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

[masterpix]

Unless the 5DV has a removable mirror box and was then both EF and RF compatible.

remember this?

Canon is Developing a Mirrorless-to-DSLR 'Converter': Report

One of Japan's largest newspapers has published a report that contains two very interesting tidbits for Canon mirrorless shooters. First, they claim that

petapixel.com

For some reason, I don't see that going to happen. The main point is that in order to do so, you need to stick the mirror as back as possible, not touch the sensor, yet lock it firmly in place. So you need to make a non rotating unit that will lock into the camera bionet while rotating and you need to do all that in a very limited space cause if you look into the DSLR, the mirror is almost the side of the bionet.

 

[Deleted member 378085]

I find it funny the way Friends of Sony (DPR) wait until the newest model to finally reveal their true reservations towards the outgoing model to pump it up. Sony is fine but I need a 50mm F1.2 and 85mm F1.2.

agreed, the funniest part, for someone who loves their Sony so much seems to have to come to a canon forum and tell themselves that. Pretty sad to be honest.

 

[Deleted member 378085]

Exactly. Canon remains the clear and overwhelming choice of most professionals.

I’ll never get it.

There are a handful of brand names today that I, personally, perceive as being absolute junk. It’s just MY opinion. Hyundai and KIA cars, Payless shoes, Huawei phones, Sony electronics (all of them). All strike me as being cheap schlock that I have no interest in buying. I’m secure in my firm belief that every car, shoe, phone, stereo, television or camera I own is superior in everyday way to anything from any of these brands.

And because I’m secure about my decisions I don’t feel the need to run to their brand pages or advertising posts or forums and crow about how much better my stuff is. I could care less about their junk.

As the song says: Real gangtahs don’t flex nuts.

Cause...

Real gangstahs know they got ‘em.

HAHA, well said and agreed! Sony seems to be going after numbers, and the non photographers are falling for it, coming on here to tell us how great their cameras are, how much DR and MP they have, if its so great then why are they hear telling us this crap?

I've played with a Sony, the menus are horrid, the feel is cheap, the agronomics are horrible. Its just feels and looks cheap.

They are so badly desperate for numbers that I looked into how the A9II can shoot 20 FPS, and really it can only shoot 12 FPS in RAW format, to get to 20 FPS it shoots in "compressed RAW", like WTF! Who the hell spends that kind of money to get a compressed RAW file? The 4 year old 1DXii still shoots faster in uncompressed RAW.

This is what I'm talking about, its a numbers game for them

I could care less if it has one more stop of DR or more MP, that in the end means the least to me. I know my 5d4 feels good in the hands, has easy menu system. Is robust and durable. Those are the most important factors for me.
In the end a capable photographer can get any shot they want with this camera or the Sony, its not the camera that is the limitation, Sony crowd think they are better just by a numbers war. This is a horrible misconception and beleif to have.

 

[AlanF]

These constant battles of Canon vs Nikon vs Sony are just so tedious. They are only cameras, tools for a job. If people can get so worked up over such trivia, no wonder different groups are killing each other over race, religion, territory and politics.

 

[Proscribo]

They are so badly desperate for numbers that I looked into how the A9II can shoot 20 FPS, and really it can only shoot 12 FPS in RAW format, to get to 20 FPS it shoots in "compressed RAW", like WTF! Who the hell spends that kind of money to get a compressed RAW file? The 4 year old 1DXii still shoots faster in uncompressed RAW.

Now now... no sane person shoots uncompressed when there are lossless compressed options (most manufacturers use lossless compressed raw formats, Canon included).

It's that Sony used to shoot only lossy compressed raws... then they included "uncompressed" because that lossy format had some drawbacks. Aand they still have only those two as options.

 

[Michael Clark]

Perhaps because they have only recently gotten seriously into it with the RF mount.
Let us see how those figures shake out in 4-5 years.

The point is, even without a FF mirrorless, Canon was already selling more total mirrorless interchangeable lens cameras than anyone else.
Sony's claim that they sold more than Canon is only true within a miniscule part of the total market: FF MILCs.

 

[dtaylor]

These constant battles of Canon vs Nikon vs Sony are just so tedious. They are only cameras, tools for a job. If people can get so worked up over such trivia, no wonder different groups are killing each other over race, religion, territory and politics.

We'll know things are bad when Congress demands troops at Nikon or a regime change at Canon.

 

[scottburgess]

It is also possible that Canon may try to further subdivide the market by offering more models, rather than less. I'm not suggesting that this will happen, I'm simply suggesting that we don't know if it would yield more profit to have several variations of basic DSLR and Mirrorless models. Certainly, Canon and Nikon both followed that philosophy with their APS-C DSLRs. In that case, it was driven by price point. It's possible they could follow a similar strategy based on features -- megapixels, video functionality, autofocus system, etc.

Canon has been employing this common strategy for years, as has been discussed in the Forums previously. If you look at the EOS Timeline, you'll see the steady growth in DSLR bodies offered by Canon.

