Re: Canon EOS R10 Sensor Measurements at PhotonsToPhotos
cocoanud wrote:
Ephemeris wrote:
cocoanud wrote:
RLight wrote:
Ephemeris wrote:
RLight wrote:
Bill,
My thanks as always for the hard work. It also shows that R10 is quite the workhorse for the size, price. I look forward to more Canons with the updated sensor design.
We picked one up for £775 on a Christmas deal for a present. Thinking back some years to what that amount bought as a DSLR it's certainly progress.
I suspect this sensor will make it's way into a few bodies. Gaining full 4K readout and RAW NR, and apparently Dual Gain is a big deal. It makes this old cat far more competitive, even without process change (BSI or stacked).
I'm impressed by the sheer amount of blood Canon has been able to squeeze from this turnip (24MP APS-C sensor). It's been around since, T6i? Long time Canon's been hanging unto 24MP APS-C without moving to BSI, etc.
Edited... Looks like Canon shifted to dual gain too, hence the DR uplift. Pretty impressive for being an old FSI process, just updated design print.
Canon needs to stop pulling an Intel though here, and shift to a newer die/process, eventually. Sheesh. Not to make jokes about Kaby Lake, etc (tick tock, tock, tock?)
I think the real secret sauce might be in DigicX ?
I have tried Digic8 based Canons (RP, G1X iii, M6 ii) and have been disappointed. R10 is the true tiny R camera which doesn’t leave me wanting… the compromises to be made are fair due to the price. M6ii was quite snappy though… but R10 has more AF prowess and is even more snappier.
I think if it makes business sense for Canon they might continue to milk this cow. Would be awesome to see G1X iv or even G3X with DigicX… Canon still need a competitor to Sony RX100 and RX10.
I would agree though, it is curious why Canon are still holding on to FSI process. Maybe their supply chain commitments run too deep… but I can only guess.
What is the purpose, or more formally what are the requirements that you think may apply to make them change from what we have today?
The one thing that (I hope) should matter will be Customers/users not being happy with the offerings when compared to competitors at a given price.
Seemsije customers are happy if sales and rentals are a way of a metric. What is the origin(s) of your speculations?
Personally, I am happy, almost over the moon, may be because I moved from m43 to Canon FF and before I fully committed to Canon I also tried a Nikon Z6 for some time and in my experience the R6 is the superior camera in more than one ways with what Canon have in the lineup.
With the R6 I can even forget about shooting RAW and can happily shoot JPEGs and not care. No complaints about the R6 and no complaints about the R10 except that we need good quality lenses for the RF-S mount... which in a way are already there albeit in the form of larger RF form factor.
As long as the IQ, AF and ergonomics are great... I don't have complaints and don't really care whether it is FSI or BSI.
It would seem you have some care?
Switch to BSI will cost Canon and that cost will be passed on to us so there's that to consider too. Having a BSI process built sensor definitely has its advantages... but are they worth the cost and will the customer benefit ? If not, then simply "Don't fix if it ain't broken"
If there are sufficient number of Canon users like me (ie satisfied happy customers) I think there may not be a reason for Canon to rush into such decisions.
An interesting example is Nikon D6 vs Canon 1Dx III , simply compare the PDR on photonstophotos.net . It seems as if that D6 sensor is not just a dual-gain sensor rather a multi-gain sensor and the 1Dx III still beats it in DR even with RAW NR and no apparent multi/dual-gain points. My guess(and I emphasise again on guess ) is Nikon must have sunk quite a lot of money into Sensor R&D costs coming up with that type of sensor whereas Canon worked it all out mostly on the DSP side of things i.e. in the DigicX. Will the customers care whether the D6 had a technically superior sensor or the end result ? And btw.. D6 is still a great camera and so is the 1Dx III ... I know where I would have put my money if I had the need for those types of cameras. (https://photonstophotos.net/Charts/PDR.htm#Canon%20EOS%201D%20X%20Mark%20III,Nikon%20D6)
Customers by systems rather than sensors and there is a lot of design work between it and the micro.
Do you think perhaps being a camera leader is enough for them to continue to deliver on their path?
Canon really took a hard beating from Sony and R6/R5 were the game changers which brought Canon back neck to neck with Sony. Sony still have them beat in other segments (for eg RX100 and RX10 type cameras) and lets not even get into lens situation.
We seemed to have missed the R and Rp. We had an R which provided many benefits over the 5dIV (our work horses). Is beating the correct description?
If Canon are able to produce knock out numbers as they have been and customers are happy then it maybe status quo for some time. That brings to the next point i.e. DigicX. All these excellent cameras ie R5, R6, R7, R10 have DigicX
We talk about Digic X. However are they all the same? Is it more a platform name?
No idea and all of that is Canon proprietary IP. All that is known is that DigicX is really an ASIC based on (or built around) ARM Cortex9 IS.
This shows some overviews of differences between the camera platforms. I don't have details on the structure of the micro but given the reliance Canon has I would assume it's custom.
https://snapshot.canon-asia.com/article/eng/canon-technology-explainer-what-is-digic
I am not sure I understand fully what you mean when you ask whether it is a platform, but I am guessing you mean whether it is customised for different product ranges ?
Indeed.
Again, don't know the answer to that. There are difference between I/O bandwidth when R6 is compared to an R10. R10 will shoot faster than R6 but R6 has much deeper buffer for e.g. In that sense the cameras themselves are platforms which use the DigicX. There could be firmware differences which when applied to different bodies change certain parameters.
Also, my understanding is that having customised ASICs for different body/products doesn't make sense. Could turn out to be way too expensive than it should be... a better way is to implement the ASIC as generally as possible so that it can be used in pretty much any camera body and then work with firmware and system integration to create different platforms (same as the R10 vs R6 example above about I/O bandwidth). But don't take my word for it... I am only conjecturing here, I don't work for Canon don't even work in electronics hardware industry for that matter.
Platform ASICS are not uncommon whereby different capabilities are not created to reduce costs.
