So sayeth DXO: "The a7R II poops on the 5DS"

[3kramd5]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

Downsampling everything to 8mp clearly disadvantages the 50mp sensor.

Actually, it is advantageous to the higher res sensor, in that it improves via noise-reducing averaging the dynamic range, which is the big ticket item.

Yes, it helps in terms of noise reduction for the sake of charts but even if there is slightly more noise in a 50mp file based on a per pixel level the extra detail of the 50mp opens new possibilities such as being able to push noise reduction further.

Within the DXO scoring methodology, downsampling helps, and the more you downsample the better. For whatever reason they don't consider sensor resolution in their camera scores, and instead consider it in their lens scores.

 

[jrista]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

Downsampling everything to 8mp clearly disadvantages the 50mp sensor.

Actually, it is advantageous to the higher res sensor, in that it improves via noise-reducing averaging the dynamic range, which is the big ticket item.

Yes, it helps in terms of noise reduction for the sake of charts but even if there is slightly more noise in a 50mp file based on a per pixel level the extra detail of the 50mp opens new possibilities such as being able to push noise reduction further.

Within the DXO scoring methodology, downsampling helps, and the more you downsample the better. For whatever reason they don't consider sensor resolution in their camera scores, and instead consider it in their lens scores.

Downsampling always helps, not just with DXO. Downsampling averages pixel data together, which improves signal strength and reduces noise. That will always improve DR. Same reason why stacking multiple frames together when doing astrophotography improves signal strength and reduces noise.

 

[that1guyy]

Who cares?

You sound mad.

 

[3kramd5]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

Downsampling everything to 8mp clearly disadvantages the 50mp sensor.

Actually, it is advantageous to the higher res sensor, in that it improves via noise-reducing averaging the dynamic range, which is the big ticket item.

Yes, it helps in terms of noise reduction for the sake of charts but even if there is slightly more noise in a 50mp file based on a per pixel level the extra detail of the 50mp opens new possibilities such as being able to push noise reduction further.

Within the DXO scoring methodology, downsampling helps, and the more you downsample the better. For whatever reason they don't consider sensor resolution in their camera scores, and instead consider it in their lens scores.

Downsampling always helps, not just with DXO. Downsampling averages pixel data together, which improves signal strength and reduces noise. That will always improve DR. Same reason why stacking multiple frames together when doing astrophotography improves signal strength and reduces noise.

Down sampling always helps... if noise and DR are your metrics.

 

[that1guyy]

*Inserts some stupid sarcastic deriding Sony*
*inserts : face

*feels accomplished*

 

[3kramd5]

*Inserts some stupid sarcastic deriding Sony*
*inserts : face

*feels accomplished*

Meh, I don't need to defend them. They have $3200 of my money, minus NRE, components, production, shipping, and dealer cut.

 

[George D.]

Conclusion: 5D4 is expected to have DR on a par with competition. Meanwhile we need a same sample image of EF35mm 1.4L II shot with a 5DS and a A7RII for the record.

With all this testing I just realized there's room for another category: Nikon cameras with Canon lenses. Surpass the DxO mark 8).

 

[moreorless]

One ironic thing of course is that it looks like Sony has sacrifed based ISO DR for higher ISO DR, about half of the advantage the 36 MP sensor had at ISO 100 over Canon is gone.

As a Nikon shooter who might well have the chance to buy this sensor in the future I have to say that its performance really doesn't live up to the "game changer" hype, the same as that Samsung 28 MP APSC one. You look back a generation and the D800 sensor was a MUCH bigger improvement in performance over the previous gen(Canons 21 MP, Nikon's 12 and 24 MP). Resolution increased by 50% over Nikons previous best and by 200% over the camera with a similar market position, DR increased and general noise performance was half a stop better.

Sony really missed the boat not getting that sensor in one of their bodies first IMHO although I spose that likely comes an a premium which given there smaller sales probably explains the massive jump in the A7R II price.

 

[StudentOfLight]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

Downsampling everything to 8mp clearly disadvantages the 50mp sensor.

Actually, it is advantageous to the higher res sensor, in that it improves via noise-reducing averaging the dynamic range, which is the big ticket item.

Yes, it helps in terms of noise reduction for the sake of charts but even if there is slightly more noise in a 50mp file based on a per pixel level the extra detail of the 50mp opens new possibilities such as being able to push noise reduction further.

Within the DXO scoring methodology, downsampling helps, and the more you downsample the better. For whatever reason they don't consider sensor resolution in their camera scores, and instead consider it in their lens scores.

