September 20, 2014, 06:09:08 PM

Author Topic: Tony Northrup - D810 vs. 5D Mk3  (Read 20863 times)

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4563
  • POTATO
    • View Profile
    • Nature Photography
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #210 on: Today at 02:31:14 AM »
Well, what actually knocked some sense into me was when I summed up the total light gathering capacity of each sensor (based on sensorgen information), then factored in noise levels and DR. Once I saw those numbers, there pretty much wasn't any denying it. And actually, based on the numbers, the D810 is still phenomenal at high ISO...the A7s only really pulls away once you get into the ISO 51200 realm and beyond. Then I factored in your low ISO pushing technique to achieve more DR at "effective high ISO" when using lower ISO, and yeah. No contest. :)

There is good detail in the A7r, but to me, it does look visibly noisier than the A7s. Especially in the the objects in the background...much noisier. (And that's in the normalized comparison.)

Regarding the 50mp MF, I guess I was considering that a different class of camera...but, when you throw the 645z into the mix, I guess it really isn't. :P So sure, that one trounces the trouncer.

My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

canon rumors FORUM

Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #210 on: Today at 02:31:14 AM »

Khalai

  • EOS M2
  • ****
  • Posts: 164
  • Let there be (flattering) light!
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #211 on: Today at 03:43:29 AM »
Thanks for the answer. I myself prefer to use LEE GNDs to avoid the need of too many files. Frankly, I'm quite bad in blending (learning about luminance masks at the moment as well as some other techniques) so I try to do all I can in the field so I can ideally work with single files.

But when I do bracket, I bracket like mad: -3,-2,-1,0,+1,+2,+3 is my usual starting point :D

I too use Lee's GNDs. I have a bunch of them, in both soft and hard grad. I really love the Lee filter system (although it failed me recently...i had my 2-stop proglass ND in...and when I was photographing rivers it just slipped out and shattered on the rocks...I am not sure why it did, but it was like a $160 filter. :(). The thing that set me off not long ago was a bunch of scenes where the skies ended up totally blown out when I exposed to preserve some detail in the shadows...and the skies were patchy...not along a nice horizon where a GND filter could actually be used to fix the problem.

I've bracketed crazy-wide like that as well, but in my experience, at least when you have bright highlights (like the highlights in water, or bright skies backing a dark foreground, or the sun in the frame), you end up with posterization or haloing if your exposures differ by a stop or more. Getting the exposure differential down to 2/3rds of a stop seems to smooth out the highlight transitions, so you don't end up with posteriation or funky CA or color issues or things like that after merging to HDR.

Ouch on that ProGlass ND, those things aren't exactly cheap :(

Strange thing about that bracketing range. I can usually push or pull my 6D files +1 2/3 or -1 2/3 w/o any trouble, no posterization or any decay. So one would think, you can easily use whole f-stops for bracketing w/o any detrimental effect on the IQ. Could it be more of a problem in the software? I honestly cannot believe, that one stop difference in editing would somehow crush any file from any recent DSLR. From the sound of it, seems like the problem lies in the software. But since I have only limited experience with Photomatix (can't really start to like this one) and HDRPro from CS6 (still learning to grasp of this one), so I may be wrong, but it just sounds so unlikely...
6D | 7D | 16-35/4L | 24-70/2.8L II | 50/1.4 (release 50/1.4 II already!) | 100/2.8L | 70-200/2.8L II

dtaylor

  • 5D Mark III
  • ******
  • Posts: 725
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #212 on: Today at 04:34:57 AM »
Normalized difference is 11.7 vs 14.8 for the D810...

And that's where any knowledgeable person stops reading. >14 stops...from a linear 14-bit ADC...kind of impossible  ;)

sarangiman

  • Canon 70D
  • ****
  • Posts: 298
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #213 on: Today at 04:46:18 AM »
Normalized difference is 11.7 vs 14.8 for the D810...

And that's where any knowledgeable person stops reading. >14 stops...from a linear 14-bit ADC...kind of impossible  ;)
Nope, that's where an unknowledgeable person stops reading, one who doesn't understand the math behind resampling.

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4563
  • POTATO
    • View Profile
    • Nature Photography
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #214 on: Today at 05:04:22 AM »
Thanks for the answer. I myself prefer to use LEE GNDs to avoid the need of too many files. Frankly, I'm quite bad in blending (learning about luminance masks at the moment as well as some other techniques) so I try to do all I can in the field so I can ideally work with single files.

