Canon to release major firmware update for the Canon EOS R5

[neuroanatomist]

Why?

Because the company that’s led the market for 20 years should be terribly concerned about competition from the company that’s lost 2/3 of its market share in the past few years?

 

[tomislavmoze]

That's really not how Clog2 works. It's not a magical fix all that gives your camera sensor more dynamic range. It's a specific tool that high-end cameras with 16+ stops of dynamic range use to get the most out of their sensors. Just adding it to any camera doesn't magically make it better. Canon said on launch day of the R5C that the reason it didn't have Clog2 is that the sensor wasn't capable of it. So adding it wouldn't make any difference.

Actually you are wrong, the sensor does have more dynamic range than what clog 3 offers and it is easy to test, shoot RAW video, transform it to clog2 and you will see just by comparing it to the Clog3. Or you can compare CLog3 with the photo files of the same camera.
There is more clipping happening shoting Clog3, in Gerald's video about R5c you can see it clearly, and Some people explain it well in the comments.
There where done so many tests regarding that and It is proven that Clog2 would bring additional dynamic range to the file in the highlights.
I think canon didn't go with it cause there is a huge amount of noise using Clog2, especially in case of R5.
R5c has a much cleaner image and actually slightly better DR when shooting video so there is definitely room to improve at least the image of R5.

 

[tomislavmoze]

Any steps that Canon take to prevent overheating are a winner for me. I don't shoot video and probably never will, but I frequently work in hot tropical climates where the camera body can become almost too hot to hold comfortably. Hot cameras produce more noise and shorten sensor life. They probably also shorten the life of memory cards and increase the risk of data corruption.

I do really think that is the main reason Canon went so hard on the overheat control on both R5 and R6, to maintain the IQ and also prevent sensors dying from shooting few hours in an overheat mode. At the end after they had enough data they did expand the limit on the R5. The R6 didn't get so much love cause of the R6mk2 which is really a beast of a camera, having an opportunity using it for a while now comparing and it with the R3,, canon did definitely produced a really great camera that can easily replace the R3 for many.

 

[codym90]

I don't see why there should be a mark 2 because how can the R5 get better. Like maybe perfect white balance lol.
-Cody
Wedding photographer in Maryland

 

[Famateur]

While the AF is on-sensor, it is not image based. It uses the phase difference between 2 adjacent subpixels, which can be thrown off by things like UV filters.

I would welcome a mode where the camera does use image data to fine tune the last steps of acquiring focus, it would fix a number of Canon AF annoyances.

Thanks for your help.

I'm still a bit confused (I must be getting slower as I age).

Because the phase-detect pixels are on-sensor, focus is evaluated on the same plane as image capture. How the light gets to the sensor is irrelevant -- it's just measured once it's there.

How can phase-detect subpixels receive the same light on the same plane as the image sensor at capture time, conclude that the image is in focus (arrows align and turn green), but still capture a back-focused image?

Why does focus peaking work accurately but the manual focus assist box/arrows do not?

In my mind, focus peaking and the focus assist box are both visual aids to alert me when focus is detected as I turn the focus ring.

How can they produce different results?

(In case it wasn't obvious, I'm only talking about manual focusing. It makes sense why a lens would need electric communication to/from the processor to know which direction to adjust focus and how far. I don't see how it is necessary for manual focusing since the processor is communicating with the operator via visual cues, then detects the changes on-sensor as the focus ring is turned.)

 

[TonyG]

Adding Clog2 doesn't magically give the sensor in your camera more dynamic range. The sensors with 16+ stop of dynamic range have clog2 as that's designed to get the most out of those sensors. It's like asking if I put formula one tyres on my vehicle does it make it a race car. No, it doesn't. Those tyres are designed and calibrated to work on an entirely different machine. The tyres don't make the Formula one car what it is but they do help get the best out it. If that makes sense? I don't drive. I don't even know why I'm making car analogy's.

Correct, it won’t magically give the sensor more dynamic range, but what it will do is utilize the sensors dynamic range.
The R5 is measured at 14.6, yet the a7siii is 13.9. Yet the a7siii has more dynamic range? Oh right it’s what the slog3 profile can do.
Your analogy should be compared to the tuning on the computer that makes the car go faster.
A more conservative tuning and you lose HP.

 

[TonyG]

I don't see why there should be a mark 2 because how can the R5 get better. Like maybe perfect white balance lol.
-Cody
Wedding photographer in Maryland

Dual cf express slots
Digital hot shoe for the new flashes
14 bit electronic shutter
Better rolling shutter

 

[Scenes]

Correct, it won’t magically give the sensor more dynamic range, but what it will do is utilize the sensors dynamic range.
The R5 is measured at 14.6, yet the a7siii is 13.9. Yet the a7siii has more dynamic range? Oh right it’s what the slog3 profile can do.
Your analogy should be compared to the tuning on the computer that makes the car go faster.
A more conservative tuning and you lose HP.