Downsampling always helps, not just with DXO. Downsampling averages pixel data together, which improves signal strength and reduces noise. That will always improve DR. Same reason why stacking multiple frames together when doing astrophotography improves signal strength and reduces noise.

Perhaps I don't understand downsampling and how averages work, but I constructed a simple experiment using excel. (see attached)

I took a black frame with a hypothetical 16 pixel camera. i.e. A perfectly accurate camera would take an image with a 0 value for each pixel. I then introduced a random noise value for each pixel. I then added pixels together first 2 pixel arrays then 4 pixel arrays then averaged to get the image noise. It appears the average image noise is the same regardless of how I scale the image down. Am I doing something wrong?

 

[Sporgon]

I think they could produce one for a $2200 retail price. Would Sony want them to when theirs is $1000 more ? I expect to see the a7rII come down in price pretty quickly.

Unless Pentax is selling them as a loss-leader to get people to buy into their system.

Possibly. Pentax have done that type of thing before: the 50mm f/1.4 eight element lens that cost more to manufacture than they sold it for, produced to raise them to the perceived level of Zeiss. And the Pentax genes do still seem to be alive in the latest digital cameras despite now being owned by Ricoh.

Regarding lenses, they do have a few FF 'limited' (high quality) primes and have just introduced two top end FF zooms.

For us Canon (and Nikon) FF dslr users who don't use the top tier cameras Pentax producing a FF dslr is very good news. It will be a feature rich camera for a good price and will keep the likes of the 6D and Nikon D610 ( or is it D630 by now ?) on their toes and keen on price.

Despite being a BLS I don't see why after development costs the R sensor will be much more expensive than the other 36 mp one. Having had a play with a Sony a7RII recently I think Sony are yanking the chain on the price. It seems to me that the major steps forward in the a7rII over the a7r is everything but the sensor itself.........

 

[romanr74]

We all refer to DXO on occasions...

No we don't...

 

[privatebydesign]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

Downsampling everything to 8mp clearly disadvantages the 50mp sensor.

Actually, it is advantageous to the higher res sensor, in that it improves via noise-reducing averaging the dynamic range, which is the big ticket item.

Yes, it helps in terms of noise reduction for the sake of charts but even if there is slightly more noise in a 50mp file based on a per pixel level the extra detail of the 50mp opens new possibilities such as being able to push noise reduction further.

Within the DXO scoring methodology, downsampling helps, and the more you downsample the better. For whatever reason they don't consider sensor resolution in their camera scores, and instead consider it in their lens scores.

Downsampling always helps, not just with DXO. Downsampling averages pixel data together, which improves signal strength and reduces noise. That will always improve DR. Same reason why stacking multiple frames together when doing astrophotography improves signal strength and reduces noise.

Perhaps I don't understand downsampling and how averages work, but I constructed a simple experiment using excel. (see attached)

I took a black frame with a hypothetical 16 pixel camera. i.e. A perfectly accurate camera would take an image with a 0 value for each pixel. I then introduced a random noise value for each pixel. I then added pixels together first 2 pixel arrays then 4 pixel arrays then averaged to get the image noise. It appears the average image noise is the same regardless of how I scale the image down. Am I doing something wrong?

You are not averaging, you are adding, to get the lower noise you have to divide by the number of pixels you have added together, in this case four. Take any one pixel of information and it has to fit in the bit depth, for 14 bit that is 16,385, if you add four together you can't have a number higher than 16,385, so you have to divide by that same number to keep your range constant.

So take your last block of four, if noise becomes visible, your noise floor, at 3 (for example) you have two noisy pixels px-ID-14 and px-ID-15, if you add the four together you get 8, then divide by four you get 2 per pixel, which you can't see. Voila, two pixels that had visible noise don't now have visible noise, but you have lost the ability to differentiate detail in those four pixels so you now have one noiseless pixel instead of two of four noisy ones.

To be sure, your DR has not increased in that you don't have a wider range, you can't see below your noise floor and the bit depth has not increased because add four and divide by four is a zero sum when confined to whole numbers. You have lowered the noise levels by averaging/downsampling though.

 

[romanr74]

People will lose it if a Pentax FF SLR comes out because it's Pentax. They've always been ahead of the curve on no AA, weathersealing, shockproofing, etc.

The problem is that it will have to almost be sold for a loss to get people hooked, wouldn't they? Pentax doesn't have much on the FF glass front, I thought. (Or do they have older film FF glass that they can use? I don't know much about them.) If they lack an FF lens portfolio, their value proposition will be painfully similar to the first Sony A7 models and they'll have to slash price for the first few generations until they amass more glass.

