--
-
"Confused? You won't be, after this week's episode of... Soap."
Speaking of soap like products I just looked at the DXOmark test figures for the A7 and A7R.
Colour depth: A7=24.8 bits... A7R=25.6 bits
Dynamic range: A7=14.2 EVs... A7R=14.1 EVs
Low light ISO: A7=2248... A7R=2746
I have read the use case score explanation here to try to better understand the difference on the low light ISO score.
Everything else being equal I would off hand think that the sensor with the lower amount of pixels (when the images from the higher megapixel sensor camera are not down sampled) would be better (the D600 scores marginally higher than the D800 for instance).
But according to DXOmark this is not the case with the A7R scoring higher than the A7.
Can somebody tell me in layman terms what I am missing here? Maybe DXOmark do downsample or...?
Thanks in advance
Xpanded
Speaking of soap like products I just looked at the DXOmark test figures for the A7 and A7R.
Colour depth: A7=24.8 bits... A7R=25.6 bits
Dynamic range: A7=14.2 EVs... A7R=14.1 EVs
Low light ISO: A7=2248... A7R=2746
I have read the use case score explanation here to try to better understand the difference on the low light ISO score.
Everything else being equal I would off hand think that the sensor with the lower amount of pixels (when the images from the higher megapixel sensor camera are not down sampled) would be better (the D600 scores marginally higher than the D800 for instance).
But according to DXOmark this is not the case with the A7R scoring higher than the A7.
Can somebody tell me in layman terms what I am missing here? Maybe DXOmark do downsample or...?
Thanks in advance
Xpanded
willie_901
Mentor
Without knowing the technical details of the technologies used in both cameras' data streams, all you could do is speculate.
Pixel size is only one of many factors that determine the the final signal-to-noise ratio. For instance, SNR differences could be due to read noise differences in the analog to digital converters. It's possible the color-filter-array lenses of A7R transmit more light (more light = more signal). So "everything else" must not be equal. Any pixel size advantage is offset by other factors.
In digital imaging we tend to focus our thoughts on noise. But this is only half the story. The signal level is just as important.
However the link you provide states, "A difference in low-light ISO of 25% represents 1/3 EV and is only slightly noticeable."
The difference here is 19%.
Finally, the largest contribution to noise in contemporary CMOS sensors is shot noise. This noise is a consequence of fluctuations inherent in the conversion of light amplitude to electrical charge. This is why a perfectly exposed (at base ISO) blue sky has noise when examined at extreme magnifications. Shadow regions are affect more by read noise from the data stream electronics.
Pixel size is only one of many factors that determine the the final signal-to-noise ratio. For instance, SNR differences could be due to read noise differences in the analog to digital converters. It's possible the color-filter-array lenses of A7R transmit more light (more light = more signal). So "everything else" must not be equal. Any pixel size advantage is offset by other factors.
In digital imaging we tend to focus our thoughts on noise. But this is only half the story. The signal level is just as important.
However the link you provide states, "A difference in low-light ISO of 25% represents 1/3 EV and is only slightly noticeable."
The difference here is 19%.
Finally, the largest contribution to noise in contemporary CMOS sensors is shot noise. This noise is a consequence of fluctuations inherent in the conversion of light amplitude to electrical charge. This is why a perfectly exposed (at base ISO) blue sky has noise when examined at extreme magnifications. Shadow regions are affect more by read noise from the data stream electronics.
--
-
Hi Willie. Yeah, I read the 1/3 EV statement. Food for thoughts.
Thanks for the last paragraph. I did not know that.
I just cannot quite fathom why Sony cannot extract (small as the difference is) the same level of ISO performance from the A7 that Nikon can get from the D600.
