The Canon EOS 5D Mark V is in the works [CR2]

Status
Not open for further replies.

Michael Clark

Now we see through a glass, darkly...
Apr 5, 2016
4,722
2,655
I'll never understand the constant carping by some. They should just get what they want and have at it. Sitting around and bitching that a Canon is not the Sony they want is mentally deranged. Some of then do it year after year after year. Stupid. They accuse us of being fanboys, yet they are the one's sticking around and complaining without switching... though they talk about switching for years. %$#&@$%^ leave. I have been here for several years now. It is always the same idiots or new accounts.

I highly suspect most of them do not own a FF or advanced APS-C camera in either brand. They've either got a Rebel or an α6xxx with a (few) cheap kit lens(es) or they do not have an ILC at all. They're arguing about which brand of camera they'll never be able to afford they want to drool after.
 
  • Like
Reactions: 1 users
Upvote 0

Ozarker

Love, joy, and peace to all of good will.
CR Pro
Jan 28, 2015
5,933
4,336
The Ozarks
I highly suspect most of them do not own a FF or advanced APS-C camera in either brand. They've either got a Rebel or an α6xxx with a (few) cheap kit lens(es) or they do not have an ILC at all. They're arguing about which brand of camera they'll never be able to afford they want to drool after.
What is weird is that I am very low income, but make a lot of sacrifices for what I have. One would think these twits could do better. I live on $1,600 a month. It isn't how much money one has. It is how one manages it.
 
  • Like
Reactions: 1 user
Upvote 0
No, they shouldn't. You are losing all the higher-frequency components of the signal (the Nyquist-Shannon theorem). Also, from the whole sensor exposed to the same "2-pixel" scene, you can recover lower signal levels for your "2-pixel" frequency (so, basically, a constant to half cycle signal) if the noise is the photon shot noise (or another kind of additive white noise).

Even a single pixel has its own dynamic range and it's not 1/30000000 of the whole sensor's DR.

That's the whole point. The DR is a measure of the amount of information you can get about the original scene.

DR is related to the information but it's defined differently. It is not the amount of information. It's the difference between the darkest and brightest parts the sensor can capture. It's a contributing factor but definitely not the amount of information.

We measure DR on a kind of image we consider as a "signal". We are typically interested in the information that has lower spatial frequencies than the Nyquist limit - otherwise we would be using the wrong instrument to the task.

No, it's the other way around: the parameters of the lens we use to measure the sensor's DR are chosen to be up to the task.

In Photostophotos method, they also use some arbitrary constants like size of the print seen a certain distance. So changing the distance or print size changes the DR. If we make the prints bigger, the DR will drop. That sounds odd doesn't it?
 
Upvote 0

SecureGSM

2 x 5D IV
Feb 26, 2017
2,360
1,231
Even a single pixel has its own dynamic range and it's not 1/30000000 of the whole sensor's DR.



DR is related to the information but it's defined differently. It is not the amount of information. It's the difference between the darkest and brightest parts the sensor can capture. It's a contributing factor but definitely not the amount of information.



In Photostophotos method, they also use some arbitrary constants like size of the print seen a certain distance. So changing the distance or print size changes the DR. If we make the prints bigger, the DR will drop. That sounds odd doesn't it?
Here is where confusion originated:

They are talking about a Dynamic Range of an image and then all their methods totally make sense and you are on about sensor / pixel level dynamic range. Obviously not the same thing. Agree?
 
Upvote 0
On the shadow side DR is bound by noise. Noise is not perfectly uniform. A quad of pixels occupying the same space as a single larger pixel are more likely to detect something of value near their noise limits. If you increase the number of samples (pixels) that contribute to a given view size, you are in fact increasing the signal against the noise and thereby improving DR. When you crop a higher MP image to a lower MP one, you're throwing away signal and decreasing DR. As you crop further you bring DR closer and closer to the DR of a single pixel.

Yep I get that concept. The only little issue is, you change the DR by changing the amount of downsampling, or, in terms of Photonstophotos method, by changing the viewing distance and/or the print size. So as I said above, their 'photographic' DR may be usable for relative comparisons (e.g. yay Sony is better than Canon), but absolute values of measured DR are meaningless.
 
Upvote 0
Here is where confusion originated:

They are talking about a Dynamic Range of an image and then all their methods totally make sense and you are on about sensor / pixel level dynamic range. Obviously not the same thing. Agree?

