Messages

31

Canon General / Re: More Lens Suggestions [CR1]

« on: February 14, 2012, 11:28:00 AM »

Where's all the new TS-Es? 

I'd love to see a TS-E 45mm upgrade (to improve optical performance), and after that a TS-E 90mm (to improve movement capability, optically it is already good).

32

EOS Bodies / Re: 5D Mark III/X Information [CR1]

« on: February 14, 2012, 10:51:01 AM »

And better color production.  I still feel like the XTi did better with flesh tones than the 5DII.

I hope you mean skin tones ;-)

33

EOS Bodies / Re: 5D Mark III/X Information [CR1]

« on: February 14, 2012, 10:41:52 AM »

I agree with you that comparisons are the only exact proof, but I don't believe Canon could allow itself to only use "mere ISO values in a marketing brouchure" and then not deliver in such high-end pro bodies, they would be the laughing stock then...I hope we will get blown away 

I think the 1Dmk4 (ISO102400) proves that Canon indeed can put in high ISO numbers in the camera just to make them show in the brochure, not caring about how it actually performs. Thus, I will remain skeptical until real RAW samples and tests appear.

34

EOS Bodies / Re: 5D Mark III/X Information [CR1]

« on: February 14, 2012, 10:23:21 AM »

Well, define compete. IMO competitive doesn't need to mean undisputed best at everything. There is no such thing from any manufacturer, and if so, it never lasts long. I think even if they don't beat the D4 in high ISO, half a stop is barrely the difference between competing and not competing.

I agree. Half stop up or down is what fanboys can argue about, but won't make a difference on the whole. I don't think Canon will come out on top concerning sensor image quality this time around either, but it will not matter. Even if the sensor performs worse in terms of ISO than the 36 megapixel Nikon many users will just be happy with the fact that they don't need to deal with 36 megapixel files. Just as high megapixels attract some users (resolution for landscape/studio), a low number attract others (faster workflow, lower storage requirements). Finding out exact ISO performance is more complicated so I think most don't do that at all.

The high megapixel count may cause the D800 to look as a highly specialized camera, regardless if it performs really well at high ISOs. A more moderate count like 22 megapixels may attract more users just because it seems more all-around and sane, regardless of its actual performance.

35

EOS Bodies / Re: 5D Mark III/X Information [CR1]

« on: February 14, 2012, 09:54:12 AM »

Those that expect large ISO improvements in RAW files are in for disappointment. From market reasons it must be less good than 1Dx, and the 18 megapixel 1Dx may be as good as the 12 megapixel D3s in RAW performance. The D3s is about 1 stop better than 5Dmk2 (plus it can do very high ISOs which 5Dmk2 can't).

I'd say that a realistic expectation is that a 22 megapixel 5Dmk3 sensor is about 0,5 stop better than 5Dmk2 up to ISO6400, plus that the ultra high ISOs (ISO12800+) is more useful than on 5Dmk2.

36

EOS Bodies / Re: 5D Mark III/X Information [CR1]

« on: February 14, 2012, 08:55:29 AM »

The true genius of the D800 was the D800E. I would gladly take an 22Mpx 5dmkIII. But I want to be able to have no anti alias filter. It makes a world of difference for sharpness, and it wouldn't hurt the sort of photography that the 5D shooters usually do. In any case, you could always buy it with the AA filter.

I don't really understand not wanting megapixels because resolution is not important, but want to skip AA filter because, ehh, resolution is important? As I see it, it is only a good idea to drop the AA filter when the resolution of the sensor is high (like on the D800), meaning that diffraction will typically kill moire/aliasing problem and thus further blurring would be unnecessary.

No AA filter is better suited for those that has slow workflows with lots of PP, such as studio/landscape, that is for a megapixel monster, not for an all-around camera.

37

EOS Bodies / Re: 5D Mark III/X Information [CR1]

« on: February 14, 2012, 08:28:03 AM »

I guess I'll stay with the 5D mark II then. Speed and AF is nice, but in those cases I almost always want more reach and use APS-C instead (7D). What I really desire is substantial sensor improvements at ISO100, and that is probably not going to happen.

