?...but definate there was a serious improvement on the K5iis
     
  9. Jonathan F/2

    Jonathan F/2 Mu-43 Hall of Famer

    Aug 10, 2011
    Los Angeles, CA
    I had a Nikon D800E and D800 and honestly, I couldn't see the difference. Good glass is still the deciding factor in image sharpness as far as I'm concerned.
     
  10. flash

    flash Mu-43 Hall of Famer

    Apr 29, 2010
    1 hour from Sydney Australia.
    Gordon
    The D800e has a AA filter ad then another to counteract the first, whatever that means. I don't see that as a camera with no low pass filter, which by definition, blurs fine detail. The EM1 on the other hand (and the a7r and M9)have no AA filter at all.

    Gordon
     
  11. lightmonkey

    lightmonkey Mu-43 Veteran

    480
    Dec 22, 2013
    unless you had a controlled A-B samples to compare.
    for end picture result, i didnt find the x100s more spectacular than x100. although it must be noted that the former uses completely different sensor array and has higher pixel count (and smaller pixel pitch). so not a good comparison there.

    my RX1 with AA-filter destroys the EM1 without AA-filter with any glass, at any aperture. both as camera-JPG and final results after optimal RAW processing.

    both are far beyond my PQ requirements.

    the only thing ive noticed, using the EM1 (vs EM5 and others) is that the RAW files are more sharpened as a baseline. i.e., it takes less sharpening in post
     
  12. I am somewhat curious to see if there is any tangible benefit to be had from the E-M1 vs the E-M5 due to the removal of the AA filter. For instance, the E-M5 will already display moire in certain circumstances.

    Full image (shot with a Panasonic Leica 25mm f1.4)

    11013000555_8131c131f9_b.



    100% crop (see the aircon fan shrouds)

    OEM5-PA120602-CROP_zpsba682a1f.



    It can also display noticeable aliasing along lines of fine, high contrast detail.

    Full image (shot with a Panasonic Leica 14-50mm f2.8-3.5)

    11099259835_aae41ab04b_b.



    100% crop

    OEM5-PB230710-CROP_zps750beca2.



    Both of these phenomena are typically more likely to occur on a sensor without an AA filter, yet these two examples came from an E-M5.
     
    • Like Like x 2
  13. snkenai

    snkenai Mu-43 Top Veteran

    523
    Sep 5, 2010
    I doubt that grandma, will notice any of the disastrous effects associated with AA/no AA, in the photo of her first grand son's first birthday party. :smile:
     
    • Like Like x 1
  14. christofp

    christofp Mu-43 Regular

    135
    Jul 21, 2012
    Please note two things when comparing E-M1 and E-M5:

    - E-M5 is said to have no or nearly no AA filter (Olympus never agreed on this officially)!
    Anyhow, weak filter or no filter: E-M5 for sure is not a good example for cameras with AA filter.

    - DPReview studio samples: dpreview changed the lens from E-M5 to E-M1: they changed from super sharp 50/f2 to more or less sharp 45/f1.8.
    If you see any difference, then you see the slight weakness of the 45mm lens, not the sharpness of filterless sensor.

    Christof
     
  15. phl0wtography

    phl0wtography Mu-43 Veteran

    227
    Apr 15, 2011
    I noticed the same thing with my E-M5 most recently. That is, heavy moire in a shot that had some canvas (fabric) in the frame. Very noticeable in jpeg, but also in the raw (LR5.3).
    Not here to open a can of worms, but Pekka Potka once did a controlled, yet unscientific test, and came to the conclusion, that the E-M5 doesn't have an AA-filter either. Admittedly, he used a D800E for the test, which is not a genuine sans-AA-camera.
    Against it stands the fact, that E-M5's .ORF-files withstand (and require, I must add) quite heavy sharpening (in the region of 80-100/25/0 - radius dependant on edge frequency - in LR5.3), while the E-M1's already look quite sharp with default LR settings applied. This also reflects Ming Thein's experience, in that he needs only one round of sharpening with E-M1 files lately.
    Either way, the E-M5, and many cameras nowadays with their really weak AA-filters are plenty sharp with the right lenses, it's the enthusiast, and those affine to technology who know a bit more than the average Joe Camera that feel handicapped by a low-pass-filter without reason.

