Canon 5Ds & 5Dr

[odd jim]

Fed up with my 5DSr. Even at 1/400th of a second I struggle to get a sharp shot. On a tripod, with great light and perfect exposure at ISO 100 it's ridiculously good, the level of detail is almost unbelievable and colours are fabulous. Handheld in normal daily lighting conditions it's absolute rubbish. The worst carry around camera I've owned and not a patch on my 5D. Noise is insanely bad over 400 ISO unless your lighting and exposure are immaculate. AF is painfully slow.

You should be able to get excellent result hand held (noise at high ISO aside).

I had no problem with one shooting with my 24-105L with normal light (landscapes), I shot it in exactly the same way I shoot my other Canons and didn't suffer with softness. All settings within comfortable ranges. That said, I much prefer my 6d for this purpose (and the 5dsr wasn't mine, I wouldn't personally buy one as it doesn't suit me), but mainly because I could pull far cleaner shadow detail from sunset shots.

 

[StewartR]

Why do we get all this about the monster 5ds and 5dsr can't get a sharp picture due to 50mp but we don't hear the same complaints about the 7d2?

The fact is it is about pixel density and it is the same. Just imagine the 5ds/r is a 7d2 with additional coverage.

Spot on. There was a similar misconception when Nikon launched the D800. Ooh, 36 megapixels, you can only use the best quality lenses, you have to use a very fast shutter speed, that kind of guff. Completely ignoring the fact that even the entry level D3x00 cameras have higher pixel densities.

 

[boristhemoggy]

you are welcome... but not worth to give much more information for such posts anyway

Which pretty much shows your mentality so do us all a favour and simply pass the post by eh? I mean if you have nothing to offer then you're just trolling right?

 

[Chipper]

User error rather than camera error

I agree with those who find this unhelpful and not in keeping with the spirit of the forum. So I guess that is your user error.

 

[HoppyUK]

Spot on. There was a similar misconception when Nikon launched the D800. Ooh, 36 megapixels, you can only use the best quality lenses, you have to use a very fast shutter speed, that kind of guff. Completely ignoring the fact that even the entry level D3x00 cameras have higher pixel densities.

It's not about pixel density per se and comparisons with the 7D2 (with similar pixel density) are not valid.

The potential sharpness from a full-frame 51mp sensor are considerably greater than an APS-C sensor with 20mp. On full-frame, the lens is able to deliver more detail and there are a lot more pixels to record it. Basically, the 5DSR is able to capture finer detail beyond that of the 7D2's sensor and beyond the limits of all but the best lenses. It is still susceptible to camera-shake and mirror-slap though, so the standards applied there must be correspondingly higher.

 

[StewartR]

It's not about pixel density per se...

Why not?

 

[HoppyUK]

Why not?

As above really. The resolution of fine detail is about lens performance and the total number of pixels (not pixel density or size per se), preferably without an AA filter. With 5DSR vs 7D2, the 5DSR wins all round - larger sensor allows the lens to perform better (basic MTF principle), it has 2.5x more pixels with no AA filter, so is able to record that detail - and more, with the best quality lenses.

The 5DSR's resolution potential is therefore much higher, but the flip-side is it's more difficult to extract every last bit of detail. Lens quality and focusing become more critical, and susceptibility to movement blur is that much higher (camera-shake, mirror-slap, subject movement).

You're right on one aspect though, pixel size/density impacts lens sharpness on all formats. Basic lens MTF theory is that when resolution demands go up, ie smaller pixels, then image contrast goes down, and contrast is the dominant factor in our visual perception of sharpness. It's the main reason why full-frame is sharper than APS-C.

 

[StewartR]

The resolution of fine detail is about lens performance and the total number of pixels (not pixel density or size per se), preferably without an AA filter. With 5DSR vs 7D2, the 5DSR wins all round - larger sensor allows the lens to perform better (basic MTF principle), it has 2.5x more pixels with no AA filter, so is able to record that detail - and more, with the best quality lenses.

Whilst I'm painfully conscious of the fact that you understand this stuff better than I do, I can't avoid the suspicion that we may be talking at crossed purposes. At least, I hope we are, because otherwise I'm really confused.

The context here is whether a 5DS "needs" better lenses than a 7D2, so it seems to me that we ought to be talking about how the two cameras perform with the same lens. But surely the 5DS having 2.5 times as many pixels is irrelevant, because it's also recording a scene which has 2.5 times the area. If that's not the case I (and I suspect many other people) would really appreciate an explanation as to how all those pixels around the edges of the 5DSR sensor affect the performance in the middle of the sensor.

You're right on one aspect though, pixel size/density impacts lens sharpness on all formats. Basic lens MTF theory is that when resolution demands go up, ie smaller pixels, then image contrast goes down, and contrast is the dominant factor in our visual perception of sharpness. It's the main reason why full-frame is sharper than APS-C.

I think you missed out the word "usually", as in "full-frame is usually sharper than APS-C". My trusty old 40D has a lower pixel density than a 5DS, so I think I'd expect it to be sharper.