It seems to me that the 5Dmk3 may be a mild evolved version, do as little as possible to please as many as possible, but no bold features. ISO performance will certainly be worse than 1DX, but will it be better than 5Dmk2? They don't need to make it better. Improving the AF is the only thing they really need to do, and then add some speed to differentiate from the competition.

38

EOS Bodies / Re: When will Canon Meet The Market Demand?

« on: February 14, 2012, 03:00:34 AM »

My greatest worry in the coming Canon cameras is the raw sensor quality. From various reports I get the feeling that Canon does not have the technology to match the best sensors found in Nikon cameras (Sony Exmor). They will certainly be "good enough", but my guess is that they will continue to lag behind in terms of high ISO performance and DR. I do hope I'm wrong though.

39

EOS Bodies / Re: When will Canon Meet The Market Demand?

« on: February 14, 2012, 01:47:43 AM »

Photography develops with the tools that become available. With high speed and fast autofocus sports images that were near-impossible to capture before is now everyday. There have been great advancements in low light photography too. The technical image quality bar has also been raised.

This means that even if the photographer still makes the pictures, she/he needs competitive tools to stay in the front (depending on which style of photography of course)

40

Lenses / Re: 36+ Mp EOS dSLR (rumored): How do existing EF lenses cope?

« on: February 10, 2012, 10:44:33 AM »

"outresolving the lens" is rather subjective thing.

Even if image degradation is evident due to limitations in resolving power, one will get still a little more resolution out of the system if sensor pixel count is increased further.

So it's all about when we reach the point when adding more pixels gives diminishing returns, and that is subjective indeed.

One extreme is that if we see evidence of resolution loss in corners of the cheapest lenses wide-open then we have reached that point, the other extreme is only to look at the sharpest lenses at ideal apertures in the center and accept rather soft pixels there.

In any way you see it, it is subjective. I think the point-of-diminishing returns is around 55-60 megapixels, a guess based on what I see on the 7D and dropoff on the 5D.

41

EOS Bodies / Re: 5D Split? 5D X & 5D Mark III [CR1]

« on: February 10, 2012, 08:13:19 AM »

So, shat is more likely - a  40+MP camera with decent video and a 22MP one with top-notch video (as this rumour states) OR a 40+MP camera with top-notch video and a 22MP one with decent video?   I'm not 100% certain, but I am certain someone can give a good technical explanation on why one case must be more feasible than the other.

4K video pixel binned 4:1 would yield about 40 MP sensor. Aiming a single product at both videographers and high res fashion/studio/landscape shooters might seem a bit odd though from a marketing perspective. But from a technical point of view I guess it makes sense. That would leave space for the 22MP all-arounder (okay in every aspect, not king on any), not mess with the 1DX (best speed, best ISO), and the combo would be competitive with Nikon's offerings. The 22MP all-arounder would be a little bit cheaper than the D800, and the high res 4K camera would be more expensive, but probably cheaper than 1DX, unless its a pro body much similar to 1Ds. I doubt though that high res is going to be as expensive as 1Ds/D3x from now on.

If this is about true, I think most will be happy. Those that want high res (like me) are probably prepared to pay a bit more (lens cost is the major part anyway), as long as it does not feel like D800 is a much better deal. $4K would be ok. The masses want that 22MP all-arounder, and if it is a few bucks cheaper than the D800 and/or has a little bit more speed it will be ok. With good AF it will be an upgrade from the 5D mark 2.

What I'm less sure about is what 4K videographers are prepared to pay. Are they expecting 5D mark 2 pricing?

42

Lenses / Re: 36+ Mp EOS dSLR (rumored): How do existing EF lenses cope?

« on: February 10, 2012, 06:54:08 AM »

Hi,
    I just wonder whether anyone notice that a 36MP DSLR (eg. Nikon D800) might face a very practical usability issue... the size of a 36MP RAW file and the time for a 75MB RAW file to write to the memory card. Even at the current highest write speed card of 100MB/s (assume the DSLR is able to write at maximum speed and a RAW file is 75MB), it'll took Nikon D800 3s to write 4 RAW files to the card and current price of high capacity, high speed CF card is not cheap.