- A

There's a huge array of vintage Pentax/Takumar lenses that could be used, many of them without an adapter and many of them very impressive, but most of them are MF, which is rather a pain on a dslr, so their appeal would likely be limited (unless the forthcoming FF body has an EVF, which would make a huge difference in that regard; I have several and use them on Sony a7 bodies, which, like other mirrorless/EVF bodies make MF easy). Some current Pentax AF lenses are old enough that they were designed for FF, but they all use slow noisy screw AF. In fact, AF hasn't been a Pentax strong-point; I've not kept up with their more recent bodies but when I switched from their then-top-of-the-line body (K5) to a 5DII (!) a few years ago I was shocked by the superiority of Canon AF in terms of speed and accuracy. Even if the new camera has better AF accuracy, they're going to lag seriously in terms of AF speed as far as many/most of their AF lenses are concerned. Those who think the mark of a good camera is its usefulness at Olympics will likely want to hold off....

I have a terribly hard time to believe that these film area lenses will perform sufficently well on such high resolution sensors...

 

[StudentOfLight]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

You are not averaging, you are adding, to get the lower noise you have to divide by the number of pixels you have added together, in this case four. Take any one pixel of information and it has to fit in the bit depth, for 14 bit that is 16,385, if you add four together you can't have a number higher than 16,385, so you have to divide by that same number to keep your range constant.

So take your last block of four, if noise becomes visible, your noise floor, at 3 (for example) you have two noisy pixels px-ID-14 and px-ID-15, if you add the four together you get 8, then divide by four you get 2 per pixel, which you can't see. Voila, two pixels that had visible noise don't now have visible noise, but you have lost the ability to differentiate detail in those four pixels so you now have one noiseless pixel instead of two of four noisy ones.

To be sure, your DR has not increased in that you don't have a wider range, you can't see below your noise floor and the bit depth has not increased because add four and divide by four is a zero sum when confined to whole numbers. You have lowered the noise levels by averaging/downsampling though.

Hi PBD, thanks for the reply.

I don't know if I'm just retarded, but I still don't get it. I included a division process in the averaging my original spreadsheet here is an update I just changed the layout to put the averages in at the bottom of the table (see attached)

Is my concept of average image noise flawed (i.e. Average image noise = sum of pixel noise divided by number of pixels)

 

[sdsr]

I have a terribly hard time to believe that these film area lenses will perform sufficently well on such high resolution sensors...

It all rather depends on what you mean by "sufficiently well". If you mean the sort of technical near-perfection of a Zeiss Otus, maybe not, but some come close to a degree you might find surprising (the hyper-critical Otus fan Ming Thein, for instance, is a big admirer of the Contax/Zeiss series from the 1970s-90s; being able to use them is among the advantages he gives for the Canon 5DS line and, especially, the Sony a7rII). With fast lenses wide open you might often find plenty to complain about, but often minor stopping down makes them indistinguishable from good modern lenses, while their "flaws" wide open can add a degree of character/atmosphere that for certain sorts of photography is (for some of us, anyway) highly desirable. I use such lenses on my a7r/a7rII more than any other (it also doesn't hurt that they're often a pleasure to handle regardless of image quality, including the older Pentax/Takumars). There are lots of examples on-line of images taken with so-called legacy lenses on cameras with high-resolution sensors, both dslrs and mirrorless (where they're easier to use), including comparisons of vintage vs new.

 

[3kramd5]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

snip

I claim zero expertise with respect to signal processing, so I'm probably wrong, but:

It seems that you're treating everything as if it's noise, or everything as if it's signal. The total reading in any given pixel is both noise AND signal. If you consider them independently, when you average multiple pixels, both values will decrease, but noise will decrease more than signal.

 

[sdsr]

It seems to me that the major steps forward in the a7rII over the a7r is everything but the sensor itself.........

It seems that way to me, too - I moved up to an a7rII for the IBIS, silent shutter, vibration-fix etc. rather than for the sensor (indeed, the sensor isn't what attracted me to this line in the first place - I prefer mirrorless/EVF and FF and at the time Sony was the only company providing both), though it seems to me that the high ISO performance of the a7rII is somewhat better than the a7r's. (Of course, if it's true, as some are reporting, that the problem Leica etc. wide-angle M mount lenses work better on the a7rII, that will presumably be a big deal for those wanting to use such lenses.)

 

[bdunbar79]

Remember, Noise = SQRT #photons

So Noise = SQRT"signal"

Signal adds linearly while noise adds only in quadrant. Signal adds faster than noise can. The more signal you have, of course the more total noise you have, but you gain signal faster than noise increases.

