Great info, as always, Kieth. For our reprography and architectural work we had the AA filters removed from our 24MP Nikon sensors for ‘pure’ detail capture. We then used the Nikon 800e that used technology that ‘emulated’ the absence of the AA filter on its 36MP sensor but had horrible moire patterns on high frequency building details and line art hatching patterns. When checking test images in DPReview I noticed that one camera’s sensor didn’t suffer from these artefacts - the 24MP Fujifilm X-Trans sensor with its unique ‘randomised’ RGB sensor filter design. We now use Fujifilm cameras exclusively with 26MP and 40MP sensors. We finally got the ideal solution. Sorry Nikon!
Excellent - it's varied a lot over camera models. There was also an element of 'I just don't need to 'fix' moire' in my choice of the 5Ds as I was doing more architectural work ;-)
@@KeithCooper I agree. Recently I acquired a ‘play camera’ Fujifilm X-A3 with the ‘poor mans’ Bayer filtered sensor but with no AA filter and I can’t believe how sharp and moire free the images are with modern processing software and the latest camera firmware!
Cameras without AA filter do not have "pure" details, that extra sharpness and details is aliasing, all of it, and it's not real sharpness. Of course it might not be a problem for you, but personally I prefer higher resolution sensor with AA filter than lower resolution ones without because real sharpness is always better than aliasing.
Thank you. Interesting take on the R5 vs 5DSR, this may explain why Canon also say that the R6 mark 2 gives more detail at 24mp than the eos R/5Dmkiv at 30mp.
Interesting stuff. The more that I begin to peek into the practical implications of these theoretical musings the more that I realize that its a game of tradeoffs and that a lot of this stuff is all in our heads.
Thanks - this is the reason I suggest trying things out and deciding what matters. Lessons I learned looking at Hi-Fi some 30-40 years ago when I did electronics design as a job and first derided super expensive speaker cables ;-)
"AA filter or not" is one of my recurring rants. And it is not about the AA filter, actually, but about the software we use for raw processing. The real problem is that some raw processing software still may not have been adapted to raw processing images from cameras without AA filter. In which case we may get noisy blurred zones with luminance and colour noise, and that software may cause a loss of detail, and it may generate digital artefacts that become extremely visible with a bit of sharpening. This is a case of very bad AI - not good enough, maybe fast, but that's no excuse if we look at how fast our cameras and their firmware generate very high quality SOOC JPEGs and how little energy these cameras use for that, compared to a 750W powerhouse desktop workstation that we run our Adobe software on.
Yes, RAW processing makes a vast difference as well. I'm still sometimes having to do two versions of a conversion and 'brush in' or out some parts. That said, I'll open up one of my 2003 RAW files and appreciate how much stuff has improved, and will continue to do so ;-) I also put my 'business hat' on and remind myself that no-one will actually notice in the real world. I have a few clients where they might - they pay for the extra work though...
@@KeithCooper - Agreed. It's a lot better today but that was necessary because of the high resolution we have today. And it's facilitated by Moore's Law 20 years, 10 cycles, 2^10=1,024 times more components in a chip per same area. But do try to upsample some of your images in Topaz Gigapixel AI, if only to prepare for a large print where this SW does the upsampling rather than the printer driver. And for better gradation, take an image into Photoshop and convert to 32 bits
Yes, I've quite a bit of stuff [more in written articles than videos] on using GP AI - especially after my printer driver resolution testing for a variety of printers showed up the old myth about 'native resolution' to be just that [with modern printers/drivers]
@@KeithCooper - I searched your website (enhanced by Google) for "printer driver resolution test" (nothing), as well as "gigapixel native" (nothing). My TGPAI point is not about the magic resolution numbers of yonder years, but about the way software upsamples. Adobe Camera Raw gave me a panorama with fluffy coloured balls in the far distance, TGPAI gave me hikers with legs, arms, heads. ACR is really not smart at detail preservation or prevention of detail loss. The recent addition of "AI Denoise" into ACR is great for noise, but not for details. As the devil is in the details, you'll see that ACR does an OK job in some genres and is very bad in others. Will people see that - I'm totally with you on that. Yes they will in large prints, walking up to them. Yes they will, pinch-zooming-in to see themselves. An experienced landscape painter told me 50..60 years ago, the illusion of depth was in losing detail for things farther away. A couple years ago, he pinched-zoomed-in on a photo that he was in, now a very old man. "It's not sharp," he shouted. I had not pixel peeped and he was right.