But when I do bracket, I bracket like mad: -3,-2,-1,0,+1,+2,+3 is my usual starting point :D

I too use Lee's GNDs. I have a bunch of them, in both soft and hard grad. I really love the Lee filter system (although it failed me recently...i had my 2-stop proglass ND in...and when I was photographing rivers it just slipped out and shattered on the rocks...I am not sure why it did, but it was like a $160 filter. :(). The thing that set me off not long ago was a bunch of scenes where the skies ended up totally blown out when I exposed to preserve some detail in the shadows...and the skies were patchy...not along a nice horizon where a GND filter could actually be used to fix the problem.

I've bracketed crazy-wide like that as well, but in my experience, at least when you have bright highlights (like the highlights in water, or bright skies backing a dark foreground, or the sun in the frame), you end up with posterization or haloing if your exposures differ by a stop or more. Getting the exposure differential down to 2/3rds of a stop seems to smooth out the highlight transitions, so you don't end up with posteriation or funky CA or color issues or things like that after merging to HDR.

Ouch on that ProGlass ND, those things aren't exactly cheap :(

Strange thing about that bracketing range. I can usually push or pull my 6D files +1 2/3 or -1 2/3 w/o any trouble, no posterization or any decay. So one would think, you can easily use whole f-stops for bracketing w/o any detrimental effect on the IQ. Could it be more of a problem in the software? I honestly cannot believe, that one stop difference in editing would somehow crush any file from any recent DSLR. From the sound of it, seems like the problem lies in the software. But since I have only limited experience with Photomatix (can't really start to like this one) and HDRPro from CS6 (still learning to grasp of this one), so I may be wrong, but it just sounds so unlikely...

I use PS CC for HDR Merge, and ACR for 32-bit float (20-stop) toning. It's one of, if not the, best options out there for tone mapping HDR images. The posterization issue with highlights is a well known problem. Bracketing more shots is the only real solution. The ideal would be to get 1/3rd stop differences between each exposure, to minimize the differential in highlights between exposures. You can't really eliminate posterization that way, but it can make it a lot less noticable.

There are plenty of issues with HDR merge. Some people really love it, spend the time to do it right, and get great results. Personally, I don't like to spend that much time on each image. I'd rather have better data to start, that was more workable and flexible, and do what I could with a single file and maybe one single GND.
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

Khalai

  • EOS M2
  • ****
  • Posts: 164
  • Let there be (flattering) light!
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #215 on: Today at 05:14:53 AM »
I'd rather have better data to start, that was more workable and flexible, and do what I could with a single file and maybe one single GND.

+1, I, too, want to minimize the time I spend behind my Eizo. Everything above 15 minutes/photo is a drag, unless it's really remarkable image worth tweaking.
6D | 7D | 16-35/4L | 24-70/2.8L II | 50/1.4 (release 50/1.4 II already!) | 100/2.8L | 70-200/2.8L II

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4563
  • POTATO
    • View Profile
    • Nature Photography
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #216 on: Today at 05:20:32 AM »
I guess I am a bit of a purist as well. I don't like to massively modify my images. I like to take photos that represent reality. I prefer less saturated colors (not dull and overly muted, but not ultra vibrant or saturated either), and I prefer things to appear as they were to my own eyes as much as I can. To that end, getting things done in-camera as best as possible is important to me.

Banded read noise makes it tough to achieve that goal. Shadows don't have banded noise, and shadows are rarely as dark in real life as they appear on screen. So, better data is of particular importance to me.
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

canon rumors FORUM

Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #216 on: Today at 05:20:32 AM »

Khalai

  • EOS M2
  • ****
  • Posts: 164
  • Let there be (flattering) light!
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #217 on: Today at 05:47:31 AM »
I guess I am a bit of a purist as well. I don't like to massively modify my images. I like to take photos that represent reality. I prefer less saturated colors (not dull and overly muted, but not ultra vibrant or saturated either), and I prefer things to appear as they were to my own eyes as much as I can. To that end, getting things done in-camera as best as possible is important to me.

Banded read noise makes it tough to achieve that goal. Shadows don't have banded noise, and shadows are rarely as dark in real life as they appear on screen. So, better data is of particular importance to me.

I agree. But when you expose for the shadows (maybe a little overexpose and then pull back in PP) and you filter the highlights (LEE has even 1.2 GNDs, I'm really considering the purchase of both hard and soft, but since one of them would cost me around ~225USD, I'm not very eager about it) or bracket and then use luminance masking or whatever, then you won't have this problem. I agree that not all compositions can be solved this way (although soft GNDs are amazing), but many of them can.

I just did a quick dirty check on rather badly exposed file (6D, ISO 800) and overexposed it +2 EV and even added +50 in shadows sliders. Yes, there is noise, but no pattern, no banding. Is 5D3 really worse? Then I'm glad I don't need the AF of the 5D3 and be happy with my 6D :)
6D | 7D | 16-35/4L | 24-70/2.8L II | 50/1.4 (release 50/1.4 II already!) | 100/2.8L | 70-200/2.8L II

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4563
  • POTATO
    • View Profile
    • Nature Photography
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #218 on: Today at 07:04:52 AM »
I guess I am a bit of a purist as well. I don't like to massively modify my images. I like to take photos that represent reality. I prefer less saturated colors (not dull and overly muted, but not ultra vibrant or saturated either), and I prefer things to appear as they were to my own eyes as much as I can. To that end, getting things done in-camera as best as possible is important to me.