Cheers,
Xpanded
defconfunk
n00b
It isn't always "can't", sometimes it is about trade offs. One of the advantages the A7 has over the A7r is PDAF pixels; it can do Phase Detection AF on chip, the A7R is limited to Contrast Detection, or the use of an adapter with a semi-transparent mirror and it's own PDAF sensor. The addition of the PDAF incurs a cost in silicon space for the sensors themselves and the wiring to get their information where it needs to be. Sony decided that the A7 would be a more well-rounded camera, knowing that the addition of PDAF on chip could cost it image capture capabilities, the A7r in contrast was about getting as good a sensor as possible.
In DSLRs like the D600, the AF sensors are mounted in the prism, so they leave more space on the sensor for image capture. The trade-off comes in having a much larger and more complicated to manufacture assembly.
willie_901
Mentor
Why should SONY worry about this? SONY's product managers might assume the signal-to-noise difference between the D610 and the A7 won't affect market share. Perhaps they decided the price-to-benefit ratio to equal the D610 was not the wisest way to spend their budget. To echo defconfunk, every camera is a compromise.
--
-
Why should not Sony worry? Why not make the best product you possibly can? That is what keeps customers coming back.
Last edited:
--
-
All true, but it makes the A7 vulnerable to people buying solely (or at least partly) based on DXOmarkings. The D600 and A7 were announced 13 months apart which is a long time in sensor technology. I cannot help thinking this is the NEX way of doing things again. Churn out new editions with only minor changes. Sony needs to make it obvious that they mean it seriously long term if they are to take a significant part of the CaNikon crowd.
user237428934
User deletion pending
Where is the need for best high Iso in a camera with contrast AF when It's not able to nail focus in a dimly lit location? Compared to a D610 and 6D the low light abilities of the A7 are just sh*t.
--
-
Because it has peaking and magnification so that you can focus manually.
user237428934
User deletion pending
Ah. Right. I always forget that some people use this camera with non native lenses.
--
-
NazgulKing
Established
That's not what happens in practice. Rather they will make the best camera that they can while maximising profit return and fulfill consumer requirements as best as possible, with a stronger preference for the former.
ray*j*gun
Mentor
I agree....its laughable to think that SONY is not capable of surpassing every benchmark that we know how to measure.... but that costs either money or size/weight. Its about bang for the buck IMO and there will always be those who want more of the bang.
--
-
Not trying to pick a fight
but this sounds very well in theory - not so sure in practice.What we have seen until recently were camera companies being quite good at minimizing production costs (always the easiest part to estimate). With the M8 (IR), M9 (broken sensors), D600 (dust), D800 (AF misalignment), various (non current) Canons (AF-issues) this does not seem to go so well any more. Classic examples of taking cost cutting too far. The profit must be seen over the entire life span of the product and not just by what the customer pays. The bean counters and engineers of this world should unite :angel:
The camera manufacturers have in recent years totally misjudged the market. Most of them producing far more than they can sell. I am still much obliged to Nikon for selling V1s at ridiculous prices and Sigma with the DP Merrill line (there are many more examples), but it hardly constitutes manufacturers fulfilling buyer's needs.
The one exception was surely the X100 which worldwide had sold 130,000 cameras for Fuji (even with sticky aperture blades) when the X100s came out. The feedback I have from Fuji is that it was a huge positive surprise to them.
The Nikon V1 was a slightly different story. Nikon went all out with the high pixel count of D800 while (more or less) at the same time making a low megapixel camera with a small sensor. Even for the majority of camera purchasers who just buy it in a Lowsalarymart it did not resonate. Somewhat ironically it is the perfect family camera that no one bought except me - I bought 3
Then there is the SD1 non-Merrill. Which buyers were they listening to? The target market of Hassy's ridiculous buy a camera and get a piece of Swedish (not even Norwegian!) wood?
I have owned (or own) registered cameras from all current manufacturers bar Samsung and Pentax. I have never been contacted about anything by any of the major brands. With the exception of Fuji and to some extent Sigma I have never seen proof of the companies listening to anyone other than a handpicked group. But there are plenty of examples of online reviewers becoming personae non gratae when voicing reasonable criticism.