Ah, yes, the search is not infallible ;-) www.northlight-images.co.uk/driver-settings-and-print-detail/ See also here and linked www.northlight-images.co.uk/upsizing-and-sharpening-for-making-a-print/ The GP AI stuff works well for a lot of upsizing - it's often so much better than older methods.
I understand your views based on your type of work. But what about fashion/clothing photography? Does not having an AA filter cause moiré artifacts on fabrics?
Yes, it's always a potential issue. With any camera, if I'm doing any fine detail work, I'll try and process a few images just to make double sure. This is indeed an occasion where the 5Ds might come out of retirement. I've not had any work with the GFX100S yet where it was an issue, but I'm on the look out for it.
Thank you its good information. I agree, I also do not enjoy math but I had to learn it to get through the Engeener school. But I really enjoy numerical solution of differential equations. If I understand it a High Mpix sensor have less need with the optical LP filter.
@@KeithCooper Your video was focused on the >50MP cameras - how do your comments translate to lower MP sensors? I would have thought that the important factor is the pixel density of the sensor. Would it be reasonable to assume that the smaller pixels for say a 24MP APS-C sensor would handle high-frequency aliasing better than a 24MP full-frame sensor?
Well, 24MP is 50-60MP equivalent [depending which APS-C] so it depends on lenses and subjects. That said, I've little experience with APS-C in current cameras of different makes.
@@KeithCooper I started with a D300 that’s an APS-C 12Mpix sensor and that's 28 Mpix on FF which is my base line. I have FF lenses that work well with that resolution. I like to be able to crop 50% to get the HB Pan format and therefore I need a 56 Mix sensor. Today I have D700 12 Mpix FF, D800E a 36 Mpix FF and I just got D850 a 45 Mpix FF and upgraded a selected few lenses for that resolution. I will finish upgrading my Nikon system with a 400mmF2.8/500mmF4.0 lens from a 300mmf2.8 screw drive AF and then I will probably get the GFX 100S to meet my goal. Nikons newZ cameras are next to as expensive as the GFX, but the inflation is high and I retire in 3 years so???
Hello Keith, thank you for another very interesting video. I have learnt such a lot from watching your channel. I am wondering if you have any thoughts on processing software with your new GFX camera. I have heard it said many times that files from GFX cameras benefit from processing in Capture One. Have you compared Capture One to Lightroom?
Thanks I've not used C1 for many years but Camera Raw+Photoshop [I dislike Lightroom a lot] and DxO Photolab+Photoshop seem to do very well C1 would cost money I don't have to test in depth...
All the major camera makers have good engineers and test equipment/processes, so I trust them to design cameras that can deliver excellent results across the range of uses that we put them through.
I've only noticed moire in a few of my 5DsR images, some bird feathers and animal hair, but I don't shoot a lot of architecture or clothing. The moire tool in Lightroom and Photoshop always fixed it, but I've heard there are some instances where the tool can't fix it. Before I bought the 5DsR, I searched long and hard to see which model to get, and decided that the model with the AA filter negated was the best choice. The examples shown online showed a significant enough difference in sharpness. I never regretted that decision, because I sometimes needed the extra detail, because I printed big, and sometimes had to crop. I've heard that the tinier the pixels are the less moire will show up. So the GFX 100S' even tinier pixels will probably make moire even less evident, if that's true.
A good sign for the GFX 100S is that portrait and fashion photographers seem to rave about how wonderful the 100S is to shoot people with. I haven't heard any complaints about moire with patterns in clothes. I haven't heard any complaints about moire from anyone so far, and I've looked at everything I could find online, video and written, since I have great interest, and am retired, so have the time.
@@KeithCooper It looks like the 100S has no moire problems at all! I just thought to look now at DPReview's Studio Test Scene to check moire. It looks like it's running perfectly again. It had many missing areas after Amazon declared they were shutting it down. Thankfully, we have new owners who hopefully will keep it going. I put in the Fuji 100/100S along side Canon 5DS R, both set to RAW and 100 ISO. If we look at the rulers with the parallel lines going from tight to further apart, the 5DsR has a lot of moire. The Fuji has none! Wow! I was shocked to see none at all, since it does not have an anti-aliasing filter. The black and white photo of the people on the left side of the Studio Test Scene also shows no moire with the 100/100S; though, plenty with the 5DsR. So with the 100S, we have optimal resolution and no moire. So it's a thing of the past for us! So glad! The 100S is standing up to be the dream camera, having hardly any downsides. ________ You said you used the Canon 50mm TS-E in your 5DsR/5Ds test. Bryan Carnathan at "The Digital Picture" does a very good job of testing lenses, and has tested all of the Canon EF lenses. There, the original 50mm TS-E looks like it may be sharp enough to do the test in the center and mid-frame, but not at the periphery, where it's somewhat soft even at f/8. The original 90mm TS-E shows a lot more moire at the center, maybe mainly because the parallel lines in the test chart are closer together. The 90mm is significantly sharper and more contrasty than the 50 at the periphery.