Banded read noise makes it tough to achieve that goal. Shadows don't have banded noise, and shadows are rarely as dark in real life as they appear on screen. So, better data is of particular importance to me.

I agree. But when you expose for the shadows (maybe a little overexpose and then pull back in PP) and you filter the highlights (LEE has even 1.2 GNDs, I'm really considering the purchase of both hard and soft, but since one of them would cost me around ~225USD, I'm not very eager about it) or bracket and then use luminance masking or whatever, then you won't have this problem. I agree that not all compositions can be solved this way (although soft GNDs are amazing), but many of them can.

I just did a quick dirty check on rather badly exposed file (6D, ISO 800) and overexposed it +2 EV and even added +50 in shadows sliders. Yes, there is noise, but no pattern, no banding. Is 5D3 really worse? Then I'm glad I don't need the AF of the 5D3 and be happy with my 6D :)

Yeah, 5D III is pretty bad. It's minimally improved over the 5D II at best. The banding might not be quite as bad, but it is still present in a significant way. The frustrating thing about it is that, unlike my 7D, Topaz DeNoise 5's debanding doesn't handle the 5D III banding well at all. With the 7D, set the vertical band separation to 8 pixels, and bam, gone. With the 5D III, it can't seem to find and remove all the bands. Sometimes it actually enhances some and removes others. The random read noise levels are also quite high. While banding is unsightly, it can be cleaned up with effort. The random read noise, however, is what actually eats away at the shadow detail. Your 6D has this too, although a bit less than the 5D III. The random read noise is where things differ hugely between an Exmor and a Canon sensor...Exmor just has FAR less. The 5D III has around 33-35e- worth of read noise, where an Exmor has around 3e-. That's a factor of 10-11.67x difference...at least a full order of magnitude. In terms of actual dynamic range, it's closer to two orders of magnitude. The reduction in read noise means that the vast majority of shadows in an Exmor are just random photon shot noise, which is very, very easy to manage.
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

raptor3x

  • EOS M2
  • ****
  • Posts: 172
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #219 on: Today at 09:50:09 AM »
Oh, and yes: it's unfortunate the A7 seems perform worse than the A7R at high ISO (~1/3 EV worse?). Not sure what happened there. Lower effective QE b/c of on-sensor PDAF? If this weren't the case, the A7 would probably be the better camera b/c it has on-sensor PDAF that helps it track (noticeably better than the A7S/R, which just hunt back and forth while trying to track, potentially leading to completely OOF images in AF-C), and EFC which means you can actually use it with telephoto lenses.

In terms of SNR I think the difference between the two is quite small, but something about the quality of the A7 noise doesn't respond well to LR's noise reduction and so in practice the difference is at least a full stop, if not more.

Also, any more detail on how you can get close to base ISO DR when using high ISO on the D810?
Bodies: 1Ds3,5D3,X-T1,A7 Lenses:  16-35L F4 IS, X 18-55 2.8-4, Σ 35 1.4 A, Σ 50 1.4 A, 24-70L II, 85 1.8, 100L IS, 135L,  70-200L F2.8 IS II

FEBS

  • EOS M2
  • ****
  • Posts: 189
  • Action & Sports
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #220 on: Today at 04:11:36 PM »
Normalized difference is 11.7 vs 14.8 for the D810...

And that's where any knowledgeable person stops reading. >14 stops...from a linear 14-bit ADC...kind of impossible  ;)
Nope, that's where an unknowledgeable person stops reading, one who doesn't understand the math behind resampling.

Or a person who thinks he understands. How could you get more then 14stops out of a 14bit ADC? No math can solve this. Yes you might interpolate, expect what the value might be, but that is not correct. 14 bits means really only 14 bits max theoretical, and practical it will be lower. There is no logic and no math to find the REAL values outside the sampled values.
Camera: 1Dx, 5D3, 7D, 300D  | Prime:14 2.8Lii, 24 1.4Lii, 50 1.4, 85 1.2Lii, 100 2.8Lmacro, 300 2.8Lii | Zoom: 16-35 4.0L, 24-105 4.0L, 70-200 2.8Lii, 200-400 1.4x 4.0L | TS-E 24 3.5Lii | Other: 1.4xIII, 2.0xIII, 2x YN568, 6x YN560iii,  YN622TX, 6x YN622, 2x YN560TX, 6x YN603, CamRanger, ...

sarangiman

  • Canon 70D
  • ****
  • Posts: 298
    • View Profile
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #221 on: Today at 04:29:40 PM »
Or a person who thinks he understands. How could you get more then 14stops out of a 14bit ADC? No math can solve this. Yes you might interpolate, expect what the value might be, but that is not correct. 14 bits means really only 14 bits max theoretical, and practical it will be lower. There is no logic and no math to find the REAL values outside the sampled values.