Even Nikon only owned fully up with the D600 problems after they were sued. The only time I have seen questionnaires like the ones camera manufacturers should make, they were made by Thom Hogan.
On the other hand I have had plenty of support requests ignored by camera manufacturers. Nikon DK has provided excellent support though I must haste to add for fairness sake.
Reading interviews by DPreview it does not seem (from my point of view at least) that the feedback loop is working seamlessly. I surely did not ask Sony to alter their hotshoe on the NEX cameras so that my external microphone does not work on newer models :bang: And who asked Sigma to make the original DP as well as the SD1 so slow? And to the guy (surely no woman could be that stupid
) who asked Olympus to leave out lens hoods and make the E-M5 viewfinder piece defect: I am not totally enamoured with you mate!With the breadth and width of the products launched by Sony it does not seem like they have focused 100% on what to do. It looks more like shotgun rain. The A7/A7R combo does indeed look like a valid response to the smaller and lighter wish, but dishing the NEX-brand which had finally gain worldwide recognition does not seem like a coherent listening-making-selling-bettering process. How many lines does Sony have today? And how many have they promised to maintain 5 years from now?
/Xpanded
--
-
I'll happily walk into your trap of irony
with this question: How much more would it have cost Sony in production costs to add 1/3 of an EV stop in high ISO performance?And how much would that have extended time to market?
Until we know this, how can we talk about bang for bucks?
Would the sales price have been 1,717 USD instead of 1,698 USD? Or 1,899? Or 1,649 USD because the new fabrication methods of the next generation sensor are cheaper?
And how much would this have influenced sales worldwide?
And how is the production cost of the A7 measured? Are the tools specific and die with the line? Or is some kind of activity based costing used?
/Xpanded
YYV_146
Well-known
IMO there are physical limitations as to how good high ISO performance can be with a mirrorless body. The sensor is always "on" and the body is smaller, hence a smaller thermal envelope and the need for a higher "operating temperature" and more noise at higher ISO. In practice I have only occasionally noticed the difference from a D610, but I still feel comfortable enough going to 6400 or even 8000 on the A7 if the situation demands it.
For the A7 cost was obviously cut on the sensors. As far as Sony sensor iterations go this is a pretty old one (and I say this owning two A7s), and production numbers from the A99, RX1 and D600 have probably cut fabrication costs down so drastically that they can sell the body so cheaply. The A7r's 36MP sensor is most likely considerably more difficult to manufacture...
Also, the $1700 price tag is pretty arbitrary, since only in the US and Europe are Sony products as expensive - The A7 retails at a little over $1400 in China, and about $1250 in Hong Kong!
YYV_146
Well-known
I'll happily walk into your trap of irony
And how much would that have extended time to market?
Until we know this, how can we talk about bang for bucks?
Would the sales price have been 1,717 USD instead of 1,698 USD? Or 1,899? Or 1,649 USD because the new fabrication methods of the next generation sensor are cheaper?
And how much would this have influenced sales worldwide?
And how is the production cost of the A7 measured? Are the tools specific and die with the line? Or is some kind of activity based costing used?
/Xpanded
My understanding is that the A7 is very cheap to make. Sensor costs are unknown, but apart from the sensor and EVF nothing on the camera is actually worth much... I suspect the price difference (to the 6d and D610)
is mostly caused by the lack of a mirror box and the prism, as well as the on sensor AF.
willie_901
Mentor
The differences you are concerned about are trivial. They have no practical impact on getting your work done. Your images' content is orders of magnitude more important. Your comfort level with how the camera operates and if you enjoy using it is also much more important.
uhoh7
Mentor
DXO is very silly.
I have a7 which I quite like. Soon after, I also bought an m9.
Read the DX oh scores and you would think the a seven is a better sensor. Yet I now shoot almost 90% with the M9. That's in spite of, not because of, the focusing system.