Yes the 50 is not up to the new 90, but neither of us owned the new 90 or similarly good 135 That said we only looked at the centre area, specifically to get as good as we could - enough to decide it wasn't worth £300 extra for a 5DSR ;-)
@@KeithCooper I was talking about the old 90 TS-E, comparing that to the old 50 (if that's the one you used), which looks plenty sharp in the center. The old 50 is the one that's somewhat soft at the periphery. The difference with the AA filter cancelled or not isn't great. For architecture, not having any moire is a great benefit, since it probably would often enough appear with the 5DsR. One of my images was an arctic tern hovering over his/her chick with glorious looking wings, which I had to crop enough so that I barely had enough sharpness with the AA filter defeated in the 5DsR. I only did some wildlife birding with the camera, and found 5 frames per second fast enough. The extra field of view with the GFX 100S would make finding and framing fast and erratic moving birds easier to capture, but the focus tracking probably isn't good enough, especially with an adapted lens. I may try it anyway, especially for the trumpeter swans that don't move as fast and are further away. But my main use will be landscape, macro flowers and some cityscape.
Hi Keith, Thank you for the very informative video. We have PTZ cameras in an event space, the stage has an LED wall behind, so when focused on our talent we have considerable Moire on the LED wall behind unless we soften focus. The PTZ cameras are panasonic aw150we. I believe the newer model has an anti alias filter built in. My question is would a Anti aliasing filter work at the front of the lens or must it be on the sensor? I have a designed a 3D printable filter housing that will clip onto the front of the camera. But wanted to know if this way would work before getting AA filters cut to size. Many thanks
I don't know this applications, but the AA filtering really need to be direct to the sensor. A diffuser on the LED wall would likely work, if not that practical though - it's likely the point source lighting of the LEDs which exacerbates the effect
I am curious about those cameras that support 4x high resolution modes, OM-D EM1 Mark II and on, Lumix G9 and GH6, Lumix S. It's claimed better colour (makes sense), and better resolution without overtaxing the lenses (I'm looking at Canon with its list of lenses not recommended for the 5Ds cameras). The E-M1x and later also do 50 megapixels hand-held, they rely on camera movement from being held. Do they show moire under difficult conditions.
Yes that Canon list for the 5Ds - mostly a marketing creation. I remember looking at it, chuckling, and thought, yes, that will help sell a few lenses and cameras - then I ignored it ;-)
AA filters are a major problem in my opinion, and each individual camera needs to be thoroughly researched on its AA filter alone. My first dslr was a cheap Nikon entry level camera15 years ago which ad 10mp. My 8mp smart phone produced much sharper images. So I got rid of it and when to the first ever 24mp Sony apsc slt and experienced fantastic results. I never went back to Nikon but kept an eye on their cameras, and found that they had some peculiar sharpness issues, which I attributed to their choice of AA filter. After Sony I switched to Pentax for the IBIS astrotracer ability, newer models went AA filterless. My first Pentax was the brilliant K-5 with 16mp that already had a very light AA filter IMO. My following K3 had the AA filter simulator which I've never used really. AA simulator is the IBIS rotating on a sub pixel level to blur the image slightly to stop aliasing. Plus Pentax IBIS does horizon correction, tilt shift mode/ability and the newer versions of GPS-less astrotracer. So when you are in Pentax land with AA filterless and AA simulation mode, its something which I haven't had to deal with. I did buy some Fuji cameras and lenses, one was bayer and the other Xtrans and I didn't feel comfortable going backwards to a camera system that didn't have IBIS and some camera models had AA filters. So I would be skeptical if the Fuji 100s didn't have even a little bit of an AA filter in the UV/Cut filter glued to the sensor. Mitsubishi owns Nikon and Fuji and I feel both brands are subpar in my experience. I think the GFX system is better than their X series, but I would never buy one because I find their lenses are to boring and clinical looking and lack some dimensionality. X-Trans sensors DO have a false colour issue according to a couple of Peta pixel articles. Maybe 100mp is enough to make aliasing go away in images. Or maybe its software. Or maybe its simply not a true AA filterless camera - you know the old false advertising as many companies do regularly. So AA filters aren't an issue when you shoot with Pentax which has all these cool IBIS modes and abilites, and Pentax solved the problem with AA simulator with varying degrees of blur.
