7D Mark II Sensor Noise & banding comparison to 7D, subjective first look

[Marsu42]

As it is, with the lenses I want to use, it's no better at AF than my 60D.

I don't know about your lenses or shooting circumstances, but for me lifting the +0.5lv af limit of the 60d (that's not very dark esp. with f4+ lenses) is a merit on its own.

...

...

 

[privatebydesign]

You make the simple assumption that there can't be any issues with the 7D II AF. What if there is? What if it is a firmware issue that needs to be addressed? You jump right out and accuse the people who are actually using the camera and noticing things about it's performance that they are crazy. You make the assumption that they are misusing the AF system. If there is a problem, how else is it going to be found, reported to Canon, and dealt with unless people test, push the system to limits, and talk about it?


None of us are going around assuming everyone else is an idiot, and proclaiming such. Guys like you, Keith, and a good number of others? Your constantly going around telling everyone who doesn't agree with you, the seasoned craftsman, is crazy for caring about details. What is it going to take to get you guys to just leave us the hell alone? I mean, do you even get it, that we want to be left the friggin hell alone? Or do you prefer to play the role of antagonist, constantly drudging things up and making an issue out of it?

You really do like the sound of your own voice don't you.

Anyway, I made no assumptions and am well aware that there could be a genuine issue with any new release, or a problem with a specific manufacturing batch, indeed nowadays it seems unlikely that there won't be some kind of issue, contrary to your belief I do not live in denial. But that doesn't mean there isn't a little craziness around some of the time!

I don't assume everybody is an idiot, I know some are, and I know many are way smarter than me (which isn't that difficult), there are many threads where I suspect i could call bullshit but don't because I am just not knowledgeable enough in a specific area to comment.

I care about details, but not in and of themselves, I know the answer to where the magic bullet is, it is inside each of us, it is not in one more stop of DR, better AF or anywhere else connected to the gear; I am not suggesting for a second gear doesn't matter, of course it does, and in some shooting situations it is paramount, but even the best gear in the hands of somebody without the time and knowledge to use it won't deliver. And that is what I feel many of the rants are about, the unrealistic expectation and the unwillingness to commit the time needed to master a hobby, any hobby. I know that if I get the same rod and reel, or gun as a fisherman or hunter I am not going to get the same results, it isn't about the type of bait or cartridge I use, it is because I have spent fifteen minutes fishing in my entire life, I don't feel the fish or the deer.

I am not surprised you shoot thousands of images birding, but I don't understand why you think you should get tens of images to "work", 99.9% of them are not worth the time or effort, those that are, are worth the time and effort! Unrealistic expectations.

When will I leave you alone? When you stop making what I see as outlandish or farcical claims that you either can't illustrate, make no difference to the actual images people shoot, or when you stop mistaking example Canon images for Exmor ones! I am not out to antagonise you, indeed you have sought out this contretemps, I have said I see myself a bit of a balance to some of the sillier claims (8"x10" prints) and to add a little to a few other threads.

 

[Loswr]

I know that if I get the same rod and reel, or gun as a fisherman or hunter I am not going to get the same results, it isn't about the type of bait or cartridge I use, it is because I have spent fifteen minutes fishing in my entire life, I don't feel the fish or the deer.

The real question is would you go to a fishing or hunting forum and blame your lack of success on the harsh color banding on the lure or the inadequate range of the gun? :-X

 

[jrista]

You make the simple assumption that there can't be any issues with the 7D II AF. What if there is? What if it is a firmware issue that needs to be addressed? You jump right out and accuse the people who are actually using the camera and noticing things about it's performance that they are crazy. You make the assumption that they are misusing the AF system. If there is a problem, how else is it going to be found, reported to Canon, and dealt with unless people test, push the system to limits, and talk about it?


None of us are going around assuming everyone else is an idiot, and proclaiming such. Guys like you, Keith, and a good number of others? Your constantly going around telling everyone who doesn't agree with you, the seasoned craftsman, is crazy for caring about details. What is it going to take to get you guys to just leave us the hell alone? I mean, do you even get it, that we want to be left the friggin hell alone? Or do you prefer to play the role of antagonist, constantly drudging things up and making an issue out of it?

You really do like the sound of your own voice don't you.

No more than you, apparently. :

I have said I see myself a bit of a balance to some of the sillier claims (8"x10" prints) and to add a little to a few other threads.

I clarified this at the time, but apparently only fragments of my posts are ever read. When I made the comment about 8x10" print, I clearly stated I was referring to "for viewing on the web." I was talking about an image downsampled to approximately 8x10" size ON SCREEN. I also clarified that when I print at home, I usually print at 13x19" which for the 5D III is just about native size. This is my issue...you guys pick posts apart so thoroughly you lose the context.


Well, whatever.