I think 75MB was the uncompressed format which noone uses, normal size is about half. Still I think it is a mistake by Nikon not to have sRAW and mRAW formats like Canon has. They still seem to think that people shoot only JPEG and design their cameras for that. They do have reduced size JPEG formats, but not RAW.

Reduced size RAW formats makes a high res camera more versatile and all-around. If I do hand-held high ISO action-packed high volume shooting I always use reduced size RAWs. Speeds up workflow and reduces disk space, without any impact on quality, since the shooting conditions won't allow getting max resolution anyway.

I imagine that D700 users that think 12 megapixels is more than enough for their shooting style are put off by D800 huge files. This would have been simply fixed by providing reduced size RAW formats. I'm sure D800 scaled down to 12 megapixels provides better image quality than D700.

43

EOS Bodies / Re: 5D Split? 5D X & 5D Mark III [CR1]

« on: February 10, 2012, 06:42:29 AM »

Where's the cinema DSLR in this (which canon actually has announced that it is in development)? Both a 5D split *and* a cinema DSLR? Seems strange.

44

EOS Bodies / Re: Latest Canon disappoints wants me to switch over to Nikon.

« on: February 09, 2012, 02:55:46 PM »

DLA (Diffraction Limited Aperture) is the result of a mathematical formula that approximates the aperture where diffraction begins to visibly affect image sharpness at the pixel level. Diffraction at the DLA is only barely visible when viewed at full-size (100%, 1 pixel = 1 pixel) on a display or output to a very large print. As sensor pixel density increases, the narrowest aperture we can use to get perfectly pixel sharp images gets wider.

If I read this , I would say 18MP on a 1.6 has a lower DLA aperture then 18MP on a full frame.
The pixel density is higher with the 18mp on a 1.6 ??

Or am I wrong ?

What makes diffraction a challenge is the resolving power, 36 megapixel on FF is just as "bad" as 36 megapixel on medium format.

The 1.6 crop is using less of the lenses glass (resolve power) then the same lens on a FF.
So I would say the FF 36 MP will resolve more detail then 36MP on a 1.6

Diffraction is not the only limiting factor. There's also resolving power of the lens. Smaller formats generally have sharper lenses, but not to fully compensate for the smaller format, so at some point lenses start to limit. There's also a limit to how large you can open up the aperture and still have good resolving power from the lens. You may in some formats hit those limits.

36 MP on 1.6 crop then resolving power of the lens is very limiting.

Another thing - a mildly diffraction-softened image responds well to deconvolution sharpening, so one should not be too afraid of it.

45

EOS Bodies / Re: Latest Canon disappoints wants me to switch over to Nikon.

« on: February 09, 2012, 02:32:06 PM »

the interior architectural shots at f/8 are even worse; the diffraction is clearly evident in preventing additional detail from being rendered.

Diffraction issue does not have anything with the format to do, if your goal is to have a large DOF. With larger format you need longer focal lengths for the same FOV, and then need smaller aperture to compensate the DOF and - voilà - no difference. That is medium format and large format struggles with diffraction in the same way. Still people seem to use those formats, at even higher pixel counts... hmm...

What makes diffraction a challenge is the resolving power, 36 megapixel on FF is just as "bad" as 36 megapixel on medium format.

With a high resolution back you can't expect the sensor to be the most limiting factor for all your shots. Instead you are pleased that you have a sensor that max out the expensive lenses you have (probably more money in those than in the camera body) and get the most of your technique. If you think your pixels are not jagged enough for a particular shot, you can always downsize. However, fairly soft pixels both enlarge and sharpen well, so since I do prints I'm not a big fan of low rez jagged pixels a la Sigma (having no AA filter on sub 30 megapixel cameras is a bad idea (tm)).

At some point the pixel count will be "excessive", which would be when center sharpness at ideal apertures (~f/4) does not get much gain. I think that limit is around 45-50 megapixels.