I believe that the Mitsubishi companies, including Nikon, are related by ownership, not structured as a hierarchy as are western companies. Imagine a group of Japanese warlords retiring form warlording, and seeking enterprises more suitable to the more peaceful times they could see coming. Imagine there were coalitions similar to European countries of recent centuries for example. So the warlords like the idea of maintaining these friendships by cross investing, so each goes about his business as he always has. One wants to make pencils and other stationary. Another likes shipbuilding, someone else automotive. Nikon is part of this, but for some reason chose not to share the name. Fujitsu has a similar structure, with computers and electronics, whitegoods, tyres but again under a different name. Canon has a more western structure.
Wow! Fuji possibly lying about not having an AA filter, I never would have guessed that was possible. I suppose someone with the technical know how and equipment could check it during a tear-down. It looks like they solved the moire problem in some way, according to DPReview's test chart, which I commented on under my comment here. I mention I got the Canon 5DsR. I never liked how Canon negated the AA filter with another lens system, instead of getting rid of it altogether. I loved the 5DsR, except for the poor dynamic range, and eventually I just got sick of being stuck in 3:2, knowing that the lenses were capable of so much more coverage, which the 100S' larger sensor can do. So I'll be using mostly my EF lenses on the 100S with the Fringer adapter; though, many won't cover the entire sensor. Pulling the rear baffle out of some help immensely. With that removed, the Canon EF 100-400 IS II is a stellar tele on the Fuji.
They are indeed a major problem for some - to others they matter not one jot I fall midway and merely suggest that people don't get too carried away by the protestations of the ultra-pixel peepers, and spend more time looking at the whole image ;-) I genuinely gave it no direct consideration in my purchase of a GFX100S - I did, indirectly, by some of the stuff I photographed when testing... Unfortunately 'clinical and boring' have no transferable meaning from my POV - the lenses I have looked at are excellent - as are some of my creaky old ones, dismissed by some as 'wholly unsuitable' for the 100S :-) :-)
@@KeithCooper We both know from personal experience that many lenses work well with the 5Ds cameras, which have pixels not much larger than the 100S' pixels. So I expect lenses that worked fine on the 5Ds to work fine on the 100S. It seems to me that the lenses that worked well on the 21 MP 5D Mk II seem to work as well on the 5DsR. If they were lousy on the 5DII they are lousy on the 5DsR, and vice versa. I think I understand your crusade against those fixated with pixel peeping, which for many can be a waste of time, and time could be better spent taking images. For some at DPReview forums, for example, many were more focused on winning debates than finding the truth. I suspect some may even be intoxicated while posting. I've been involved in some of those discussions. Some were fruitful, some not, but I learned a lot not just about cameras but people. Some of the commenters are actually mean, and it's more about winning and defending their brand than the truth. Their hobby seems to be about winning debates, not photography. That's unhealthy, and is a pain for the moderators. I really appreciate the technical know-how people who focus on the facts, and who have high standards. I've mostly printed 24 x 36 inch prints, which require getting a lot of things right. For me, it's important to know the specifics in order to produce work that I'm pleased with. I mentioned before that I actually quit taking photos and printing for many years, and am now just getting back into it, partly because of this great oversized sensor 100S and because of the higher reflectivity dye ink prints; though, I'm going to now be limited to 13 inches, which is okay for now, because I'm retired and not working, so have limited funds.
Yes - I do have a genuine interest in all the technical stuff, but it's almost always out of technical curiosity or to support my more general photography. I know what you mean about some on DPR - it's one of the main reasons we decided against having forums on the Northlight Images web site ;-) Also because we'd have enforced a 'real names only' policy [I have a 20+ year policy of mainly ignoring emails where people won't use their real name...]
Great info, as always, Kieth. For our reprography and architectural work we had the AA filters removed from our 24MP Nikon sensors for ‘pure’ detail capture. We then used the Nikon 800e that used technology that ‘emulated’ the absence of the AA filter on its 36MP sensor but had horrible moire patterns on high frequency building details and line art hatching patterns. When checking test images in DPReview I noticed that one camera’s sensor didn’t suffer from these artefacts - the 24MP Fujifilm X-Trans sensor with its unique ‘randomised’ RGB sensor filter design. We now use Fujifilm cameras exclusively with 26MP and 40MP sensors. We finally got the ideal solution. Sorry Nikon!