You keep making the assumption that I am not interested in investing the time to master something. The whole concept about Outliers is pretty clear that it is 10,000 hours of dedicated practice, with express intent and effort to always learn something new every hour of time you invest in whatever it is you are trying to master. Repetitive, mundane "task" work is mastered rather quickly, and once mastered, it is no longer going to contribute to your growth. You have to actually find an area you are weak at, say focusing, and work it with the express intent of LEARNING more about focusing. If you are weak at tracking, you have to work it with the express intent of LEARNING more about tracking. That is what it takes to become an outlier. That 10,000 hours isn't just total mundane hours spent...it is 10,000 hours of dedicated, explicit practice that intentionally pushes your boundaries and forces you to constantly break into new territory. To grow as an artist, for example, you have to actually be an artist. Removing banding and blotch, extra steps to denoise, or running an HDR merge...that isn't art, it's just the busywork that you have to perform to get up to the point of then being an artist, of working the photo to produce an actual work of art. It's useless extra.


I'm not afraid to work and spend the time to master something. I've spent the last eight months doing exactly that with my astrophotography, and I have some great examples to demonstrate real-world growth in that area, as an astrophotographer, and an artistic one. I don't like all the extra busywork. DSLR's increase the busywork I have for astro...moving to mono CCD will eliminate a LOT of that busywork. Depending on exactly what kind of astrophotography I want to do, some cameras would be better than others...it's a careful choice, no different than with regular photography. Canon's read noise is often a source of busywork I have for regular photography (namely landscapes)...moving to a camera that has lower read noise would eliminate much of that busywork.

 

[Sporgon]

When will I leave you alone? When you stop making what I see as outlandish or farcical claims that you either can't illustrate, make no difference to the actual images people shoot, or when you stop mistaking example Canon images for Exmor ones! I am not out to antagonise you, indeed you have sought out this contretemps, I have said I see myself a bit of a balance to some of the sillier claims (8"x10" prints) and to add a little to a few other threads.

You forgot to mention the bit about citing (blatantly) multiple exposure D800 pictures exhibited on 500px, as single frame exposures.........

Now that jrista has actually used an Exmor based camera in the field he must realise that he was wrong on this front, but hasn't admitted it........yet.

 

[jrista]

When will I leave you alone? When you stop making what I see as outlandish or farcical claims that you either can't illustrate, make no difference to the actual images people shoot, or when you stop mistaking example Canon images for Exmor ones! I am not out to antagonise you, indeed you have sought out this contretemps, I have said I see myself a bit of a balance to some of the sillier claims (8"x10" prints) and to add a little to a few other threads.

You forgot to mention the bit about citing (blatantly) multiple exposure D800 pictures exhibited on 500px, as single frame exposures.........

Now that jrista has actually used an Exmor based camera in the field he must realise that he was wrong on this front, but hasn't admitted it........yet.

I admitted that one or two might have been multiple exposures based on an author comment or something like that. As for the rest, sorry, but I don't believe they are "proven" multiple exposures. They do look tonemapped to me...but that's what you do with tons of dynamic range. It doesn't matter if it's 14 or 20 stops worth, you still have to tonemap it into the limited DR you can see on screen or in a print. I haven't admitted anything because I don't believe most of those are anything other than single, tonemapped shots...even after having used an Exmor based camera.

 

[Loswr]

You forgot to mention the bit about citing (blatantly) multiple exposure D800 pictures exhibited on 500px, as single frame exposures.........

When one is convinced of the superiority of Exmor, one sees that superiority.......everywhere. Even in Canon images! :

 

[Valvebounce]

Hi Folks.
It amazes and saddens me at how fast two intelligent beings can descend the spiral to the bottom. It appears from outside this "argument" that neither is listening to the other, each has misunderstood part or all of a post and both like your own voice!
By the way this from someone who respects both of you as valuable contributors to this community. Perhaps we could draw a line under this and move back to the original topic?
-------------------------------------------------------------------------------------------------------------------------------------

Cheers, Graham.

 

[Aglet]

I don't know about your lenses or shooting circumstances, but for me lifting the +0.5lv af limit of the 60d (that's not very dark esp. with f4+ lenses) is a merit on its own.

FWIW, I measured +0.1 EV with my Sekonic 558 on an indoor target i was using for some tests.
with the 100-400 L mounted and stroked out to the 400mm end, the 60D and the 7d2 were both able to AF, and oddly, the 60D did it subjectively faster every time.
I think there may be some minor issues the 7d2 needs to have addressed with some firmware tweaks.
I'm pretty sure 7d2 would deliver better IQ than the good old, non-ML-equipped, 60D.

 

[jrista]

I don't know about your lenses or shooting circumstances, but for me lifting the +0.5lv af limit of the 60d (that's not very dark esp. with f4+ lenses) is a merit on its own.