 

[HoppyUK]

Whilst I'm painfully conscious of the fact that you understand this stuff better than I do, I can't avoid the suspicion that we may be talking at crossed purposes. At least, I hope we are, because otherwise I'm really confused.

The context here is whether a 5DS "needs" better lenses than a 7D2, so it seems to me that we ought to be talking about how the two cameras perform with the same lens. But surely the 5DS having 2.5 times as many pixels is irrelevant, because it's also recording a scene which has 2.5 times the area. If that's not the case I (and I suspect many other people) would really appreciate an explanation as to how all those pixels around the edges of the 5DSR sensor affect the performance in the middle of the sensor.

I think you missed out the word "usually", as in "full-frame is usually sharper than APS-C". My trusty old 40D has a lower pixel density than a 5DS, so I think I'd expect it to be sharper.

Slightly at cross purposes, but the post I responded to went beyond "The context here is whether a 5DS 'needs' better lenses than a 7D2" and introduced other factors. I touched on this in my last paragraph. I know where you're coming from, but you're confusing pixels/resolution and lines-per-mm/resolution that are not the same thing, and overlooking the effect that physical sensor size/area has on lens performance. Put another way, the 5DSR doesn't 'need' better lenses, but it's capable of recording a lot more detail when presented with a really good lens. Correspondingly, actually capturing that extra detail demands immaculate technique, with faster shutter speeds and so on, or camera-shake, mirror-slap and subject movement will blur it.

If you get a chance to compare a decent lens on a 40D and 5DSR, then the 5DSR will be clearly better, no question. I did something similar some years ago now, when I was looking to upgrade my 40D (10mp, APS-C) and to the then-new 7D (18mp, APS-C). Shot side by side, the 7D was very slightly better, but not much, and you had to look damn close. But the shop also had a 5D2 (21mp, full-frame) so I put the same lens on that and did the same tests. The 5D2 was immediately and obviously better. It turned out to be quite an expensive little comparison as I had to change all my EF-S lenses too.

Coincidentally, I've just responded to a similar point on another thread, post #37 here https://www.talkphotography.co.uk/threads/nikon-af-s-nikkor-85mm-f-1-8g.630007/ so I've copied it below.

It is a common misunderstanding that sharpness is defined by the number of pixels. That has an impact of course, but it's relatively minor compared to the size/area/format of the sensor - size matters, and bigger is better. It's physics (as opposed to marketing talk ) and the bottom line is basically this - using the same lens at same settings, on average images will be about 10-12% sharper on full-frame.

What we call 'sharpness' has two components - resolution (the fineness of detail) and contrast (how clearly those details are shown). Of the two, contrast has the most influence on our visual perception of sharpness. The physics of it is, as resolution demands go up, so image contrast goes down. It's like a car that will accelerate from 0-60 in six seconds, takes twice that time (or more) to get from 60-120. The more you ask, the harder it gets, and lenses are the same. DX cameras have smaller sensors, so they demand higher lens resolution - 1.5x higher for DX vs FX.

Good explanation here, but just look at the first couple of diagrams that show as the black and white lines get closer, so the contrast gets lower. Blacks get lighter, whites get darker, ultimately turning to grey porridge http://www.cambridgeincolour.com/tutorials/lens-quality-mtf-resolution.htm

Lens MTF tests (Modulation Transfer Function) compare resolution vs contrast. Here's one example of MTF graphs for four 85mm f/1.4 lenses, plotting contrast at different resolution levels - 10-lines-per-mm, 20, 30, 40, 50 https://www.lensrentals.com/blog/2016/04/sony-fe-85mm-f1-4-g-master-lens-mtf-and-variance/ Just look at the Y-axis up the left (which is the centre of the frame, you can ignore everything else*) and note how with every increase in resolution of 10-lpmm, contrast drops by roughly 0.1 (that's actually 10%, with 1.0 at the top being 100% theoretical maximum). The orange 20-lpmm and green 30-lpmm lines are good ones to look at. That's a decent standard for high quality working, and conveniently, the difference between them is 1.5x - exactly the Nikon DX crop factor. In other words, if you put the same lens on a DX camera with 20mp sensor, the green line shows the standard of sharpness; but on a 20mp FX camera, you'll get the orange line and a jump of maybe 10-12% in sharpness.

*Those graphs are only a snapshot of lens performance. In particular, they're at f/1.4 only, with % contrast up the left, and mm from the centre of the frame along the bottom. It's worth noting the Zeiss Otus 85/1.4 with outstanding sharpness in the centre at f/1.4, though the others catch up and get much closer at f/2.8-4. That's what you get for £4k

 

[Craig_85]

It's not about pixel density per se and comparisons with the 7D2 (with similar pixel density) are not valid.

The potential sharpness from a full-frame 51mp sensor are considerably greater than an APS-C sensor with 20mp. On full-frame, the lens is able to deliver more detail and there are a lot more pixels to record it. Basically, the 5DSR is able to capture finer detail beyond that of the 7D2's sensor and beyond the limits of all but the best lenses. It is still susceptible to camera-shake and mirror-slap though, so the standards applied there must be correspondingly higher.