Excellent - it's varied a lot over camera models. There was also an element of 'I just don't need to 'fix' moire' in my choice of the 5Ds as I was doing more architectural work ;-)
@@KeithCooper I agree. Recently I acquired a ‘play camera’ Fujifilm X-A3 with the ‘poor mans’ Bayer filtered sensor but with no AA filter and I can’t believe how sharp and moire free the images are with modern processing software and the latest camera firmware!
The MF cameras, however, have a standard Bayer pattern.
Cameras without AA filter do not have "pure" details, that extra sharpness and details is aliasing, all of it, and it's not real sharpness. Of course it might not be a problem for you, but personally I prefer higher resolution sensor with AA filter than lower resolution ones without because real sharpness is always better than aliasing.
so you prefer a sensor with a filter that softens the image. We go on real world, printable results not theories! @@gurratell7326
Thank you. Interesting take on the R5 vs 5DSR, this may explain why Canon also say that the R6 mark 2 gives more detail at 24mp than the eos R/5Dmkiv at 30mp.
Thanks - they do a lot of subtle things with these filters, so it's never a simple difference.
Interesting stuff. The more that I begin to peek into the practical implications of these theoretical musings the more that I realize that its a game of tradeoffs and that a lot of this stuff is all in our heads.
Thanks - this is the reason I suggest trying things out and deciding what matters.
Lessons I learned looking at Hi-Fi some 30-40 years ago when I did electronics design as a job and first derided super expensive speaker cables ;-)
Thanks for the extended discussion on the topic.
Thanks - glad it was of interest
"AA filter or not" is one of my recurring rants. And it is not about the AA filter, actually, but about the software we use for raw processing.
The real problem is that some raw processing software still may not have been adapted to raw processing images from cameras without AA filter. In which case we may get noisy blurred zones with luminance and colour noise, and that software may cause a loss of detail, and it may generate digital artefacts that become extremely visible with a bit of sharpening.
This is a case of very bad AI - not good enough, maybe fast, but that's no excuse if we look at how fast our cameras and their firmware generate very high quality SOOC JPEGs and how little energy these cameras use for that, compared to a 750W powerhouse desktop workstation that we run our Adobe software on.
Yes, RAW processing makes a vast difference as well.
I'm still sometimes having to do two versions of a conversion and 'brush in' or out some parts.
That said, I'll open up one of my 2003 RAW files and appreciate how much stuff has improved, and will continue to do so ;-)
I also put my 'business hat' on and remind myself that no-one will actually notice in the real world. I have a few clients where they might - they pay for the extra work though...
@@KeithCooper - Agreed. It's a lot better today but that was necessary because of the high resolution we have today. And it's facilitated by Moore's Law 20 years, 10 cycles, 2^10=1,024 times more components in a chip per same area. But do try to upsample some of your images in Topaz Gigapixel AI, if only to prepare for a large print where this SW does the upsampling rather than the printer driver. And for better gradation, take an image into Photoshop and convert to 32 bits
Yes, I've quite a bit of stuff [more in written articles than videos] on using GP AI - especially after my printer driver resolution testing for a variety of printers showed up the old myth about 'native resolution' to be just that [with modern printers/drivers]
@@KeithCooper - I searched your website (enhanced by Google) for "printer driver resolution test" (nothing), as well as "gigapixel native" (nothing). My TGPAI point is not about the magic resolution numbers of yonder years, but about the way software upsamples. Adobe Camera Raw gave me a panorama with fluffy coloured balls in the far distance, TGPAI gave me hikers with legs, arms, heads. ACR is really not smart at detail preservation or prevention of detail loss. The recent addition of "AI Denoise" into ACR is great for noise, but not for details. As the devil is in the details, you'll see that ACR does an OK job in some genres and is very bad in others.
Will people see that - I'm totally with you on that. Yes they will in large prints, walking up to them. Yes they will, pinch-zooming-in to see themselves. An experienced landscape painter told me 50..60 years ago, the illusion of depth was in losing detail for things farther away. A couple years ago, he pinched-zoomed-in on a photo that he was in, now a very old man. "It's not sharp," he shouted. I had not pixel peeped and he was right.