    EDIT: I added a snapshot I took on a recent walk that shows heavy Moiré in the black down jacket. Just noticed that readers like Tapatalk and the like downsample the pics too much to notice Moire. You wanna look at the original files through a browser.

    E-M5 PL25/1.4 1/320@f/4.0, ISO 200
    [​IMG]
    100%-crop
    [​IMG]
     
  16. Just for reference, the two examples I posted above used the same sharpening settings in Lightroom 4.4

    Amount: 65
    Radius: 0.8
    Detail: 25
     
  17. stripedrex

    stripedrex Do or do not. There is no try.

    373
    Jun 8, 2012
    Long Island, NY
    Alex
    While I think this is an extreme example, here's what a stronger AA filter removed can supposedly do (I don't know how valid these sources are):
    http://www.maxmax.com/nikon_d300HR.htm

    Sorry to keep pointing back to flickr but I just tested myself on various searches that I know would include cameras mixed with aa and non-aa cameras and I have about a %90 success rate and visually picking out which came from non-aa bodies viewing pix full screen on a 21 inch 1080p monitor. Just the very premise of removing moire by putting a filter in to slightly blur it seems obvious enough it'll affect the image. I'm at a point I don't need convincing it seems manufacturers are not removing aa filter as a gimmick there's a very tangible result taking it out and if my main audience (wife) and myself see it, it's worth the investment because I've made a commitment and taken up photography as a hobby with the goal of capturing life and family moments with the best of my ability (learning) and the best equipment i can comfortably afford (for best preservation). I'm on the bridge where I'll likely make a longer system commitment within 1 to 2 years and know now my next camera must be without an aa filter. Nay say all you want I feel my eyes are not fooling me =p.
     
  18. tosvus

    tosvus Mu-43 Top Veteran

    632
    Jan 4, 2014
    No offense to maxmax, but they sell camera modifications. It is not the first place I would look for good examples. I'm not saying it is intentional, but it could be that either them or the customer (if he took the pictures), did something different to cause it. (shot on different days, not exactly same position, post processing differences?). The other theory would be that on newer cameras the AA filters have been dialed back quite a bit. I have only anecdotal evidence, but I've read posts saying they they removed the AA filter on an older camera, but felt it was not necessary on newer ones.
     
  19. Qiou87

    Qiou87 Mu-43 Regular

    150
    Jul 15, 2013
    Paris, France
    Recently I had my Olympus E-PL3 converted to IR. When they do this conversion, they remove the IR-blocking filter along with the AA-filter, then replace this with the IR-filter blocking everything shorter than 715nm in my case.

    The increase in sharpness is really visible, even at 1920x1200. The old sensor of the E-PL3 had probably a strong AA-filter, which explains why I would spot the difference whereas newer cameras with weaker AA-filters would show less of a difference.

    I do not have a before/after shot but I could provide some 100% magnification shots for the pixel peepers here if anyone's interested - without any added sharpness on the RAW file of course. I used to apply +40 or +50 sharpness in Lightroom but now I actually go below the standard value and set +10 or +15 (on RAW-files). Or course I tested this with the same lens (20mm f/1.7, closed down a little).

    I'm not saying you should try and have the AA-filter of your camera removed though, as I do not really see the point. Added sharpness in my case is just a side effect (I wanted an IR camera, not an overly sharp one) and I believe current editing tools are more than sufficient to add crazy amounts of sharpness for those who think it will make their pictures look better. IMHO a bad picture, even if it is crazy sharp, will remain bad ; similarly I've never heard anyone say "Your picture is amazing but lacks sharpness".
     
  20. Dramaturg

    Dramaturg Mu-43 Top Veteran

    614
    Jun 7, 2013
    Ukraine
    Yevgen
    I do a lot of video shooting and my footage of high contrast details (with E-M5) reveals the same amount of moire as my friend's Nikon D800E does. Even with sharpness set to -2. My conlusion is that even if there is an AA filter at E-M5 it is a very weak one.