You're right in that the computer analogy is better. So, what you're saying is that you have a 3070 and if you put 4090 drivers on it, it will somehow perform like the 4090. That's incorrect. Let's say you're correct at 14.6 stops. Clog2 is for cameras with 16+ You might 'feel' it give you more than clog3. You might prefer the way it initially looks as the gamma curve is different in the balance of shadows and mid tones. But that doesn't mean it's factually correct.

 

[Scenes]

Actually you are wrong, the sensor does have more dynamic range than what clog 3 offers and it is easy to test, shoot RAW video, transform it to clog2 and you will see just by comparing it to the Clog3. Or you can compare CLog3 with the photo files of the same camera.
There is more clipping happening shoting Clog3, in Gerald's video about R5c you can see it clearly, and Some people explain it well in the comments.
There where done so many tests regarding that and It is proven that Clog2 would bring additional dynamic range to the file in the highlights.
I think canon didn't go with it cause there is a huge amount of noise using Clog2, especially in case of R5.
R5c has a much cleaner image and actually slightly better DR when shooting video so there is definitely room to improve at least the image of R5.

Your mistaken. Raw is developed with Clog2.. That doesn't mean it shoots clog2. The R5c has a much cleaner image and better DR than the R5 because it has a different sensor and processor than the R5. They're similar but they are very different, hence the increased capabilities of the R5C.

 

[neuroanatomist]

The R5c has a much cleaner image and better DR than the R5 because it has a different sensor and processor than the R5.

I’m not a video shooter, but when you make the claim that the R5 C has a different sensor than the R5, you pretty much trash your credibility across the board.

Canon says: “The EOS R5 C and the EOS R5 share the same 45MP full-frame CMOS sensor…”

 

[David - Sydney]

Dual cf express slots
Digital hot shoe for the new flashes
14 bit electronic shutter
Better rolling shutter

Dual CFe cards will generate more heat but would allow dual recording for video. Cost vs USH-II SD cards isn't very different. Unlikely I think.
New hot shoe would also fix a relatively common mechanical weak point
14 bit electronic shutter would be great but can only be handled by the R5's current Digic X if the fps is lower than 20fps due to processor bandwidth issues. Would be great to have 14 bit/12fps
Better rolling shutter would be a new sensor and probably stacked to get the read out speeds like R3/A1/Z9. Hard to find definitive comparisons for times though (video and stills). I wouldn't be holding my breath.

 

[sanj]

The R5C also has a different sensor and CPU. They're similar but not the same as the R5 which is why the R5C can do more.

Different sensor? I thought it was all about adding a fan...

 

[Famateur]

how did you get the arrows with a nikon lens, on a r5 ?

In AF menu 2, I set "Focus guide" to "On".

The lens is attached using a Photodiox adapter. It's brilliant with focus peaking. Makes achieving focus fairly quick. The IBIS in the R5 is an extra bonus. Plus, I get to enjoy an old 85mm F1.8 and a 105mm F2.5. Neither are exceptionally sharp, but they're fun to use and produce their own character, if that's not too overused of a term.

 

[Famateur]

I should probably contribute to something on-topic.

I'm not putting a lot of hope into the mysterious "above and beyond" feature speculation, but I'd be thrilled if the AF system was updated to allow tracking from all AF methods like the newer bodies do.

What can I say -- I'm easy to please.

(Sometimes I have to remind myself what a blessing it is to have such a capable camera in the first place. What a time to enjoy photography!)

 

[David - Sydney]

Different sensor? I thought it was all about adding a fan...

and removing IBIS and cinema menus (with lag to reboot between)

 

[koenkooi]

Thanks for your help.

I'm still a bit confused (I must be getting slower as I age).

Because the phase-detect pixels are on-sensor, focus is evaluated on the same plane as image capture. How the light gets to the sensor is irrelevant -- it's just measured once it's there.

How can phase-detect subpixels receive the same light on the same plane as the image sensor at capture time, conclude that the image is in focus (arrows align and turn green), but still capture a back-focused image?

Why does focus peaking work accurately but the manual focus assist box/arrows do not?

In my mind, focus peaking and the focus assist box are both visual aids to alert me when focus is detected as I turn the focus ring.

How can they produce different results?

(In case it wasn't obvious, I'm only talking about manual focusing. It makes sense why a lens would need electric communication to/from the processor to know which direction to adjust focus and how far. I don't see how it is necessary for manual focusing since the processor is communicating with the operator via visual cues, then detects the changes on-sensor as the focus ring is turned.)

I’m at the same point, I know that DPAF can get thrown off by things that don’t impact contrast detect, but I don’t know why. If someone has a link to an explanation,
please share it

 

[Del Paso]

Thanks for your help.

I'm still a bit confused (I must be getting slower as I age).