 

[privatebydesign]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

You are not averaging, you are adding, to get the lower noise you have to divide by the number of pixels you have added together, in this case four. Take any one pixel of information and it has to fit in the bit depth, for 14 bit that is 16,385, if you add four together you can't have a number higher than 16,385, so you have to divide by that same number to keep your range constant.

So take your last block of four, if noise becomes visible, your noise floor, at 3 (for example) you have two noisy pixels px-ID-14 and px-ID-15, if you add the four together you get 8, then divide by four you get 2 per pixel, which you can't see. Voila, two pixels that had visible noise don't now have visible noise, but you have lost the ability to differentiate detail in those four pixels so you now have one noiseless pixel instead of two of four noisy ones.

To be sure, your DR has not increased in that you don't have a wider range, you can't see below your noise floor and the bit depth has not increased because add four and divide by four is a zero sum when confined to whole numbers. You have lowered the noise levels by averaging/downsampling though.

Hi PBD, thanks for the reply.

I don't know if I'm just retarded, but I still don't get it. I included a division process in the averaging my original spreadsheet here is an update I just changed the layout to put the averages in at the bottom of the table (see attached)

Is my concept of average image noise flawed (i.e. Average image noise = sum of pixel noise divided by number of pixels)

Yes, you don't average the first group.

So take your first four pixels, say the noise floor is 4, ID-2 and ID-4 are both noisy pixels, at 100% view those pixels are garbage. Add the four together and divide by four and the resulting value is 3, so that block of four pixels, that is now one number is no longer noisy, at 100% view that down sampled one pixel (the four have become one) is not noisy but the picture is 1/4 the size it was.

This is how multiple exposures reduces noise on a same size basis, take various exposures of the same thing, add them together and divide by the number of exposures and you get less noise and retain the number of pixels. Basic astrphotography.

Is my concept of average image noise flawed (i.e. Average image noise = sum of pixel noise divided by number of pixels)

Yes this concept is wrong. Remember, you are not adding up all the values and dividing by the total number of pixels, you are only adding the down sampled pixels together and dividing by that number of pixels to get a new averaged pixel value. Then the number of actual pixels that fall into the range of noise is lower, but so is detail!

P.S. You can't have decimal places in your averages (you'd need more bit depth), so in the adjusted table below the .5 and above would be rounded up, .49 and lower would be rounded down. So you'd actually have 6 - 3 - 1.

 

[StudentOfLight]

Re: Sony A7R II scores 98, new king of DxO Mark (by a nose)

You are not averaging, you are adding, to get the lower noise you have to divide by the number of pixels you have added together, in this case four. Take any one pixel of information and it has to fit in the bit depth, for 14 bit that is 16,385, if you add four together you can't have a number higher than 16,385, so you have to divide by that same number to keep your range constant.

So take your last block of four, if noise becomes visible, your noise floor, at 3 (for example) you have two noisy pixels px-ID-14 and px-ID-15, if you add the four together you get 8, then divide by four you get 2 per pixel, which you can't see. Voila, two pixels that had visible noise don't now have visible noise, but you have lost the ability to differentiate detail in those four pixels so you now have one noiseless pixel instead of two of four noisy ones.

To be sure, your DR has not increased in that you don't have a wider range, you can't see below your noise floor and the bit depth has not increased because add four and divide by four is a zero sum when confined to whole numbers. You have lowered the noise levels by averaging/downsampling though.

Hi PBD, thanks for the reply.

I don't know if I'm just retarded, but I still don't get it. I included a division process in the averaging my original spreadsheet here is an update I just changed the layout to put the averages in at the bottom of the table (see attached)

Is my concept of average image noise flawed (i.e. Average image noise = sum of pixel noise divided by number of pixels)

Yes, you don't average the first group.

So take your first four pixels, say the noise floor is 4, ID-2 and ID-4 are both noisy pixels, at 100% view those pixels are garbage. Add the four together and divide by four and the resulting value is 3, so that block of four pixels, that is now one number is no longer noisy, at 100% view that down sampled one pixel (the four have become one) is not noisy but the picture is 1/4 the size it was.

This is how multiple exposures reduces noise on a same size basis, take various exposures of the same thing, add them together and divide by the number of exposures and you get less noise and retain the number of pixels. Basic astrophotography.

I've ask the question before in a different thread regarding how "noise floor" is calculated but unfortunately got no reply. I thought that since this example is a dark frame that the average value of all the "supposed-to-be-zero-value pixels" would be the noise floor. Hence I calculated the average value of 2.5 initially.