I doubt the development costs for a DSLR are a large fraction of the price: if they were, by now most of the small manufacturers would be out of business. Most of the cost is the sensor, with the processor likely taking second or third place.

 

[analoggrotto]

agreed, the funniest part, for someone who loves their Sony so much seems to have to come to a canon forum and tell themselves that. Pretty sad to be honest.

I've often hypothesized that they are all from some sort place that never fully recovered and moved on from the cold war.

 

[MadScotsman]

I've often hypothesized that they are all from some sort place that never fully recovered and moved on from the cold war.

Someone else on here put forth the hypothesis that they don't actually own a camera.

I made a snide comment about the Sony customer culture, and their reply was along the lines of "they'd have to actually buy the camera to qualify as a customer". Which struck me as very true. If you actually owned the camera you thought was so much better, why would you be wasting your time trolling?

You wouldn't, ergo, you don't.

 

[slclick]

I've shared and compared various past time and professional forum behavior with friends who frequent other sites and it is clear that among boats, cycling, fishing, photo, automobile, running, music, musical instrument, gun, political and religious that the most mean spirited and close minded are the gun and politic forums (that should go without saying) but in a close 2nd place is cars. Canon vs Nikon is nothing compared to Ford trucks vs Chevy trucks.

From my personal history, the road cycling and drum forums I have been on are far more uplifting, positive, helpful and understanding than most camera sites. Still, CR is in itself the most friendly of all the camera forums and I have dabbled in others always to leave and never come back. Ignorance (usually chest thumping over products they never used or owned) and downright derisive opinions of anyone with film history or true experience and knowledge of the industry just burns me up.

Some folks love to burst into a hobby, spend a ton of cash and instantly prop themselves up in their own minds as if the credit card bill gives them cred. The internet fuels this persona. Seriously...your bad images, failures in the field, silly mistakes and poor gear choices make you a better photographer, we all go through it, some of us every single time. Believing that the gear can propel you from 1st time noob to pro shooter is a pathetic lie many tell themselves and is glaringly obvious. I am very glad that is much more of a thing over at DPR and not so much here. There are quite a few amazing and learned participants on this site and I am glad to have their experience to grow from. (Just don't talk about sensor science all day long for 16 pages, ugh just shoot me)

 

[analoggrotto]

Someone else on here put forth the hypothesis that they don't actually own a camera.

I made a snide comment about the Sony customer culture, and their reply was along the lines of "they'd have to actually buy the camera to qualify as a customer". Which struck me as very true. If you actually owned the camera you thought was so much better, why would you be wasting your time trolling?

You wouldn't, ergo, you don't.

There's another good point, who is really buying what and what are they using it for.

I throughly and almost bitterly regret not getting the 5D4 the day it was released and selling my 5D3 as it's replacement was arriving, but I let the DPR goofuses convince me that it wasnt a major step up, but it is. And its still not that far behind Sony's finest over at PhotonstoPixels.

Here's the killer : the A7 was released about a year after the 5D3 (2012) and that camera had poor low light AF, poor DR and and almost nonexistent tracking AF. Six years later, barely 1 year after Canon's Full Frame Mirrorless and we are only 1 notch behind in those categories and we have F1.2 Lenses while the 90D's sensor wins praise from Sony's-R-Us. Canon is just getting started.

Canon is that woman that doesnt care about makeup, wears whatever she found on the floor but is still a total endurance freak in bed.

 

[Quarkcharmed]

All the papers I've read so far on digital sensors point out they're linear. A single pixel can have more than 14 stops of well capacity, but it's then clamped and converted lineary in ADC. So the number of bits in ADC is the upper bound for the resulting dynamic range.

Because there is no 1:1 correspondence between the bit depth of a digital sensor and the maximum DR it can do. It's a fallacy that has been around since the dawn of the digital age.

A "digital"sensor is really a collection of a bunch of analog charge collectors. When the analog charges accumulated by each detector are converted to digital information, how much distance is one "step" (that is, how much more charge is needed to increase a digital value by "one") is strictly arbitrary. The difference between the noise floor and full well capacity can be divided between as little as one bit (it's either "off" or "on" at some arbitrary charge level) or as many bits as one desires. Sure, the bit depth can limit the number of expressible steps between a minimum and maximum digitally recorded value, but there's nothing that says the difference between 0 and 16,383 that requires 14-bits has to be exactly 14 stops. We can choose to assign "0" to an analog voltage value that is half full well capacity and "16,383" to an analog value that is three-quarters of FWC. Anything less than half of FWC is recorded as "0" during ADC, and anything more than 75% of FWC is recorded as 16383 during ADC.

 

[Joules]

When the analog charges accumulated by each detector are converted to digital information, how much distance is one "step" (that is, how much more charge is needed to increase a digital value by "one") is strictly arbitrary.