I'm not arguing that you absolutely have more than 14 bits of data even after normalization. First of all, I personally don't use image data from SNR=1 regions. So in my book the D810 doesn't really even have 14 stops of usable DR per-pixel. I'm just saying that even for an area that has SNR less than 1, the SNR goes up when you average multiple samples. So let's just say that both the 11.7 and 14.8 numbers are inflated, for the sake of argument. They're still comparable, b/c SNR goes up by the sqrt of samples averaged. The per-pixel values are not comparable, not for any reasonable sort of viewing.

jrista

  • Canon EF 400mm f/2.8L IS II
  • *******
  • Posts: 4563
  • POTATO
    • View Profile
    • Nature Photography
Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #222 on: Today at 04:40:30 PM »
Normalized difference is 11.7 vs 14.8 for the D810...

And that's where any knowledgeable person stops reading. >14 stops...from a linear 14-bit ADC...kind of impossible  ;)
Nope, that's where an unknowledgeable person stops reading, one who doesn't understand the math behind resampling.

Or a person who thinks he understands. How could you get more then 14stops out of a 14bit ADC? No math can solve this. Yes you might interpolate, expect what the value might be, but that is not correct. 14 bits means really only 14 bits max theoretical, and practical it will be lower. There is no logic and no math to find the REAL values outside the sampled values.

It's purely a noise thing. This is in a normalized context. I prefer to know the literal dynamic range of the hardware itself myself as well. That would be DXO's Screen DR measurement, which tells you the per-pixel dynamic range of a given sensor. For cameras with 14-bit ADC units, none of them have a Screen DR above 14 stops.

A normalized context is used to create a comparable basis. Comparing the noise levels of an image with smaller pixels to the noise levels of an image with larger pixels is ignoring the frame size. This is basic equivalence. Sensor size and quantum efficiency are by far the primary factors that affect noise levels in an image. Since a D810 has more total smaller pixels than say a 5D III with fewer total larger pixels, you have to resample the larger image to the same dimensions as the smaller image. By resampling, noise is averaged out, which reduces the per-pixel noise levels. For COMPARISON purposes, this is the only fair way of determining how one camera compares to another. Otherwise your comparing noise produced at a higher frequency with noise produced at a lower frequency...which is basically comparing apples to apples.

Downsampled comparisons have their limitations. For one, they can be misleading as to how much dynamic range on a sensor is actually usable on-scene. If your standing in front of a landscape, and you meter 14.8 stops of DR from the deepest shadows to the brightest highlights, you won't be able to capture that scene with a D810. Even though it's 8x12 "normalized" DR is 14.8, the hardware DR is 13.8. Your a full stop short with the D810, and you would either need to use a 1-stop GND to balance the scene DR, or use HDR, to capture the entire thing. Additionally, you won't be able to lift the single shot with the GND by seven stops in post (as would be indicated by a 14.8 stop Print DR number). You would be able to lift it by at most 5.8 stops, however trying to lift even a D810 shot that much is going to encounter read noise. Realistically, you probably have 4-5 stops of shadow lifting ability without unsightly read noise (which in the case of Exmor-based cameras, is pretty much just random color noise, still no banding.)

Sarangiman is talking about the normalized DR values from DXO. Those values are only valid if you assume an 8x12" 300ppi downsamplng target. When a 5D III and D810 are downsampled to that image size, the 5D III has  11.7 stops of DR and the D810 has 14.8 stops of DR (engineering DR...the raw measure from the RMS of noise to the saturation point...whether all of that DR is fully usable depends on too many factors, which completely reduces any comparison to mush. The actual usable range is dependent upon the tools you use to process, the precision of those tools algorithms, your capability at maximizing the effectiveness of those tools, your personal tolerances for noise, etc....so there is no real objective measure of "usable dynamic range.")
My Photography
Current Gear: Canon 5D III | Canon 7D | Canon EF 600mm f/4 L IS II | EF 100-400mm f/4.5-5.6 L IS | EF 16-35mm f/2.8 L | EF 100mm f/2.8 Macro | 50mm f/1.4
New Gear List: SBIG STT-8300M | Canon EF 300mm f/2.8 L II

canon rumors FORUM

Re: Tony Northrup - D810 vs. 5D Mk3
« Reply #222 on: Today at 04:40:30 PM »