Yes, they're different, I agree. My point is as in my message above, "their 'photographic' DR may be usable for relative comparisons (e.g. yay Sony is better than Canon), but absolute values of measured DR are meaningless."
 
Upvote 0
You mean the workflow used by the majority of landscape photographers over the past 4-5 decades?

Depends on the equipment. With the large format film cameras, they'd use the rear screen, pretty close to the Live View, only upside down.

Dusk/dawn aren't all that tricky exposure wise. It's not one of the situations where I consider exposure preview to be a significant advantage. I have no issue composing or visualizing dusk/dawn landscapes through an OVF. If I use LV it's typically because I'm on a tripod and the height makes LV more convenient.

OVF is great for action shooting, LV for landscapes. Obviously it's not impossible to use OVF for landscape photography, but LV is so much better. To me it does provide a significant advantage. Especially with a tripod, as you said. 99% of my landscape shots are taken from a tripod. Very rarely I'd shoot a landscape handheld, in such cases I might use the OVF.

My 5Ds has OVF level indicators, did Canon drop that from the 5D IV?

No they didn't but in-OVF level indicator seems to be less accurate, also not visible in the dark. Also OVF on 5DIV only covers 97-98% of the frame and.

LV will show the very brightest stars, but you can see and compose the Milky Way in an OVF with a fast prime. (Assuming you're at a site where the Milky Way is not light polluted away to begin with.)

I can set focus on a bright star in the LV (and I won't be able to focus on the stars reliably in the OVF). Mlky Way itself can be positioned in the frame using the OVF, but if you care about your foreground, you'll see nothing through the OVF. What I normally do is shooting samples at a very high ISO so I can see the composition and the foreground, then adjust the camera iteratively until the satisfactory result. In this case, the OVF can be used for initial rough composition.
 
Upvote 0

Joules

doom
CR Pro
Jul 16, 2017
1,801
2,247
Hamburg, Germany
It's the difference between the darkest and brightest parts the sensor can capture.
How do you define the darkest part your sensor can capture? Is it true black (pixel value of 0 for each channel)? Would that not lead to every sensor having almost exactly 14 stops of DR?

My understanding of Photons to Photos definition of this is the part where the noise transitions from being not to distracting too distracting. How visible the noise is is clearly related to your viewing distance / enlargement. So he adjusts the measured SNR to take this into account. And then the point where the adjusted SNR drops below his threshold is what he defines as his "darkest part".

I don’t understand how your concept of this part works. Could you please elaborate or give some site that uses that way of measuring?
 
Upvote 0
I normally look at images and not at individual pixels. Anything smaller than pixel peeping (100%, 1:1) is going to represent an increase in DR over the strict DR of a single pixel measurement.

But again what does the absolute DR value from Photostophotos tell you? Say for 5DSr it's 9.8 at ISO 100. And you're putting an image on Instagram. But that DR is calculated for a 8"x10" print.
 
Upvote 0

Joules

doom
CR Pro
Jul 16, 2017
1,801
2,247
Hamburg, Germany
And you're putting an image on Instagram. But that DR is calculated for a 8"x10" print.
The dynamic range we are talking about is the maximum that you could pull out of a given raw file. Measuring doesn't tell you how much DR is in your output or how much of it can be displayed on the device / medium used to view it. I personally don't care about that either, when I look at measurements like this I want to know what my gear is capable of. If I or the people viewing my images can make use of that is a seperate thing and probably much harder to measure.
 
Upvote 0
OVF is great for action shooting, LV for landscapes. Obviously it's not impossible to use OVF for landscape photography, but LV is so much better. To me it does provide a significant advantage. Especially with a tripod, as you said. 99% of my landscape shots are taken from a tripod. Very rarely I'd shoot a landscape handheld, in such cases I might use the OVF.

For me, I like to eye ball landscape compositions through the OVF. I might do that off and on for hours or even days whilst scouting with no battery drain and not a single shot fired. Then if the light actually plays ball, I'll set up the tripod and engage LV to take the shots, and retake until the best light is gone with all the benefits of live histogram and contrast AF.

The problem with pre-dawn is that whilst I can see the scene with my eyes, and through the OVF, the LV is totally black or just a silhouette even at 30 sec F2.8. Right now, using OVF I can compose in the near dark before dawn, focus, and be all set on the tripod ready to simply adjust exposure and hit the shutter if it's worth it in a few minutes time. That's the method I suspect would be an issue with a mirrorless system, and one reason a 5DV DSLR would be useful at least to me.
 