Ah, yes, the search is not infallible ;-)
www.northlight-images.co.uk/driver-settings-and-print-detail/
See also here and linked
www.northlight-images.co.uk/upsizing-and-sharpening-for-making-a-print/
The GP AI stuff works well for a lot of upsizing - it's often so much better than older methods.
I understand your views based on your type of work.
But what about fashion/clothing photography? Does not having an AA filter cause moiré artifacts on fabrics?
Yes, it's always a potential issue. With any camera, if I'm doing any fine detail work, I'll try and process a few images just to make double sure.
This is indeed an occasion where the 5Ds might come out of retirement. I've not had any work with the GFX100S yet where it was an issue, but I'm on the look out for it.
Thank you its good information. I agree, I also do not enjoy math but I had to learn it to get through the Engeener school. But I really enjoy numerical solution of differential equations. If I understand it a High Mpix sensor have less need with the optical LP filter.
It's high MP and sensor size
@@KeithCooper Your video was focused on the >50MP cameras - how do your comments translate to lower MP sensors? I would have thought that the important factor is the pixel density of the sensor. Would it be reasonable to assume that the smaller pixels for say a 24MP APS-C sensor would handle high-frequency aliasing better than a 24MP full-frame sensor?
Well, 24MP is 50-60MP equivalent [depending which APS-C] so it depends on lenses and subjects.
That said, I've little experience with APS-C in current cameras of different makes.
@@KeithCooper I started with a D300 that’s an APS-C 12Mpix sensor and that's 28 Mpix on FF which is my base line. I have FF lenses that work well with that resolution. I like to be able to crop 50% to get the HB Pan format and therefore I need a 56 Mix sensor. Today I have D700 12 Mpix FF, D800E a 36 Mpix FF and I just got D850 a 45 Mpix FF and upgraded a selected few lenses for that resolution. I will finish upgrading my Nikon system with a 400mmF2.8/500mmF4.0 lens from a 300mmf2.8 screw drive AF and then I will probably get the GFX 100S to meet my goal. Nikons newZ cameras are next to as expensive as the GFX, but the inflation is high and I retire in 3 years so???
Hello Keith, thank you for another very interesting video. I have learnt such a lot from watching your channel. I am wondering if you have any thoughts on processing software with your new GFX camera. I have heard it said many times that files from GFX cameras benefit from processing in Capture One. Have you compared Capture One to Lightroom?
Thanks
I've not used C1 for many years but Camera Raw+Photoshop [I dislike Lightroom a lot] and DxO Photolab+Photoshop seem to do very well
C1 would cost money I don't have to test in depth...
@@KeithCooper Thank you anyway Keith.
All the major camera makers have good engineers and test equipment/processes, so I trust them to design cameras that can deliver excellent results across the range of uses that we put them through.
Yes - it's sometimes a bit more work to decide if it matches what you want to do
@KeithCooper yes, that's always the trick. As one expert once (ok, more than once) said, "it depends. "
I've only noticed moire in a few of my 5DsR images, some bird feathers and animal hair, but I don't shoot a lot of architecture or clothing. The moire tool in Lightroom and Photoshop always fixed it, but I've heard there are some instances where the tool can't fix it.
Before I bought the 5DsR, I searched long and hard to see which model to get, and decided that the model with the AA filter negated was the best choice. The examples shown online showed a significant enough difference in sharpness. I never regretted that decision, because I sometimes needed the extra detail, because I printed big, and sometimes had to crop.
I've heard that the tinier the pixels are the less moire will show up. So the GFX 100S' even tinier pixels will probably make moire even less evident, if that's true.
A good sign for the GFX 100S is that portrait and fashion photographers seem to rave about how wonderful the 100S is to shoot people with. I haven't heard any complaints about moire with patterns in clothes. I haven't heard any complaints about moire from anyone so far, and I've looked at everything I could find online, video and written, since I have great interest, and am retired, so have the time.
I've looked on a few GFX images and not found any significant issues - of course some would say my lenses are no good anyway ;-)
@@KeithCooper It looks like the 100S has no moire problems at all! I just thought to look now at DPReview's Studio Test Scene to check moire. It looks like it's running perfectly again. It had many missing areas after Amazon declared they were shutting it down. Thankfully, we have new owners who hopefully will keep it going.
I put in the Fuji 100/100S along side Canon 5DS R, both set to RAW and 100 ISO. If we look at the rulers with the parallel lines going from tight to further apart, the 5DsR has a lot of moire. The Fuji has none! Wow! I was shocked to see none at all, since it does not have an anti-aliasing filter.