Which is why I specifically mentioned linear RAW data. In that case, the analog charge step is constant. Going from a digital value of a to a+1 should be a step as big as going from b to b+1 for any 0 <= a, b < 2^14. Right? I'd like to get it right if I'm wrong, so please correct me if that is the case. But my understanding is this:

An exposure increase by one stop means the analog charge value is doubled. In a linear conversion to digital, my digital value must also double. So if you want to allow the recorded brightness to double n times, you must also double your maximal digital value n times. Doubling a digital value means shifting it by one bit obviously and so to keep your numbers from being cut off, your bit count has to increase by one.

If we're not talking about a linear conversion between charge and digital values, it is clear that there is not limit for how much DR fits into any given number of bits. But I was talking about linear data, as that is what we are used to in camera sensors. And for those, does my above reasoning not hold up for the claim that a 14 bit file holding linearly converted data can't exceed 14 stop DR?

 

[Michael Clark]

All the papers I've read so far on digital sensors point out they're linear. A single pixel can have more than 14 stops of well capacity, but it's then clamped and converted lineary in ADC. So the number of bits in ADC is the upper bound for the resulting dynamic range.

The number of bits in ADC set the upper bound for the number of steps possible between the analog charge value assigned to "0" and the analog charge value assigned to 2^n. There's nothing that says how large or small each of those steps have to be. Nor is there anything that says how far above an absolute "zero" analog charge (that doesn't really exist) an analog charge should be set as the "0" digital value. There's also nothing that says the 2^n digital value can only be assigned to a full well value.

 

[Michael Clark]

I'm sorry, but I'm not familiar with Adam and his Zone System I think.

And for linear RAW data, why is 1 bit different than 1 stop of DR, if noise is ignored?

Which is why I specifically mentioned linear RAW data. In that case, the analog charge step is constant. Going from a digital value of a to a+1 should be a step as big as going from b to b+1 for any 0 <= a, b < 2^14. Right? I'd like to get it right if I'm wrong, so please correct me if that is the case. But my understanding is this:

An exposure increase by one stop means the analog charge value is doubled. In a linear conversion to digital, my digital value must also double. So if you want to allow the recorded brightness to double n times, you must also double your maximal digital value n times. Doubling a digital value means shifting it by one bit obviously and so to keep your numbers from being cut off, your bit count has to increase by one.

If we're not talking about a linear conversion between charge and digital values, it is clear that there is not limit for how much DR fits into any given number of bits. But I was talking about linear data, as that is what we are used to in camera sensors. And for those, does my above reasoning not hold up for the claim that a 14 bit file holding linearly converted data can't exceed 14 stop DR?

The vast majority (maybe all) of consumer cameras (consumer in the sense that includes even "pro" cameras used to make photos instead of take scientific measurements) do not ignore noise and they do some noise processing to the analog signal before it is digitized. Some do it more than others, which is why some of them have reputations as "star eaters" when they eliminate the signal from weak stars as noise. There's no true "0" analog value for a photosite that collects no electrical charge from photons. By the time the analog signal from such photosites reach the ADC, some of the values can even be negative. So even in the most simplistic methods the starting point for where the first amount of analog charge that is assigned a digital value greater than "0" is more or less arbitrary.

 

[Quarkcharmed]

The number of bits in ADC set the upper bound for the number of steps possible between the analog charge value assigned to "0" and the analog charge value assigned to 2^n. There's nothing that says how large or small each of those steps have to be. Nor is there anything that says how far above an absolute "zero" analog charge (that doesn't really exist) an analog charge should be set as the "0" digital value. There's also nothing that says the 2^n digital value can only be assigned to a full well value.

The number of electrons and therefore the voltage in a pixel is a linear function of number of photons, and the voltage is then converted to a digital value via ADC. Now with the steps, they're normally defined as the smallest detectable level of the signal, effectively the noise, and the DR is defined as max level/noise level, so by definition a linear ADC can't produce a DR larger than its bitness.

 

[Joules]

The number of bits in ADC set the upper bound for the number of steps possible between the analog charge value assigned to "0" and the analog charge value assigned to 2^n.

Yes, I see where I got confused now. I had assumed a proportionality of 1 between analog and digital, but you are clearly right on that this assumption is flawed. You can express a linear function that maps an analog to a digital value for any given range of analog and digital values.

d(a) = (a - cMin) * 2^n / (cMax - cMin)

where a is the analog value, d(a) the digital equivalent, cMin and cMax the min and max value for a we want to record and n is the number of bits.

Glad you brought that up, thanks.

 

[Kit.]

The number of electrons and therefore the voltage in a pixel is a linear function of number of photons, and the voltage is then converted to a digital value via ADC. Now with the steps, they're normally defined as the smallest detectable level of the signal, effectively the noise, and the DR is defined as max level/noise level, so by definition a linear ADC can't produce a DR larger than its bitness.

The problem with this approach is that if the number of photons in the signal as measured by such a linear sensor can be exactly zero, the DR is infinite.

 

[Quarkcharmed]

The problem with this approach is that if the number of photons in the signal as measured by such a linear sensor can be exactly zero, the DR is infinite.

But it cannot...