March 31, 2015, 12:04:40 PM

Poll

Would you like a camera without low pass filter?

YES! sharper is always better
27 (25.7%)
Yes, I'm willing to take the risk of aliasing/moiré
29 (27.6%)
No, the tradeoff doesn't appeal to me
39 (37.1%)
NO, don't see any need & never touch a running system
10 (9.5%)

Total Members Voted: 105

Voting closed: March 02, 2015, 12:50:31 PM

Author Topic: POLL: aa, or not to aa (5ds vs. 5dsr) ?  (Read 4199 times)

Danielle

  • Rebel T5i
  • ****
  • Posts: 142
    • D.Judd Photographics
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #15 on: February 01, 2015, 09:49:03 PM »
50mp is getting possibly serious enough to go without. But I'm also willing to bet presuming these numbers are correct that either camera might be phenomenal. I said I'm willing to take the risk of aliasing. In the real world, I'm unsure especially of price!
In the end, only the image matters... Not what equipment you used to get there.

canon rumors FORUM

Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #15 on: February 01, 2015, 09:49:03 PM »

helpful

  • EOS M2
  • ****
  • Posts: 222
  • Ecclesiastes 3:11
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #16 on: February 01, 2015, 10:02:33 PM »
I selected the second option. Sharper is not always better, so that's not an accurate reason. But sharpness when needed is well worth the risk of moire, which is a risk that can be avoided by a knowledgeable photographer, and which technology causes to become smaller and smaller even in the cases when due to photographer error it wasn't avoided.

I have used both the D800E and D810 considerably, and I find that it boils down to this:

Main reason to go AA-free: Sharpness, and overall drastic image improvement at the level of tight crops. I totally disagree with the predominant opinion that it is an insignificant improvement (see below). I can crop so much smaller areas of the picture and still obtain a sharp result. That's the main factor for me. Even in the photo where Ken Rockwell is trying to show there is almost no difference, I see the opposite, at least a doubling of sharpness. Branches and twigs that look like fuzz in the D800 image suddenly jump out in the D800E image. If you need to crop out one small face from a huge 36 MP file, you'll depend on a D810 or D800E (or 5Ds R) to do the job. You can't post fuzz and keep your job.

Main reason to keep AA: Most of the time it never really matters. Extra sharpness is wasted in 90% or more of the applications of my images. Those who say it is an insignificant difference are talking about these 90% of applications where sharpness doesn't matter that much anyway, and I would agree with them in these cases.

Fake reason to keep AA: Fear of moire.

Moire is something that a photographer can learn to avoid the same as a photographer learns to focus. In fact, focusing itself can be used to control moire. In fact, focusing alone is a much, much bigger issue than moire ever could be.

I lose more photos due to missed focus than I ever will to moire showing up. But do I feel the need to use only f/11 wide angle focus free lenses? Of course not. And the exact same thing holds true about moire and AA filters.

I do not feel the need to blur every pixel in my picture with an AA filter because of a crippling fear of moire, anymore than I need to avoid f/1.4-2.8 lenses because of a crippling fear of being out of focus.

I will always buy the camera option that is AA-free, and with the current state of image processing, I don't even need to feel the slightest bit of anxiety about doing so.

5DIII, 5DII, 7D x5, 6D, T2i, T3, 1D X, 10-22mm, 16-35mm II, 18-55mm II, 18-135mm IS x2, 70-200mm f/2.8L II, 24mm f/1.4L II, 50mm f/1.4, 50mm 1/1.8 II x2, 85mm f/1.8 x2, 100mm f/2 x2, 135mm f/2L x2, 200mm f/2.8L II x2, 1.4X III, 2.0X II, 60mm f/2.8 Macro, etc. only had room to list a few Canon items

torger

  • EOS M2
  • ****
  • Posts: 236
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #17 on: February 02, 2015, 02:04:03 AM »
I shoot medium format were there you can't get a camera with AA filter. It's a pain. People has come to believe that being without low pass filter is a feature, but believe me it's not. It's a bad idea, anyone remotely familiar with signal processing knows it.

People think it's only about moire, but the main problem (unless you shoot fabrics) is actually false colors around small details. I rather have a correct smooth image than an aliased image riddled with false colors. With small apertures you can compensate some, but with the Canon you'd need f/11 or even f/16 to really kill aliasing. Yes you can live with it in landscape as you shoot with small apertures and false colors is in that case only obvious to the trained eye (moving water can be a big issue though), but it's just so unnecessary when there is an alternative.