Because the phase-detect pixels are on-sensor, focus is evaluated on the same plane as image capture. How the light gets to the sensor is irrelevant -- it's just measured once it's there.

How can phase-detect subpixels receive the same light on the same plane as the image sensor at capture time, conclude that the image is in focus (arrows align and turn green), but still capture a back-focused image?

Why does focus peaking work accurately but the manual focus assist box/arrows do not?

In my mind, focus peaking and the focus assist box are both visual aids to alert me when focus is detected as I turn the focus ring.

How can they produce different results?

(In case it wasn't obvious, I'm only talking about manual focusing. It makes sense why a lens would need electric communication to/from the processor to know which direction to adjust focus and how far. I don't see how it is necessary for manual focusing since the processor is communicating with the operator via visual cues, then detects the changes on-sensor as the focus ring is turned.)

Focus peaking only suggests more or less sharp areas. But if you look closer (on pixel level), you'll notice that the definition of "sharp" is quite vague, while the "triangle" method yields a much higher "on the spot" precision.
Focus peaking can be helpful, but not for determining with precision which part of a flower, for instance, is in perfect focus. I'd rather use the "loupe" function for critical work, like macros.

 

[Famateur]

Focus peaking only suggests more or less sharp areas. But if you look closer (on pixel level), you'll notice that the definition of "sharp" is quite vague, while the "triangle" method yields a much higher "on the spot" precision.
Focus peaking can be helpful, but not for determining with precision which part of a flower, for instance, is in perfect focus. I'd rather use the "loupe" function for critical work, like macros.

I'd love to have that increased precision, but for some reason, the triangles always indicate focus is achieved when it's actually back-focused. It's a head scratcher...

 

[Bonich]

My understanding is that the loss of DR with ES typically amounts to about 1 stop. The loss of DR is due to a reduction in bit depth (12 bit mechanical, 10 bit electronic in most cases), and that is done purely to increase readout speed. AFAIK, there is no reason why electronic shutter can't run on 12 bit and thereby retain the full DR. Presumably it's dictated by sensor design and processor power. Anyone with greater knowledge please contribute.

You are right the reduced bit depth to be a contribution to max readout speed.
This is why I think to have a "slow mode" for frame rates less than 10 FPS to be separated from high frame rates (incl. video) would make more sense than coupling bit depth to shutter mode.
I would like to shoot exposure bracketing, focus bracketing, time lapses, long exposures, .... all with pure electronic shutter but at max bit depth.
There is no technical limitation hindering this strategy, simply implementation issues while writing SW.

 

[tomislavmoze]

Your mistaken. Raw is developed with Clog2.. That doesn't mean it shoots clog2. The R5c has a much cleaner image and better DR than the R5 because it has a different sensor and processor than the R5. They're similar but they are very different, hence the increased capabilities of the R5C.

Shooting raw allows you to use the maximum dynamic range from the sensor, In case of canon it is from 12bit raw since it does not use the full 14 bit for video, probably to contain faster readout speeds.
When you shoot raw with any camera you can chose to transcode the image in to the different gama curves, because of that you can compare different log profiles since you can transcode raw to both Clog3 and Clog2, if you do that you can obviously see that there is more info in shadows when you chose to transcode it to clog2.
That does not mean that it will allow the sensor to produce 16 stops of dynamic range but it will allow it to produce its maximum, and maximum for clog 3 is 11,5 stops and for clog2 with noise reduction applied in post production on R5 is almost 13 stops so with Clog2 transcoding you get one more stop of DR, but you need to ad noise reduction to get that result.
With The R5c in clog2 you get 16 stops ( According to Gerald Undone test, which is really well done) without NR applied and with NR applied it is closer to 14 stops. which is an amazing difference.

And no the R5 and R5c do not have different sensors, maybe R5c has an improved processor although canon never claimed that as with R6mk2 so I doubt even that. Where did you pull out that info?
It was clearly stated that the only difference is software and codecs in combination with cooling which allows the better IQ.
Canon cinema development team is just better at utilising sensors for video which showed the full potential of the R5 sensor for video.

Correct, it won’t magically give the sensor more dynamic range, but what it will do is utilize the sensors dynamic range.
The R5 is measured at 14.6, yet the a7siii is 13.9. Yet the a7siii has more dynamic range? Oh right it’s what the slog3 profile can do.
Your analogy should be compared to the tuning on the computer that makes the car go faster.
A more conservative tuning and you lose HP.

There is one difference, Canon uses 12bit image for video, Sony uses 14bit image and some of the differences come from that. Also that is the reason most of the Sony bodies have much worse readout speeds and stronger rolling shutter but do produce more DR in video. A7sIII being the best from Sony due to having only 12mpx sensor and it doesn't oversample from higher resolutions to 4K it is basically 1/1 readout.
I do wish canon will introduce an Rc camera with R3 sensor, that would be a big win for them in that hybrid, entry level market.