p.3 #1 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
More interesting info from northlight-images.co.uk !
10th Several more people with testing experience of the new 5Ds have written to agree with some of the comments in the previous days. In particular, one commented that: "Canon's new 50.6mp sensor at low ISO will perform much better then any other EOS camera currently in the pro line. The colour filters on the sensor are designed to produce a higher level of colour accuracy and separation, the sensor itself runs at a significantly lower temperature. This will come at a price since high ISO performance will drop significantly. The new sensor will capture 14 stops of DR (just like the 5D III) ... However it will produce remarkably cleaner results when lifting deep shadows"....Show more →
p.3 #3 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
RobDickinson wrote:
' The new sensor will capture 14 stops of DR (just like the 5D III) .'
Which 5dIII was that?
Or according to what definition of dynamic range?
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
p.3 #4 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
alundeb wrote:
Or according to what definition of dynamic range?
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
That's a curious definition; it's like quoting the 0-60 acceleration rate of an engine without the drive train attached (the sensor vs the sensor complex/ADC). It also doesn't do single-exposure shooters much good except if they're using ML's dual ISO.
p.3 #5 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
RobDickinson wrote:
' The new sensor will capture 14 stops of DR (just like the 5D III) .'
alundeb wrote:
Which 5dIII was that?
Or according to what definition of dynamic range?
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
That would be fine and dandy if only the 5D3 (and all Canon sensors) had the same read noise at base ISO as they do at highest ISO. As matters stand, however, the read noise at base ISO is way, way, way higher than at the highest ISO, and that's why the 5D3 has 11 stops of engineering DR at base ISO:
p.3 #6 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
snapsy wrote:
That's a curious definition; it's like quoting the 0-60 acceleration rate of an engine without the drive train attached (the sensor vs the sensor complex/ADC). It also doesn't do single-exposure shooters much good except if they're using ML's dual ISO.
Yes, it is a curious definition but not entirely uncommon.
p.3 #7 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
takowasa wrote:
That would be fine and dandy if only the 5D3 (and all Canon sensors) had the same read noise at base ISO as they do at highest ISO. As matters stand, however, the read noise at base ISO is way, way, way higher than at the highest ISO, and that's why the 5D3 has 11 stops of engineering DR at base ISO:
p.3 #9 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
You don't willy-nilly limit higher ISO performance unless you've found some way to clean-up the lower end. If fixing/eliminating banding and shadow noise was one of the main targets for; fixing that ugly in the 5D3 sensor, like the 6D did to a degree, then many of the complaints will have been addressed, and DR although might not be technically wider, at least those darker areas might be lifted while preserving more detail.
These are niche bodies, are for those that don't shoot with high ISOs and if they have significantly cleaned up, eliminated banding and noise from 100 thru 6400, IMHO that's a good, very good advancement and along with the doubling of MPs could be stellar.
For most of what I shoot; gardens/flowers/plants and city/land/seascapes, still life subjects, high ISO or FPS isn't needed, in fact I usually shoot in good light, in my makeshift studio or in daylight. Unfortunately I'm somewhat restricted time wise, there aren't many early or late shooting sessions. My injury requires assistance to just get dressed and in my chair. This limitation really does suck but it has helped me greatly with time management and understanding light and how best to shoot when it's not so perfect, no magic.
So while I understand comments regarding the need for cleaner high ISO performance like Peter mentioned the 1DX has in abundance; it's simply stellar, I'm one who welcomes cleaner low ISO performance at the expense of super high ones.
It's my hope that Canon has other plans for the next 1DX, I doubt 50MP is part just yet, perhaps 24-36MP's with the same or even cleaner higher ISO with a killer rate of 14-16 FPS. Maybe their consolidation of the 1D series was a mistake and we'll see two 1D bodies, one each dedicated, designed specifically for action and non-action much like prior.1D's were.
We'll just have to see how these new 5D images pan out, I hope they're stellar! And unfortunately with Canon sensors a one size fits all just doesn't seem possible yet, so the 'perfect' camera for every subject will not be.. Who knows what the 5D Mark IV or 1D X Mark II will bring us, 'maybe closer to perfect', only Canon has those plans!
p.3 #10 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
RobDickinson wrote:
' The new sensor will capture 14 stops of DR (just like the 5D III) .'
alundeb wrote:
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
An Imaging Resource review of the 5D3 says that its total dynamic range is not much less than the Nikon D800 but goes on to demonstrate that the 5D3's shadow noise makes much of that range unusable. Their quality dynamic range figures more closely match the clear Nikon/Canon low ISO dynamic range difference. Is shadow noise the same as read noise, i.e. is that where it would be noticeable?