And really at 50 megapixel you need that tiny bit of extra sharpness? False colors is a real image quality issue, while slight softening at 50 megapixel resolution is not, especially when you make a print. Being without AA filter is only a pixel peep thing. AA filters sharpen up well, as it's the same smoothness over the whole image surface. You have a problem with sharpness if the lens does not perform and delivers uneven sharpness over the surface, but that's a different story.

Here's one example article showing the effects of moire and false colors:

http://echophoto.dnsalias.net/ekr/index.php/photoarticles/78-aliasing-and-supersampling-why-small-pixels-are-good
« Last Edit: February 02, 2015, 02:06:25 AM by torger »

torger

  • EOS M2
  • ****
  • Posts: 236
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #18 on: February 02, 2015, 02:31:18 AM »
Being a contributor to RawTherapee I know a thing or two about demosaicing, and sensors without AA filters is not making it easier. You will get false details and colors, when you know what happens on the software level and how much errors there is in the final image it's much harder to think AA-free cameras is a good idea than if you don't know...

However a little disclaimer - my experience comes mainly from medium format where lenses are sharper and microlenses are either not there or are less effective. It may happen that thanks to softer lenses and better coverage of the microlenses combined with a landscape aperture at f/11 the false color issue is so much reduced it will be negligible to some. If you're the unsure kind I'd wait and see how it looks.
« Last Edit: February 02, 2015, 02:33:46 AM by torger »

Marsu42

  • Canon EF 400mm f/2.8L IS II
  • *********
  • Posts: 6037
  • It's a PONY!
    • Der Tierfotograf
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #19 on: February 02, 2015, 03:05:49 AM »
However a little disclaimer - my experience comes mainly from medium format where lenses are sharper and microlenses are either not there or are less effective. It may happen that thanks to softer lenses and better coverage of the microlenses combined with a landscape aperture at f/11 the false color issue is so much reduced it will be negligible to some. If you're the unsure kind I'd wait and see how it looks.

Interesting experiences from the mf guys. Imho the 5ds-r cannot be that bad color-wise, and I don't remember reading about it on the d800e reviews (even though I'm not in the market for one of these, so I didn't go into the details). My guess is that the main issue will be moiré, which is sufficiently bad to make two camera releases prudent.

torger

  • EOS M2
  • ****
  • Posts: 236
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #20 on: February 02, 2015, 03:20:18 AM »
Here's an example of false colors in an irregular landscape subject, the magenta and cyan pixels are not there in reality:

http://www.ludd.ltu.se/~torger/photography/img/ltr-colormoire.jpg

this requires the lens to be very sharp of course, and results vary depending on demosaicers. In general the demosiacers producing the finest details get more problems with false colors. Lightroom makes more false colors than Phase One's Capture One for example.

Water is another landscape classic where moire/false colors can show:

http://echophoto.dnsalias.net/ekr/Articles/Aliasing2/seawater_a.png

So there's more than fabrics.

When you shoot winter subjects it can be visible around sharp edges, thin tree branches on a white background for example.

The examples I've seen from the D800E when it doesn't produce any aliasing always show a pretty soft lens or much stopped down...

Marsu42

  • Canon EF 400mm f/2.8L IS II
  • *********
  • Posts: 6037
  • It's a PONY!
    • Der Tierfotograf
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #21 on: February 02, 2015, 03:38:05 AM »
Here's an example of false colors in an irregular landscape subject, the magenta and cyan pixels are not there in reality

Riiiiiight... I guess you really have to look for this, other than moiré on the infamous cloth everybody knows from tv. Probably for many people, a softer lens (esp. on high-res 50mp) will indeed work as an aa filter substitute, giving enough blur to hide this issue. But if the Canon 5ds-r performs as your samples, it'd be a reason for me not to get it.

canon rumors FORUM

Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #21 on: February 02, 2015, 03:38:05 AM »

rs

  • 5D Mark III
  • ******
  • Posts: 737
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #22 on: February 02, 2015, 04:21:50 AM »
Here's an example of false colors in an irregular landscape subject, the magenta and cyan pixels are not there in reality

Riiiiiight... I guess you really have to look for this, other than moiré on the infamous cloth everybody knows from tv. Probably for many people, a softer lens (esp. on high-res 50mp) will indeed work as an aa filter substitute, giving enough blur to hide this issue. But if the Canon 5ds-r performs as your samples, it'd be a reason for me not to get it.
This is a very divisive subject.

