EOS-1D X Mark II Dynamic Range [CR2]

[gsealy]

I like all the complaining, wishing, and speculating. It tells me what people are thinking and what is probably coming down the line sooner or later. Many of these posts lead to some detailed technical discussions. Not all people express themselves nicely, but that's okay. I can work around all that.

 

[sanj]

It was just a matter of time.

It's STILL a matter of time. Canon's concept of DR, and the actual measurements of DR, have thus far proven to be very different things. It will only "have been" a matter of time when it's actually a fact in a physical camera we can buy and hold in our hands.

In time.

 

[sanj]

As a long time occasional lurker I just dropped in to say what a bunch of insufferable arsetwats the majority of you are.

I've never seen such an argumentative, self absorbed, smug, infantile and downright pathetic group as that which regularly posts here. Nothing is more important than being right, starting a fight, arguing over nothing or scoring points.

Never mind the signal to noise ratio of the new vaporware 1DXII, how about y'all try to improve it here on this site.

It's not been a pleasure folks, you can stick it up yer fundamental.

Love & kisses,

bollo

I'm always wondering if these are reappearances of ancient personalities. I've even forgotten who most of them are, but apparently they don't like Canon.

Another one for the road:
Nikon Lens Selection Sucks, Nikon Menues Suck, Nikon Live View Sucks, Nikon Control Dials Suck, Nikon Editing Software Sucks, Nikon Chromatic Aberration Sucks, Nikon Video Mode Sucks, Nikon Lens Naming Sucks...
Sony has flimsy lens mounts and is indecisive, Samsung uses 12bit RAW in burst mode, The 654Z cost more than a 1Dx, Oh and Nikon doesn't have an RT flash, and Nikon's wireless functionality Sucks. Fuji only makes crop sensors and Micro 4/3 is even worse! But as bad as they are, it's still not as bad as the Nikon 1 series.
Nikon, Really, Sucks.

Feeling better? Must be!

 

[ScottyP]

This is such a dramatic jump, from out of nowhere, that I don't think I believe the rumor.

Not only that, but if there really were a giant tech/fab leap about to happen, would Canon premiere a drastically new, drastically different sensor on a 1-series camera? Usually they are conservative with the 1-series sensors and they try out new stuff lower down the product chain (dual pixel AF in the 70d, for example), right?

If this rumor IS true, I wonder if it is a Sony sensor, or at least some kind of Sony collaboration.

 

[K]

This is such a dramatic jump, from out of nowhere, that I don't think I believe the rumor.

Not only that, but if there really were a giant tech/fab leap about to happen, would Canon premiere a drastically new, drastically different sensor on a 1-series camera? Usually they are conservative with the 1-series sensors and they try out new stuff lower down the product chain (dual pixel AF in the 70d, for example), right?

If this rumor IS true, I wonder if it is a Sony sensor, or at least some kind of Sony collaboration.

Every time a new Canon is on the verge of release, there is speculation of a Sony sensor.

Canon has shown that they will do everything possible to not use a non-Canon sensor. Even if that sacrifices some ISO and DR performance.

This could be a pride thing. It could be the economics of paying another company for the sensor. It could be anything. Canon doesn't disclose their reasoning. But the fact remains, Canon prefers to use their own sensors.

Now, Canon sensors are not bad at all. For being "behind" on specs and all the nerd-tests - they are either the equal of or the superior of other sensors on the market when it comes to final IQ result. Photos from Canon are great. Simple as that.

Whether that is due to other factors (processing, lenses etcetera) - that matters not to me. Because the camera is a SYSTEM. As a photographer, I only care about the photo I get out of the camera. If Canon is doing that with magic fairy pixel dust algorithms, or with a better sensor - I don't care. I get a RAW file that looks fantastic.

No one shoots photos with just a sensor. The Sonikonians and DXO monkeys seem to conveniently forget that they aren't walking around with ONLY a CMOS. There's a camera body, image processing, electronics, lenses and much more that delivers that final IQ result.

I do concede that in lab tests of the most extremes, the Sony / Nikon sensors come out on top by a very slight margin. But again and again, this comes up in real photography almost never. The gap between Sony/Nikon and Canon would have to be very significant to lead to practical IQ differences in most images.

This difference is NOT seen within the same generation or even within the span of 2 generations in some instances.

You would have to compare a 2-3 generation old camera of the same line/level to have noticeable differences in IQ without pixel peeping.

 

[deleteme]

So just what is the theoretical limit of DR assuming 16 bit capture?
I am also guessing that the usable DR is also affected by the assumptions of base noise levels.

In the end it seems that while we will see improvements, most of us will be underwhelmed by the arrival of DR nirvana.

 

[TeT]

it could be Pride... that would explain a lot of Canon's product decisions.