  • Like
Reactions: 1 users
Upvote 0
How do you define the darkest part your sensor can capture? Is it true black (pixel value of 0 for each channel)? Would that not lead to every sensor having almost exactly 14 stops of DR?

The sensor's DR is not about absolute brightness values but about the difference. Roughly speaking, if a scene has 15 stops between the darkest and brightest parts, and your camera only has 12 stops of DR, you can adjust your exposure and capture either the first darkest 12 stops, or the last brightest 12 stops, or in the middle. Beyond those 12 stops, your sensor doesn't capture any detail.

My understanding of Photons to Photos definition of this is the part where the noise transitions from being not to distracting too distracting. How visible the noise is is clearly related to your viewing distance / enlargement. So he adjusts the measured SNR to take this into account. And then the point where the adjusted SNR drops below his threshold is what he defines as his "darkest part".

I don’t understand how your concept of this part works. Could you please elaborate or give some site that uses that way of measuring?
I don't have a concept, it's their concept. My critique is that their measurements can be used for relative comparison (again as above, Sony vs Canon) but the absolute DR values on that site are meaningless as they're only applicable to specific downsampling ratio.
 
Upvote 0
The dynamic range we are talking about is the maximum that you could pull out of a given raw file. Measuring doesn't tell you how much DR is in your output or how much of it can be displayed on the device / medium used to view it. I personally don't care about that either, when I look at measurements like this I want to know what my gear is capable of. If I or the people viewing my images can make use of that is a seperate thing and probably much harder to measure.

No, we were talking about specific DR measurement method from http://www.photonstophotos.net/Charts/PDR.htm
Those charts don't tell you the maximum you could pull out of your raw files. The charts are based on a model where you print the image on 8"x10" paper and view it from a certain distance (I understand it's just the math model - they don't actually print the images).
 
Upvote 0
Partial Spec List based on logic & experience:

Shutter lag will be so slightly improved that it won't make any difference.
Expect 40MP and an S or R version to be about 100 MP after a year or so..
1/200ths of a sec, top sync with dedicated strobes.. All other lights, like studio lighting will be 1/125ths, tops!
Expect no swivel screen.
7.7fps (since the 5D2 was 3.9fps, they couldn't even do a solid 4fps, so sorry, no 8fps... it will fall short on that too, just to annoy you!)
The AF system will remain the same layout, with a nice small little area of AF coverage. If you want more AF coverage, then buy a nifty new 90D, (which is not a combo of the 80D and 7D2 line... just a minor update to the 80D)..
No brightly illuminated AF point, so you'll have to dink around with trying to locate which AF sensor is selected, while you miss shots.
If you want a better performing camera, then buy a 1DX3, for more than 2x as much!
Expect a short burst of RAW photos...maybe 2.5 secs, tops!
It will use the same batteries as always... nothing new there, so at least all your old batteries will still work..
Expect a new flash for 900.00 with such slightly better performance, you'll only buy one if you find a good deal on a used one, from someone who gave up on Canon and switched to either Sony or Nikon... who have capable 2nd tier cameras, that totally out perform anything Canon has to offer...
 
Upvote 0
For me, I like to eye ball landscape compositions through the OVF. I might do that off and on for hours or even days whilst scouting with no battery drain and not a single shot fired. Then if the light actually plays ball, I'll set up the tripod and engage LV to take the shots, and retake until the best light is gone with all the benefits of live histogram and contrast AF.

It's a good technique but as you said, you're doing the actual shooting with the LV.

The problem with pre-dawn is that whilst I can see the scene with my eyes, and through the OVF, the LV is totally black or just a silhouette even at 30 sec F2.8.

Agree, there are cases when the LV fails and the OVF works, but they're quite rare. In manual, I'd just increase ISO until the preview is bright enough, so I can compose. However it still fails in very dark conditions. Disabling exposure simulation may help although it also disables the histogram.

Right now, using OVF I can compose in the near dark before dawn, focus, and be all set on the tripod ready to simply adjust exposure and hit the shutter if it's worth it in a few minutes time. That's the method I suspect would be an issue with a mirrorless system, and one reason a 5DV DSLR would be useful at least to me.

I'd rather have a flip screen. Also there are reviews on the EVF in EOS R and they say it works very good at night.
 
Upvote 0
Status
Not open for further replies.