Indeed the Sensorgen data show that the 5D3 low ISO read noise is more than 5 times that of the D800. By ISO 800 the 5D3 and D800 have about the same low read noise values.
The anecdotal quote that Rob repeated above goes on to say that the 5Ds has much less read noise than the 5D3 which might yield a better quality dynamic range to use Imaging Resoure's term. Lets hope. I remember reading the same comment about the 7D2's low iso read noise compared to the 7D but haven't seen dramatic difference in the test data yet. I'll add that 7D2 users seem happier in general than did 7D users.
p.3 #11 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
RustyBug wrote:
Late to the party (and a bit OT),
I couldn't find a metamerism index for the Kodak SLR/C. I'd be curious to know what it is, so I'd have some basis for how much (color accuracy) gain / loss I'd occur when I should finally update to something more current ... like the 5DS or 5DSR.
I wouldnt be suprised if it was somewhere close to 90. Nobody tested it (or I havent found it). By default, it doesnt have accurate colors, but it separates quite insane amount of shades/hues and has really pleasing (Kodak-like) output. Also Kodak did some crazy things, like per-pixel color correction built into firmware (based on each camera sensor). I dont think that any manufacturer today would do something like that.. (Leica can, but they dont seem that fond of colors). Plus they didnt have only good CFA, but sensor itself had specific response to each wavelength of light. No wonder it cost that much when new..
But anything over 85 SMI is usually "good enough". A900 seems to be pretty close to SLR/n or c.
Really shame manufacturers care usually so little about colors, while most of photos is done in color..
If 5Ds is going to have 1Ds MK3 CFA, then it will become quite interesting camera for me.. Thats if I needed 50 mpix, which I dont.
p.3 #12 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
alundeb wrote:
Or according to what definition of dynamic range?
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
So this might explain why chuck said "same Dr as 5d3" but in reality it has more usable Dr.
p.3 #13 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
+1 @ GL finding info on it.
I did find a thread which suggested it's DR is around 11.x ... about on par with contemporary MFD and Canon's latest. Not quite up to Sony's latest & greatest, but for an 11 year old camera ... how can I complain.
Better color and can't complain @ DR ... Canon hasn't given me a reason to set it down just yet. Granted, ISO / noise becomes an issue extraordinaire once you get off base ISO, but I'm used to shooting ISO 50 @ film, so 160 is "high speed" stuff for me.
We'll see what the 5DS / 5DSR brings ... it'll be interesting to see by how much it can best the SLR/C @ color & DR @ best vs. best. Granted, it will definitely be more versatile on many levels.
p.3 #14 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
alundeb wrote:
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
Yes, and that is absurd. I never did understand the reasons behind such calculation.
At high ISO the sensor doesn't reach the saturation. It has no sense at all to compare the signal at the highest saturation level (ISO 100) with the little read noise detected when only 1/64 of the sensor capacity is in use (ISO 6400). On Canon sensors, which use a distant off-chip ADC, the read noise mostly depends on the sensor saturation: the more the signal registered, the more the noise introduced while delivering it to the ADC. The day Canon will fix this problem, yes, we may truly get by then the 15 stops.
p.3 #15 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
RobDickinson wrote:
' The new sensor will capture 14 stops of DR (just like the 5D III) .'
alundeb wrote:
R.N Clark at clarkvision.com defines "sensor dynamic range" as the ratio between full well capacity at lowest ISO and read noise at highest ISO, at the pixel level. That definition gives the 5DIII 14.7 stops of dynamic range.
Jeff Nolten wrote:
An Imaging Resource review of the 5D3 says that its total dynamic range is not much less than the Nikon D800 but goes on to demonstrate that the 5D3's shadow noise makes much of that range unusable.
I'm not sure what that means, since the shadow noise is the noise floor used for the calculation of the DR: DR = log2 (saturation / noise floor).
