Canon @ CP+: Mirrorless strategy, development focus and Dual Sensing IS tech
posted Sunday, April 1, 2018 at 7:12 PM EDT
For a good few years now, we've had a team from IR on the ground at the CP+ tradeshow in Japan each spring, attending an annual event which represents something of a homecoming for many of the world's biggest camera makers. And each year, we make sure to schedule time around the show to meet with executives from each of these companies, both to discuss all of their newest products and technologies, and to pass on feedback and questions from our readers. (As well as from our own experiences, of course.)
This year, we kick off our CP+ 2018 interview series with the fruits of IR founder and publisher Dave Etchells' visit to the always-impressive Canon booth. Dave was fortunate to be able to meet with three of Canon's senior executives for a discussion on a wide range of topics including the company's mirrorless plans, what's next for its professional DSLR cameras, and some very interesting insight into the company's Dual Sensing IS technology. Representing Canon Inc. were Go Tokura (Executive Officer and Chief Executive of Image Communication Products Operations), Naoya Kaneda (Advisory Director, Group Executive, ICB Optical Business Group, Image Communication Business Operations), and Yoshiyuki Mizoguchi, Group Executive, ICB Products Group, Image Communications Business Operations.
Without any further ado, let's get right down to the interview!
Dave Etchells/Imaging Resource: I appreciate very much you making the time. Domo arigato.
My first question has to do with mirrorless and the market split. Last year, you mentioned seeing growing demand for mirrorless; even though the market is going down, mirrorless is increasing. And in particular you said that for the Japanese market, there was a fifty-fifty split between customers for mirrorless and DSLR.
In the last year, have you seen a shift one way or the other in that? And how about in other market regions? In the US, is there still more demand for EOS DSLRs than for mirrorless?
Yoshiyuki Mizoguchi/Canon Inc.: Generally, there's not that significant a difference, but having said that, although we said fifty-fifty, it has grown slightly beyond 50-some percent. In terms of other markets, if you look at the US for example, for 2017 it was 20-some percent. For Europe, it's in the mid-30%, [and] China [is] also mid 30%. But overall, when we compare against 2016, there are slight increase for all [markets]. We can say that for Japan and other regions, the mirrorless market is increasing on a slight level.
DE: Interesting that the US is still so much lower than Europe and [Asia].
YM: That seems to be continuing.
DE: Yes. And while you continue to develop your mirrorless solutions, we can't help but feel that DSLRs are Canon's first love and your main focus. But on the other hand, competitors who don't have Canon's sort of market strength with DSLRs have embraced mirrorless very aggressively, and they're rapidly developing pro-level performance that used to only be available with DSLRs. Where does mirrorless fit in Canon's current strategy, and do you see Canon mirrorless cameras ever becoming mainstream professional offerings?
Go Tokura/Canon Inc.: Basically, our strategy remains the same. And what we mean when we talk about our strategy, we have a full lineup strategy, and what that means is that we want to provide the hierarchy for all products. For the optical SLR as well as the mirrorless [cameras].
But having said that, we can see the volume zone to be in the entry level, and when we say that, it's really important for us to maintain our share in that entry level market. And the demand for that is the mirrorless, hence the reason we have started to focus on [mirrorless].
But going back, because we [have] the full lineup strategy it is always in our mind that we actually want to satisfy different levels with both mirrorless and optical SLRs, so that we can satisfy the demands in all areas. We want to win in all the areas.
[Ed. Note: I found this interesting, that Canon's ultimate plan is to have full lines of products from entry-level through professional, for both SLRs and mirrorless models.]
DE: Ah, that's very interesting, so you see the demand for mirrorless being strong in the entry level currently?
GT: Yes, exactly.
DE: Interesting. But as we go into the future and technology advances, do you see mirrorless moving into higher-end applications, and moving into the professional world as well?
GT: Yes of course.
DE: Haha, I ask long questions, with short answers sometimes.
GT: Sorry, sorry! <laughs> So, to give a background to the short answer, as the technology [is] advancing, customer demands [are changing]. Once again, because we have the full lineup strategy, we want to fill up the gap where we still haven't satisfied the demand. So in other words, the high-end mirrorless [market] is in our scope.
DE: We've seen still image capture rates are increasing and becoming similar to video frame rates, and it seems we're coming to a point where the two different modes are converging. You yourselves, in your 5D Mark IV, have an option to extract still frames from 4K video sequences. Panasonic has 4K and 6K Photo.
How do you think things will evolve from here? Will we see continued convergence, and one day will continuous shooting just be "capture 4K video and extract frames"? Or will there be a continuing role for the continuous mode still-image capture approach? How do you see this whole area evolving?
YM: Just a recap: [For] video, it has tended to be that you have lower pixel numbers but you have a higher frame rate. On the other hand, with still photos, you have higher pixel numbers but with a lower frame rate. But [alongside] the arrival of 4K and then 8K [ultra high-definition video], the frame rate has increased for still photos as well.
So yes, there is a movement where we tend to extract still images from a video. And with 8K, obviously it gives a large enough pixel number than it can withstand to be a proper still photo, so that is going to be a big sort of a turnaround, see?