 

[K-amps]

They just did it on the C300 II. 15 stops of DR, same 14bit RAW files. I assume then some logarithmic encoding. And no, stop with this Sony and Samsung stuff. Canon made the sensor in the C300II themselves on what is obviously a new fab process. The 1DX2 will use the same process.

I wonder how they would choose to encode an image with more than 16 stops of DR.

There are really two ways.

One is to use a linear encoding with more than 16 bits.

The other is to use a non-linear (logarithmic) encoding the way the new Lightroom HDR merge works (16 bit floating point for about 30 stops).

or save as an 18 bit (or more) RAW file

Isn't that the first one I said?

I doubt they will use Sony Sensors, or Samsung etc... Face saving and all.... The C300 mk.ii uses a sensor with similar QE/SN as the 6D. The difference would be the DPAF. Those extra pixels can be used to expose the shade areas for example; expose half the file at +4ev, and the non DPAF pixels expose at 0 EV; then the RAW file is cooked to spread this exposure latitude over one file (i.e Average out normal pixels and the DPAF ones shot at +4Ev. Very much like the ML Dual_iso hack. Except that this will be done at a hardware layer, meaning no loss of resolution like in Dual_iso. This will certainly increase DR, but I do not expect a commiserate increase in high ISO performance (there might be some Digic magic in play ofcourse) but for RAW files, I doubt it will be much more than the 5D3... unless they use both Sony tech plus DPAF, then Nikon fanbois can kiss our combined rears....

I hope they use 16bit files with greater color depth as well... I struggle capturing deep reds in flowers with my 5D3, there is of loss of detail. Other channels are handled better. I do not recall my old 5Dc struggling with reds...

 

[ritholtz]

...
Every time a new Canon is on the verge of release, there is speculation of a Sony sensor.
...

Yup and Zeiss appear to have mistakingly released a pic showing Sony with a 56MP sensor:
http://www.sonyalpharumors.com/sr4-did-zeiss-just-leak-the-sony-a7rii-with-56-megapixel-sensor/
I wonder if it will be formally announced before the 5Ds/5DsR are available for sale?

I do concede that in lab tests of the most extremes, the Sony / Nikon sensors come out on top by a very slight margin. But again and again, this comes up in real photography almost never.

If you want to compare the noise levels of Canon's sensor with those of Sony/Nikon/Samsung, read this thread:
http://www.canonrumors.com/forum/index.php?topic=26199.0

... all using the "latest" cameras from each manufacturer. Is the difference in noise a "slight margin"?

That is exposure latitude test in DPR where they will underexpose +5EV then push it later. Latest 70d, 7D2 and 6D are doing fine upto +3EV. There is no banding either. Actually Nikon d7100 is the one with the green horizon banding even though it has class leading DR and ISO performance. DPReview is keep on pushing d7100 over all the crop cameras including 7d2 with the assumption of great shadow pushing performance, DR and ISO. I think, that is the reason why they are avoiding adding d7100 to their exposure latitude test.

They kept on referencing 70d and 7d2 not so good performance when some on pushes unnecessarily underexposed shots by 5EV in each of their review. I am yet to understand their methodology where they are mixing up low iso DR, underexposing 5 stops then pushing later to preserve highlights, banding and pushing shadows by few stops. I am still trying to find out how to push shadows by x stops. Because DPP and dxo optics has shadow slider in terms of numbers/levels instead of stops. If you look at their real world example, it has lot of noise where they pushed shadows in a landscape. 70d, 7d2 can match upto +3EV push with any other crop cameras. Beyond +3EV push they are bad. Even though they gave up something in low ISO DR, 70d and 7d2 sensors have latest duel pixel tech which can do video AF where as Nikon still stuck with old CDAF for video AF.

Sony A7II you are refering in above links is not comparable to Nikon and Canon full frame sensors in terms of high ISO noise. It is barely beating d7100 and 7d2. Check these comparisons done by photographylife.
https://photographylife.com/reviews/sony-a7-ii/6

 

[rs]

The C300 mk.ii uses a sensor with similar QE/SN as the 6D. The difference would be the DPAF. Those extra pixels can be used to expose the shade areas for example; expose half the file at +4ev, and the non DPAF pixels expose at 0 EV; then the RAW file is cooked to spread this exposure latitude over one file (i.e Average out normal pixels and the DPAF ones shot at +4Ev. Very much like the ML Dual_iso hack. Except that this will be done at a hardware layer, meaning no loss of resolution like in Dual_iso.

DPAF is not used for dual ISO on the C300 mk II. It would create quite horrendous bokeh in bright and dark regions. The effect would have a very pronounced negative effect on the image at any viewing size (unless everything is in focus), unlike magic lanterns dual ISO which only has a minor impact when viewed at 100%.