If a system comprising of a 5Ds R, lens, and user/conditions all combine to produce images sharp enough to reveal the extra 'detail' that would be missing with an AA equipped sensor, moire can be a very real issue. This will only bother people who are irked by the difference between real detail and false detail. And if it's next to impossible or even impossible to excite moire with this system due to ultimate sharpness being unobtainable for one reason or another, what possible advantage does removing the AA filter give anyone?
5D II | 16-35/4 | 24-70 II | 70-200 II | 100L | 40 | Sigma 50/1.4 | 40D | 10-22 | 580 EX II | 1.4x TC II

RLPhoto

  • Canon EF 300mm f/2.8L IS II
  • ********
  • Posts: 3701
  • Gear doesn't matter, Just a Matter of Convenience.
    • My Portfolio
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #23 on: February 02, 2015, 08:49:59 AM »
With my MF system, I've seen enough moire to want an AA filter in my 35mm system. I'd want the AA filter as 50mp would be give plenty of resolution anyway.

StudentOfLight

  • Canon 7D MK II
  • *****
  • Posts: 498
  • I'm on a life-long journey of self-discovery
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #24 on: February 02, 2015, 11:11:09 AM »
With my MF system, I've seen enough moire to want an AA filter in my 35mm system. I'd want the AA filter as 50mp would be give plenty of resolution anyway.
Out of interest, what is the pixel pitch on the MF system? If for instance you were to take the same picture from the same spot with smaller pixels wouldn't the moire be eliminated?
Fantasy Gear:
TS-E: 45mm f/2.8 L-II,  EF: 40mm f/0.8,  100mm f/1.4,  35-85mm f/1.8, 
EF with 1.4xInt: 100-300mm f/4 ,  500mm f/5.6 L

jdramirez

  • Canon EF 300mm f/2.8L IS II
  • ********
  • Posts: 2809
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #25 on: February 02, 2015, 11:23:05 AM »
You know what's funny... Is this entire time I didn't give a Damn that there was an as filter over my mkiii's sensor.  I'm guessing I wouldn't notice... in the grand scheme of things... if I'm happy with the performance the mkiii, I'd be happy with either of the refreshers.
Upgrade  path.->means the former was sold for the latter.

XS->60D->5d Mkiii:18-55->24-105L:75-300->55-250->70-300->70-200 f4L USM->70-200 f/2.8L USM->70-200 f/2.8L IS Mkii:50 f/1.8->50 f/1.4->100L-> 85mm f/1.8 USM-> 8mm -> 85mm f/1.2L mkii

crashpc

  • Rebel T5i
  • ****
  • Posts: 115
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #26 on: February 02, 2015, 01:26:13 PM »
The problem is, that I get AA artifact on Canon AA-filter-equipped body anyway. That way I rather get AA-filter-less body, and get rid of AA different way, and when I need sharpness, I get it. Now I can´t get it...

RLPhoto

  • Canon EF 300mm f/2.8L IS II
  • ********
  • Posts: 3701
  • Gear doesn't matter, Just a Matter of Convenience.
    • My Portfolio
Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #27 on: February 02, 2015, 02:59:41 PM »
With my MF system, I've seen enough moire to want an AA filter in my 35mm system. I'd want the AA filter as 50mp would be give plenty of resolution anyway.
Out of interest, what is the pixel pitch on the MF system? If for instance you were to take the same picture from the same spot with smaller pixels wouldn't the moire be eliminated?
I have no idea. It's a 1.1x 645 crop sensor with 39 megapixels. Maybe someone can do the calculation.

I've seen moire in my 5D3 files occasionally! It's usually a tight pattern on a suit is when I saw it. My Hassy will show it occasionally with a suit, or anything woven super tight and small. Usually you can correct it by standing a little forward or back but sometimes it's just a pain.

canon rumors FORUM

Re: POLL: aa, or not to aa (5ds vs. 5dsr) ?
« Reply #27 on: February 02, 2015, 02:59:41 PM »