The black and white photo of the people on the left side of the Studio Test Scene also shows no moire with the 100/100S; though, plenty with the 5DsR.
So with the 100S, we have optimal resolution and no moire. So it's a thing of the past for us! So glad!
The 100S is standing up to be the dream camera, having hardly any downsides.
________
You said you used the Canon 50mm TS-E in your 5DsR/5Ds test. Bryan Carnathan at "The Digital Picture" does a very good job of testing lenses, and has tested all of the Canon EF lenses. There, the original 50mm TS-E looks like it may be sharp enough to do the test in the center and mid-frame, but not at the periphery, where it's somewhat soft even at f/8. The original 90mm TS-E shows a lot more moire at the center, maybe mainly because the parallel lines in the test chart are closer together.
The 90mm is significantly sharper and more contrasty than the 50 at the periphery.
Yes the 50 is not up to the new 90, but neither of us owned the new 90 or similarly good 135
That said we only looked at the centre area, specifically to get as good as we could - enough to decide it wasn't worth £300 extra for a 5DSR ;-)
@@KeithCooper I was talking about the old 90 TS-E, comparing that to the old 50 (if that's the one you used), which looks plenty sharp in the center. The old 50 is the one that's somewhat soft at the periphery.
The difference with the AA filter cancelled or not isn't great. For architecture, not having any moire is a great benefit, since it probably would often enough appear with the 5DsR.
One of my images was an arctic tern hovering over his/her chick with glorious looking wings, which I had to crop enough so that I barely had enough sharpness with the AA filter defeated in the 5DsR.
I only did some wildlife birding with the camera, and found 5 frames per second fast enough. The extra field of view with the GFX 100S would make finding and framing fast and erratic moving birds easier to capture, but the focus tracking probably isn't good enough, especially with an adapted lens. I may try it anyway, especially for the trumpeter swans that don't move as fast and are further away.
But my main use will be landscape, macro flowers and some cityscape.
Hi Keith, Thank you for the very informative video. We have PTZ cameras in an event space, the stage has an LED wall behind, so when focused on our talent we have considerable Moire on the LED wall behind unless we soften focus. The PTZ cameras are panasonic aw150we. I believe the newer model has an anti alias filter built in. My question is would a Anti aliasing filter work at the front of the lens or must it be on the sensor? I have a designed a 3D printable filter housing that will clip onto the front of the camera. But wanted to know if this way would work before getting AA filters cut to size. Many thanks
I don't know this applications, but the AA filtering really need to be direct to the sensor.
A diffuser on the LED wall would likely work, if not that practical though - it's likely the point source lighting of the LEDs which exacerbates the effect
I am curious about those cameras that support 4x high resolution modes, OM-D EM1 Mark II and on, Lumix G9 and GH6, Lumix S. It's claimed better colour (makes sense), and better resolution without overtaxing the lenses (I'm looking at Canon with its list of lenses not recommended for the 5Ds cameras).
The E-M1x and later also do 50 megapixels hand-held, they rely on camera movement from being held.
Do they show moire under difficult conditions.
Yes that Canon list for the 5Ds - mostly a marketing creation. I remember looking at it, chuckling, and thought, yes, that will help sell a few lenses and cameras - then I ignored it ;-)
AA filters are a major problem in my opinion, and each individual camera needs to be thoroughly researched on its AA filter alone. My first dslr was a cheap Nikon entry level camera15 years ago which ad 10mp. My 8mp smart phone produced much sharper images. So I got rid of it and when to the first ever 24mp Sony apsc slt and experienced fantastic results. I never went back to Nikon but kept an eye on their cameras, and found that they had some peculiar sharpness issues, which I attributed to their choice of AA filter.
After Sony I switched to Pentax for the IBIS astrotracer ability, newer models went AA filterless. My first Pentax was the brilliant K-5 with 16mp that already had a very light AA filter IMO. My following K3 had the AA filter simulator which I've never used really. AA simulator is the IBIS rotating on a sub pixel level to blur the image slightly to stop aliasing. Plus Pentax IBIS does horizon correction, tilt shift mode/ability and the newer versions of GPS-less astrotracer.
So when you are in Pentax land with AA filterless and AA simulation mode, its something which I haven't had to deal with.
I did buy some Fuji cameras and lenses, one was bayer and the other Xtrans and I didn't feel comfortable going backwards to a camera system that didn't have IBIS and some camera models had AA filters.