Their quality dynamic range figures more closely match the clear Nikon/Canon low ISO dynamic range difference. Is shadow noise the same as read noise, i.e. is that where it would be noticeable?
The read noise is the noise added by the sensor and other hardware -- it does not include photon noise. When the read noise is used as the noise floor, it is usually called "engineering DR" and is also usually given as a per-pixel measure.
Indeed the Sensorgen data show that the 5D3 low ISO read noise is more than 5 times that of the D800. By ISO 800 the 5D3 and D800 have about the same low read noise values.
Yep. That's why Canon's low ISO DR lags so far behind Sonikon but does just as well, and even better on occasion, at high ISO.
The anecdotal quote that Rob repeated above goes on to say that the 5Ds has much less read noise than the 5D3...
I had not heard that. It would be wonderful, if true, but I really doubt it.
...which might yield a better quality dynamic range to use Imaging Resoure's term. Lets hope. I remember reading the same comment about the 7D2's low iso read noise compared to the 7D but haven't seen dramatic difference in the test data yet. I'll add that 7D2 users seem happier in general than did 7D users.
which help a lot for high ISO performance. At higher ISOs, the pixel read noise is also a bit lower further improving high ISO performance. However, while its pixels also have a greater FWC over the 7D, they also have greater read noise at base ISO, which results in the same base ISO DR. All in all an overall improvement, though.
More importantly, however, is that, like the 6D vs 5D3, the 7D2 has significantly less banding which makes the DR it has significantly more usable.
p.3 #17 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
I would guess a 5D2 DR at its best. At wide DR scenes, less shadows noise and better colors are only two things I hope for this coming 5Ds/r. Here ya go:
p.3 #18 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
Read noise, shadow noise, noise floor, DR = log2 (saturation / noise floor), charts, graphs... Holy cow my head is spinning!
I do appreciate the effort from you guys / gals because I'm really trying to learn. And, it's obvious you guys know your stuff. However, I have a much simpler solution - just deliver my pre-ordered 5DsR asap and I can tell you guys in about an hour if it's any good or not. Yes, it's the old-fashion way by actually using it but hey... Okay, I've been wanting to say that in about ten different 5Ds threads I'm following. I feel much better now. Carry on...
p.3 #19 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
Jim Servies wrote:
Read noise, shadow noise, noise floor, DR = log2 (saturation / noise floor), charts, graphs... Holy cow my head is spinning!
I do appreciate the effort from you guys / gals because I'm really trying to learn. And, it's obvious you guys know your stuff. However, I have a much simpler solution - just deliver my pre-ordered 5DsR asap and I can tell you guys in about an hour if it's any good or not. Yes, it's the old-fashion way by actually using it but hey... Okay, I've been wanting to say that in about ten different 5Ds threads I'm following. I feel much better now. Carry on... ...Show more →
No need for your head to be spinning. What you're reading is the language of people who make or measure sensors for a living or are in related engineering fields where they've learned the lingo related to sensors. It's like a chemist going to the Makeup Forum and talking about the chemistry and charts underlying various lipstick brands. Interesting if you have to produce lipstick at the factory, but not that interesting if you're deciding which lipstick to wear. If you're a photographer, you learn much more from using a camera than from reading this stuff and trying to "learn" it from people online. That's why, as you say, use a camera for an hour and you'll know whether it's good or not.
p.3 #20 · 5DSR Low ISO DR 1.5-2 stops better than 1DX ...
If we presume the 5Ds has the same pixel architecture and per-pixel read noise as the 7DM2, the 5Ds's engineering ISO 100 dynamic range will be .66EV higher than the 7DM2 on a normalized sensor-area/MP basis. The 7DM2 ISO 100 DR is 11.11EV at its native resolution (the "screen" tab in DxO). We can calculate the 5Ds's DR by normalizing its 50.6MP down to the 7DM2's 20.17MP, using the following formula from DxO:
= log2(sqrt(50.6/20.17))
= log2(1.58388)
= 0.663463 EV
The DR normally quoted by others from DxO is the "print" value, which is further normalized to 8MP, using the same formula as above but sqrt(native_mp/8_mp). The 7DM2's 8MP-normalized DR is 11.78EV, which means the 5Ds's 8MP-normalized DR would be 12.44EV. By comparison, the 1DX's 8MP-normalized DR is 11.77EV.