FWIW, I measured +0.1 EV with my Sekonic 558 on an indoor target i was using for some tests.
with the 100-400 L mounted and stroked out to the 400mm end, the 60D and the 7d2 were both able to AF, and oddly, the 60D did it subjectively faster every time.
I think there may be some minor issues the 7d2 needs to have addressed with some firmware tweaks.
I'm pretty sure 7d2 would deliver better IQ than the good old, non-ML-equipped, 60D.

Well, is it possible the 7D II has firmware that is explicitly designed to throttle AF speed on lenses that need it to happen more slowly? The old 100-400 has it's issues...mine seemed to hunt more than I ever liked. Maybe they are forcing a slower AF speed now to limit hunting. Have you tried other lenses?


Also, is the 7D II consistently accurate and precise? I mean, is the issue just that it's slow, or is it actually misfocusing? The thing I'm most interested in is whether Canon's new 65pt AF system inherited the 19pt AF system's inherent "jitter"...in that, frame to frame, it would always adjust focus ever so slightly, resulting in some frames just being enough OOF that you could tell...then popping back into perfect focus the next frame.

 

[risc32]

I'm not totally against this lens cap photo comparison, but i'd much prefer photos of a real life scene. i just don't trust that the camera bodies are being honest with what they are showing me. in regards to what the cameras might be doing, in the words of han solo, "i can imagine quite a bit."
now back to "bidness" as usual (insert DR range comments)

what? 2 replies while i was typing, if they are on topic i'm drinking a second beer! or should that be the other way around?
- this just in, CR forum drinking game, so everytime someone types something about DR we all take a shot of jager!

 

[Aglet]

Here's another series of screenshots from Iridient Developer, v3 beta.

These are 25% linear scale of the black-cap-shots which give an impression of what the raw files are like when pushed even harder so you can better see the residual horizontal and vertical banding structures.

Files are saved as jpg with 4-4-4 subsampling to better preserve the speckle appearance which 4-2-2 otherwise smears. hopefully that doesn't get messed up when uploaded here.

These are pushed FIVE stops.
Then another has +100 Fill Light added
Deep Shadow Fine Tune has NO tint correction applied

done for both 100 and 1600 Iso

one 1600 iso is pushed 5 stops, no fill light, and then Deep Shadow Tint-Correction is set to low. Notice how it does a nice job of removed most of the red noise speckle. No idea how well this would translate to shadow detail loss, however. Perhaps someone can try that and post the results.

The file names are 7d2_iso-EV-fillLight-444.jpg

EDIT - just checked full size image recovered from clicking on the sample here and it's pretty much what it should look like.

 

[Marsu42]

Files are saved as jpg with 4-4-4 subsampling to better preserve the speckle appearance which 4-2-2 otherwise smears. hopefully that doesn't get messed up when uploaded here.

This site converts in-thread previews from jpeg to png, so to preserve the best quality simply upload as png right away. You can still link the original full-res jpeg files from an imagehoster in addition to that. To save filesize or for detail comparisons use compressed tiff or (imho best choice) lossless jpeg2.

 

[bgosselin]

For the expert on that subject. Where is that noise coming from? Is it completly random or it's pixel dependent? What I mean. If you shoot with your caps on twice. With the same exposure time and same iso. Will both raw look exactly the same when push 4 stops?
If it does. Would it be possible to manipulate the raw file so you correct the exposition per pixel? I know that is a lot of pixel to be corrected. But with computer power that must be something that can be done.

I seen on my mark iii an option that reduce noise for long exposure. What does canon do then? They just filter out dim light in long exposure?

Feel free to send me to articles on this subject that could make me understand it better.

 

[FEBS]

For the expert on that subject. Where is that noise coming from? Is it completly random or it's pixel dependent? What I mean. If you shoot with your caps on twice. With the same exposure time and same iso. Will both raw look exactly the same when push 4 stops?
If it does. Would it be possible to manipulate the raw file so you correct the exposition per pixel? I know that is a lot of pixel to be corrected. But with computer power that must be something that can be done.

I seen on my mark iii an option that reduce noise for long exposure. What does canon do then? They just filter out dim light in long exposure?

Feel free to send me to articles on this subject that could make me understand it better.

Here a link to CPN:

http://cpn.canon-europe.com/content/education/infobank/exposure_settings/digital_noise.do

 

[Lawliet]

Is it completly random or it's pixel dependent? What I mean. If you shoot with your caps on twice. With the same exposure time and same iso. Will both raw look exactly the same when push 4 stops?

Neither - It's not fixed, the results change from shot to shot. It's also not random, the readout electronics Canon uses tend to introduce patterns into the noise. You have areas with higher noise next to patches with a cleaner signal.
Now fixed pixels could easily suppressed, and completely random noise can be cleaned up by statistical means or simply left in as it has a mostly local effect.
The patterns OTOH are bad because they remain perceivable even with aggressive noise reduction or image size reduction and introduce large scale artifacts.