So I would be skeptical if the Fuji 100s didn't have even a little bit of an AA filter in the UV/Cut filter glued to the sensor. Mitsubishi owns Nikon and Fuji and I feel both brands are subpar in my experience. I think the GFX system is better than their X series, but I would never buy one because I find their lenses are to boring and clinical looking and lack some dimensionality. X-Trans sensors DO have a false colour issue according to a couple of Peta pixel articles.
Maybe 100mp is enough to make aliasing go away in images. Or maybe its software. Or maybe its simply not a true AA filterless camera - you know the old false advertising as many companies do regularly.
So AA filters aren't an issue when you shoot with Pentax which has all these cool IBIS modes and abilites, and Pentax solved the problem with AA simulator with varying degrees of blur.
I believe that the Mitsubishi companies, including Nikon, are related by ownership, not structured as a hierarchy as are western companies. Imagine a group of Japanese warlords retiring form warlording, and seeking enterprises more suitable to the more peaceful times they could see coming. Imagine there were coalitions similar to European countries of recent centuries for example. So the warlords like the idea of maintaining these friendships by cross investing, so each goes about his business as he always has. One wants to make pencils and other stationary. Another likes shipbuilding, someone else automotive. Nikon is part of this, but for some reason chose not to share the name. Fujitsu has a similar structure, with computers and electronics, whitegoods, tyres but again under a different name. Canon has a more western structure.
Wow! Fuji possibly lying about not having an AA filter, I never would have guessed that was possible. I suppose someone with the technical know how and equipment could check it during a tear-down. It looks like they solved the moire problem in some way, according to DPReview's test chart, which I commented on under my comment here.
I mention I got the Canon 5DsR. I never liked how Canon negated the AA filter with another lens system, instead of getting rid of it altogether.
I loved the 5DsR, except for the poor dynamic range, and eventually I just got sick of being stuck in 3:2, knowing that the lenses were capable of so much more coverage, which the 100S' larger sensor can do.
So I'll be using mostly my EF lenses on the 100S with the Fringer adapter; though, many won't cover the entire sensor. Pulling the rear baffle out of some help immensely. With that removed, the Canon EF 100-400 IS II is a stellar tele on the Fuji.
They are indeed a major problem for some - to others they matter not one jot
I fall midway and merely suggest that people don't get too carried away by the protestations of the ultra-pixel peepers, and spend more time looking at the whole image ;-)
I genuinely gave it no direct consideration in my purchase of a GFX100S - I did, indirectly, by some of the stuff I photographed when testing...
Unfortunately 'clinical and boring' have no transferable meaning from my POV - the lenses I have looked at are excellent - as are some of my creaky old ones, dismissed by some as 'wholly unsuitable' for the 100S :-) :-)
@@KeithCooper We both know from personal experience that many lenses work well with the 5Ds cameras, which have pixels not much larger than the 100S' pixels. So I expect lenses that worked fine on the 5Ds to work fine on the 100S. It seems to me that the lenses that worked well on the 21 MP 5D Mk II seem to work as well on the 5DsR. If they were lousy on the 5DII they are lousy on the 5DsR, and vice versa.
I think I understand your crusade against those fixated with pixel peeping, which for many can be a waste of time, and time could be better spent taking images.
For some at DPReview forums, for example, many were more focused on winning debates than finding the truth. I suspect some may even be intoxicated while posting.
I've been involved in some of those discussions. Some were fruitful, some not, but I learned a lot not just about cameras but people. Some of the commenters are actually mean, and it's more about winning and defending their brand than the truth. Their hobby seems to be about winning debates, not photography. That's unhealthy, and is a pain for the moderators.
I really appreciate the technical know-how people who focus on the facts, and who have high standards.
I've mostly printed 24 x 36 inch prints, which require getting a lot of things right. For me, it's important to know the specifics in order to produce work that I'm pleased with.
I mentioned before that I actually quit taking photos and printing for many years, and am now just getting back into it, partly because of this great oversized sensor 100S and because of the higher reflectivity dye ink prints; though, I'm going to now be limited to 13 inches, which is okay for now, because I'm retired and not working, so have limited funds.
Yes - I do have a genuine interest in all the technical stuff, but it's almost always out of technical curiosity or to support my more general photography.
I know what you mean about some on DPR - it's one of the main reasons we decided against having forums on the Northlight Images web site ;-) Also because we'd have enforced a 'real names only' policy [I have a 20+ year policy of mainly ignoring emails where people won't use their real name...]