M9 is a lot sharper and I like everything about the images created. I like the colors, and the dynamic range is perfectly fine once adjusted in light room.
The D XO scores of the M9 are simply very very misleading, for real world photography.
I will never trust their scores for anything again. I'm not sure what they are measuring, but it has nothing to do with creating images.
I have a7 which I quite like. Soon after, I also bought an m9.
Read the DX oh scores and you would think the a seven is a better sensor. Yet I now shoot almost 90% with the M9. That's in spite of, not because of, the focusing system.
M9 is a lot sharper and I like everything about the images created. I like the colors, and the dynamic range is perfectly fine once adjusted in light room.
The D XO scores of the M9 are simply very very misleading, for real world photography.
I will never trust their scores for anything again. I'm not sure what they are measuring, but it has nothing to do with creating images.
NazgulKing
Established
Well, that's the reason why I said "with a strong preference for the former". Companies are more inclined to resort to maximizing profits and to offload consumer requirements under most circumstances. Just looking at the spate of complaints made by a few A7/7r users over the light leak issue is another example of lose tolerances in the name of cheaper manufacturing.Not trying to pick a fight
What we have seen until recently were camera companies being quite good at minimizing production costs (always the easiest part to estimate). With the M8 (IR), M9 (broken sensors), D600 (dust), D800 (AF misalignment), various (non current) Canons (AF-issues) this does not seem to go so well any more. Classic examples of taking cost cutting too far. The profit must be seen over the entire life span of the product and not just by what the customer pays. The bean counters and engineers of this world should unite :angel:
The camera manufacturers have in recent years totally misjudged the market. Most of them producing far more than they can sell. I am still much obliged to Nikon for selling V1s at ridiculous prices and Sigma with the DP Merrill line (there are many more examples), but it hardly constitutes manufacturers fulfilling buyer's needs.
The one exception was surely the X100 which worldwide had sold 130,000 cameras for Fuji (even with sticky aperture blades) when the X100s came out. The feedback I have from Fuji is that it was a huge positive surprise to them.
The Nikon V1 was a slightly different story. Nikon went all out with the high pixel count of D800 while (more or less) at the same time making a low megapixel camera with a small sensor. Even for the majority of camera purchasers who just buy it in a Lowsalarymart it did not resonate. Somewhat ironically it is the perfect family camera that no one bought except me - I bought 3
Then there is the SD1 non-Merrill. Which buyers were they listening to? The target market of Hassy's ridiculous buy a camera and get a piece of Swedish (not even Norwegian!) wood?
I have owned (or own) registered cameras from all current manufacturers bar Samsung and Pentax. I have never been contacted about anything by any of the major brands. With the exception of Fuji and to some extent Sigma I have never seen proof of the companies listening to anyone other than a handpicked group. But there are plenty of examples of online reviewers becoming personae non gratae when voicing reasonable criticism.
Even Nikon only owned fully up with the D600 problems after they were sued. The only time I have seen questionnaires like the ones camera manufacturers should make, they were made by Thom Hogan.
On the other hand I have had plenty of support requests ignored by camera manufacturers. Nikon DK has provided excellent support though I must haste to add for fairness sake.
Reading interviews by DPreview it does not seem (from my point of view at least) that the feedback loop is working seamlessly. I surely did not ask Sony to alter their hotshoe on the NEX cameras so that my external microphone does not work on newer models :bang: And who asked Sigma to make the original DP as well as the SD1 so slow? And to the guy (surely no woman could be that stupid
With the breadth and width of the products launched by Sony it does not seem like they have focused 100% on what to do. It looks more like shotgun rain. The A7/A7R combo does indeed look like a valid response to the smaller and lighter wish, but dishing the NEX-brand which had finally gain worldwide recognition does not seem like a coherent listening-making-selling-bettering process. How many lines does Sony have today? And how many have they promised to maintain 5 years from now?
/Xpanded