Personally I would expect the technique required and the available sharpness from a 5ds and a 7d2 with the same lens shot from the same position and not downsized to be identical at the pixel level.

If you used a 1.6x longer lens or walked closer with the 5ds then there would be an impact.

Obviously the 5dsr should achieve a greater sharpness due to the cancelling of the anti aliasing filter.

IMHO it is about pixel density, I happen to think the OP's issue is lens or lens drive related as he mentions slow autofocus, well it is not normally so seems a little odd.

Where the full frame requires a better lens is because it sees further towards the edges of the lens where it is inherently weaker.

 

[HoppyUK]

<snip>

If you used a 1.6x longer lens or walked closer with the 5ds then there would be an impact.

<snip>

That is the only valid comparison. If you don't adjust focal length when comparing formats, then you're not looking at the same picture! Changing the shooting distance to keep the main subject the same size in the frame can also work, so long as perspective changes don't skew things.

 

[StewartR]

That is the only valid comparison. If you don't adjust focal length when comparing formats, then you're not looking at the same picture! Changing the shooting distance to keep the main subject the same size in the frame can also work, so long as perspective changes don't skew things.

Sorry Richard, I disagree.

The thought experiment that I am doing, and that others are doing (including @Craig_85 I think), is to compare a 7D2 image with the central portion of a 5DS image. Same lens, same settings, same 22mm x 15mm chunk of silicon, same framing and composition, same pixel count. We're expecting the images to be the same.

I assert that that is a valid and meaningful comparison. Some people claim that the 5DS (and the D800 before it) needs the best lenses and the best technique, but - and this is critical - they don't say the same about crop bodies with lower pixel counts. I say that if the lens and/or technique is found wanting on the 5DS, then by reference to the above thought experiment it will also be found wanting on the 7D2.

 

[snerkler]

I've just stumbled across the latter part of this discussion and have found it very interesting, albeit a bit confusing. Would someone in the know mind clarifying a couple of things for me please? I read above (and other places) that DX is more demanding on a lens than FX and therefore generally results in less sharp images, why is this? I thought it was due to pixel density/size but the D750 has less pixel density than the D810 so would expect less demands on a lens, yet it's less sharp. Or in this case does the lack of AA filter on the D810 make more of a difference than the pixel density?

Are 'we' now saying that it's a myth that it's harder to get sharp shots with a high MP camera? Makes me wonder how folk get on with MF?

 

[HoppyUK]

Sorry Richard, I disagree.

The thought experiment that I am doing, and that others are doing (including @Craig_85 I think), is to compare a 7D2 image with the central portion of a 5DS image. Same lens, same settings, same 22mm x 15mm chunk of silicon, same framing and composition, same pixel count. We're expecting the images to be the same.

Then we are talking at cross purposes then Stewart, at least on one aspect, so apologies

In the comparison you describe, ie the same lens in a 1.6x crop taken from the centre of a 5DS vs whole sensor area of the 7D2 (same physical sensor area and same pixel count) there would be no difference. The only caveat would be the different effect of the AA filter (it may be stronger in the 7D2) and of course there is a notable difference in very fine detail if you compare with and without AA filter, eg 5DS vs 5DSR, or Nikon D800 vs D800E/D810.

I assert that that is a valid and meaningful comparison. Some people claim that the 5DS (and the D800 before it) needs the best lenses and the best technique, but - and this is critical - they don't say the same about crop bodies with lower pixel counts. I say that if the lens and/or technique is found wanting on the 5DS, then by reference to the above thought experiment it will also be found wanting on the 7D2.

Yes, valid and meaningful if you want to make the point that a high-res full-frame camera makes the same demands of lens performance as APS-C. It is wrong to say otherwise (actually, APS-C makes more demands in practise - see my reply to Snerkler below). However, and this is where I think the comparison falls short, that is not how full-frame cameras are used. When the whole sensor area from the 5DS/R and 7D2 are compared, the full-frame cameras are more demanding of technique in order to extract maximum sharpness, simply because there is more to be had from them.

 

[HoppyUK]

I've just stumbled across the latter part of this discussion and have found it very interesting, albeit a bit confusing. Would someone in the know mind clarifying a couple of things for me please? I read above (and other places) that DX is more demanding on a lens than FX and therefore generally results in less sharp images, why is this? I thought it was due to pixel density/size but the D750 has less pixel density than the D810 so would expect less demands on a lens, yet it's less sharp. Or in this case does the lack of AA filter on the D810 make more of a difference than the pixel density?

DX is more demanding of lens performance than FX, simply because the sensor is smaller. This applies regardless of pixel count (though that also has a lesser impact). I explained the reasons in post #449 above as best I can, with appropriate links.

Are 'we' now saying that it's a myth that it's harder to get sharp shots with a high MP camera? Makes me wonder how folk get on with MF?

The more detail you want, the harder it is to get. Focusing must be dead accurate (especially when depth-of-field may be reduced with higher res cameras) and any movement blur (camera-shake, mirror-slap, subject movement etc) will take the edge off sharpness and be more visible.