The implementation used on the C300 mk II has the output from each pixel simultaneously fed into two amplifiers - one for the shadows/mid tones, and the other for mid tones/highlights. This way each segment of the picture sees the entire captured dynamic range, not split up into alternate pixels, and much more importantly, not split into alternate phases.

 

[LetTheRightLensIn]

I hope they use 16bit files with greater color depth as well... I struggle capturing deep reds in flowers with my 5D3, there is of loss of detail. Other channels are handled better. I do not recall my old 5Dc struggling with reds...

If you are not using a wide gamut monitor in wide gamut mode that can cause a lot of troubles with reds. Sure the color filter in the camera can make some difference, but the most critical difference is sRGB vs wider gamut color spaces.

 

[sanj]

...
Every time a new Canon is on the verge of release, there is speculation of a Sony sensor.
...

Yup and Zeiss appear to have mistakingly released a pic showing Sony with a 56MP sensor:
http://www.sonyalpharumors.com/sr4-did-zeiss-just-leak-the-sony-a7rii-with-56-megapixel-sensor/
I wonder if it will be formally announced before the 5Ds/5DsR are available for sale?

I do concede that in lab tests of the most extremes, the Sony / Nikon sensors come out on top by a very slight margin. But again and again, this comes up in real photography almost never.

If you want to compare the noise levels of Canon's sensor with those of Sony/Nikon/Samsung, read this thread:
http://www.canonrumors.com/forum/index.php?topic=26199.0

... all using the "latest" cameras from each manufacturer. Is the difference in noise a "slight margin"?

Do you have a similar comparison at ISO 800 and at ISO 2500. Please I want to see that.

 

[MrToes]

I'll take some more DR and less shadow noise!!!

 

[Hjalmarg1]

<p>Beyond the increased ISO capabilities of the EOS-1D X Mark II, we’re told that dynamic range is going to see a massive improvement, and possibly have the most stops of any DSLR currently on the market. It’s always possible that Sony will do something in the meantime with their new sensors to increase their lead in stops of DR.</p>
<p>It’s also sounding like a new DIGIC 7 processor will make its way into the EOS-1D X Mark II, instead of going with the DIGIC 6 processor that was introduced in 2013. The current EOS-1D X uses dual DIGIC 5+ processors.</p>
<p>More to come…</p>

I hope it comes true and better DR is also incorporated in all other Canon DSRL camera bodies.

 

[Aglet]

Now, Canon sensors are not bad at all. For being "behind" on specs and all the nerd-tests - they are either the equal of or the superior of other sensors on the market when it comes to final IQ result. Photos from Canon are great. Simple as that.

uhmmm... Lots of talented/lucky people make have made great photos using Canon gear, but that doesn't always equate to any form of technical superiority, it just reflects the state of the market numbers, which are, again, no metric equating to quality.

Whether that is due to other factors (processing, lenses etcetera) - that matters not to me. Because the camera is a SYSTEM. As a photographer, I only care about the photo I get out of the camera. If Canon is doing that with magic fairy pixel dust algorithms, or with a better sensor - I don't care. I get a RAW file that looks fantastic.

you really otta play with some ABC camera raw files to appreciate what the limitations are compared to canon's raw files shot for the same scene. ABC raw files will give you more processing latitude with less work.
I don't care so much about what comes out of the camera. I care about the quality of the raw file because some of the images I create may take a great deal of manipulation in post to produce the effect I'm after. A clean raw file is a good thing to have, otherwise it's a lot of work to fix the garbage noise from a poor sensor system.

No one shoots photos with just a sensor. The Sonikonians and DXO monkeys seem to conveniently forget that they aren't walking around with ONLY a CMOS. There's a camera body, image processing, electronics, lenses and much more that delivers that final IQ result.

Do you think that ABC cameras are lagging so far behind Canon, as a system, that they are handicapped in some way? While I'll concede that Canon has a good variety of niche lenses, it's again more likely due to their ability to fund that much R&D because of their size, not just engineering talent.

I do concede that in lab tests of the most extremes, the Sony / Nikon sensors come out on top by a very slight margin. But again and again, this comes up in real photography almost never. The gap between Sony/Nikon and Canon would have to be very significant to lead to practical IQ differences in most images.

As you may be aware, at the lower iso ranges, the difference is a modest 3 stops. having that much more clean data available provides a lot more choice over how to create the final image.

This difference is NOT seen within the same generation or even within the span of 2 generations in some instances. You would have to compare a 2-3 generation old camera of the same line/level to have noticeable differences in IQ without pixel peeping.

true, canon's IQ has not noticeably improved, model-to-model, for nearly a decade. With the only significant improvement in the very latest models finally reducing FPN by cleaning up the noise-inducing circuitry.
OTOH, SoNikon's improvements have been steady and incremental over the same period, adding up to today's 3 stop lead in some models.

 

[K-amps]

The C300 mk.ii uses a sensor with similar QE/SN as the 6D. The difference would be the DPAF. Those extra pixels can be used to expose the shade areas for example; expose half the file at +4ev, and the non DPAF pixels expose at 0 EV; then the RAW file is cooked to spread this exposure latitude over one file (i.e Average out normal pixels and the DPAF ones shot at +4Ev. Very much like the ML Dual_iso hack. Except that this will be done at a hardware layer, meaning no loss of resolution like in Dual_iso.

The implementation used on the C300 mk II has the output from each pixel simultaneously fed into two amplifiers - one for the shadows/mid tones, and the other for mid tones/highlights. This way each segment of the picture sees the entire captured dynamic range, not split up into alternate pixels, and much more importantly, not split into alternate phases.

Let me understand this better, The amplifier reading the shadows would need to amplify more since the pixel is not going to collect more photons beyond the shutter closing up. This would denote an operation similar to high iso (i.e. Signal amplification), so would that introduce noise?... or perhaps noisier than the mid to highlight capture which requires less amplification?

 

[Aglet]

ahem... so... how many stops of DR do you need to shoot a BIF anyway?...
Let's make it more relevant, a backlit black one?...
After all, you're properly exposing it, right?...

I'd like to see a Nikon sample of that if you have one. Depending on time of day it could be 18-20 stops of DR to get the bird and the sky. More DR would help in some circumstances.

Me too. But I don't have the fast-focusing long lenses for BiF shooting, it's not my area of interest to put that kind of loot into a big lens I don't have a use for.

 

[Lee Jay]

The C300 mk.ii uses a sensor with similar QE/SN as the 6D. The difference would be the DPAF. Those extra pixels can be used to expose the shade areas for example; expose half the file at +4ev, and the non DPAF pixels expose at 0 EV; then the RAW file is cooked to spread this exposure latitude over one file (i.e Average out normal pixels and the DPAF ones shot at +4Ev. Very much like the ML Dual_iso hack. Except that this will be done at a hardware layer, meaning no loss of resolution like in Dual_iso.

The implementation used on the C300 mk II has the output from each pixel simultaneously fed into two amplifiers - one for the shadows/mid tones, and the other for mid tones/highlights. This way each segment of the picture sees the entire captured dynamic range, not split up into alternate pixels, and much more importantly, not split into alternate phases.

Let me understand this better, The amplifier reading the shadows would need to amplify more since the pixel is not going to collect more photons beyond the shutter closing up. This would denote an operation similar to high iso (i.e. Signal amplification), so would that introduce noise?... or perhaps noisier than the mid to highlight capture which requires less amplification?

For a given number of photons captured you want to read out at the *highest* possible ISO, not the lowest. This will give you the lowest noise. The reason low ISOs have less noise is that more light is captured, not that the electronics are less noisy.

 

[Loswr]

But I don't have the fast-focusing long lenses for BiF shooting, it's not my area of interest to put that kind of loot into a big lens I don't have a use for.

You really should consider getting a supertele lens. Unlike smaller lenses where the lens cap merely sits on the front of the lens, supertele lens cap/covers fit completely over the front and extend down the sides like a glove, which ensures complete blockage of any light entering the lens. That could have a profoundly beneficial impact your lens cap photography.

 

[K-amps]

The C300 mk.ii uses a sensor with similar QE/SN as the 6D. The difference would be the DPAF. Those extra pixels can be used to expose the shade areas for example; expose half the file at +4ev, and the non DPAF pixels expose at 0 EV; then the RAW file is cooked to spread this exposure latitude over one file (i.e Average out normal pixels and the DPAF ones shot at +4Ev. Very much like the ML Dual_iso hack. Except that this will be done at a hardware layer, meaning no loss of resolution like in Dual_iso.

The implementation used on the C300 mk II has the output from each pixel simultaneously fed into two amplifiers - one for the shadows/mid tones, and the other for mid tones/highlights. This way each segment of the picture sees the entire captured dynamic range, not split up into alternate pixels, and much more importantly, not split into alternate phases.

Let me understand this better, The amplifier reading the shadows would need to amplify more since the pixel is not going to collect more photons beyond the shutter closing up. This would denote an operation similar to high iso (i.e. Signal amplification), so would that introduce noise?... or perhaps noisier than the mid to highlight capture which requires less amplification?

For a given number of photons captured you want to read out at the *highest* possible ISO, not the lowest. This will give you the lowest noise. The reason low ISOs have less noise is that more light is captured, not that the electronics are less noisy.

Thanks, I had it backwards. I thought hi-ISO meant high amplification, thus higher noise....