DE: So you think 8K will be a transition point?
YM: That will be one of the milestones, I guess you can say. What that would mean is yes, there is a [convergeance] that would probably increasingly happen where we have video and still [capture] coming together, and that the whole concept of image capturing might have a new concept being introduced.
But having said that, we still believe there is a role that still image capturing plays, and also a role that video image capturing plays. They will remain as they are, because with still photography there's always a demand for that proper focus for that moment. And that will never die out, we believe.
DE: Yes clearly, still will never go away but for continuous shooting, we may see that becomes a video with extraction. Hmm, I guess that's not a question. <laughs> I'm just trying to summarize what I think the situation is...
YM: We still believe that for continuous photography and video coming together, there needs to be a lot more technological progress that has to happen for that to be a proper feature. So, I think it will remain as is for the time being.
DE: We seem to still be seeing kind of a megapixel race. Many manufacturers are making cameras with higher and higher resolution. Your EOS 5DS and 5DS R cameras first came out in 2015. Can you share what the response has been to these cameras? Have they met your expectations in terms of the size of the customer base? And has the 5D Mark IV impacted potential customers for the 5DS and 5DS R, or are those really two completely different market segments?
YM: In terms of the 5DS and 5DS R, people who look for higher pixel numbers, they do like it very much. We had a really good market response. And in terms of the launch of the 5D Mark IV, from what we have seen so far, we haven't seen any difference in the take up of customers who want higher pixel numbers. People who want high pixel numbers still go to our 5DS and 5DS R.
DE: On a related note, we've seen various competitors develop quote-unquote "high-res" capture modes that use sensor shifting to increase resolution. Is this something Canon is exploring at all, or are you looking at a different direction entirely when it comes to, like say, very high-resolution photography?
YM: As you know the sensor shift technology, it might work for still photography, but for video, it makes no sense.
DE: I know, yes.
YM: As we have already announced the hundred million-pixel numbers. We are pursuing that route, but we believe that it's important for the technology to satisfy both still and video. Unless it does that, I don't think the technology would be successful. [Ed. Note: Canon has publicly demonstrated both 120 and 250-megapixel image sensors in the sub-frame APS-H format after first revealing their development at Canon Expo 2015.]
DE: Talking about sensors, they're obviously at the heart of digital cameras, and that's an area where Canon was an early leader with your CMOS technology. Now, though, competitors have taken advantage of huge production volumes of cell phone sensors to advance their underlying technology.
Canon hasn't had a presence in that market in a way that would help drive its technology. But on the other hand, three years ago you acquired security camera maker Axis. They have a very large volume for security cameras. Do the Axis cameras use Canon sensors and if so has the volume demand there for low-light and high-dynamic range sensors helped to advance your underlying sensor technology in ways that benefit the photographic market?
GT: In terms of the Axis sensor technology, that hasn't really fed into our camera side of the business. But having said that, at the head office we have the B2B business segment where we also have our network visual system cameras. And for this [business] area, we actually have learned from Axis in terms of getting the high sensors [Ed. Note: We're guessing they meant either high-resolution or high dynamic range, but we're not sure which] into the cameras, and they're working accordingly with Axis.
Having said that, we still believe that the technology within the security camera segment and also the consumer segment, they all have ways to sort of merge and if need be, learn from each other to really elevate the standards.
DE: Yes. Well, you obviously can't talk about future products, but it's been a couple of years now since you introduced your current flagship, the 1D X Mark II, in February 2016. Even by current standards, it's an amazing, super-fast camera with excellent autofocus.
Where do you see customer demand leading you going forward in that part of the market? What are pro level customers asking for the most? Is it higher shooting speeds, even more powerful autofocus, expanded video capabilities... they probably all have some level of demand, but is there anything that dominates? Just generally, where do you see that highest-end professional market trending?
YM: When we say professional photographers, we're talking about various types of subjects that they actually take photography for. So it's difficult to say "this one feature is the one that really stands out." But if we were to say, maybe the sensitivity, that might be something where they're seeking more progress or advancement.
Also, speed and sensitivity in auto focus; that they all seek. But at the same time, I think reliability is very important. And also, that they can shoot in no matter what kind of condition they are. Whether it be rain and whatnot. So, water / dust resistance factor is an important part. And I think that's what all professional photographers would seek for as a function, something that they could rely on. A camera that they can rely on, that they can shoot anywhere.
GT: With the 2018 PyeongChang Olympics, the Winter Olympics which just ended, we hear that all of our cameras performed very well, even under the -20°C (-4°F) conditions. Having said that, we've heard some talk about other competitors' cameras who didn't do so well in the conditions. So again, it goes back to the reliability aspect.
DE: Confidence in reliability, yes. I have sometimes made a joke that professional photographers want to use a Canon 1D body to take pictures, and then also use it as a hammer to pound their tent pegs into the ground :-)
YM, GT: <Laughter> Please don't do that!
DE: Turning to a particular technical question, we were a little confused by the Dual Sensing IS feature that's in the M50 and some of your PowerShot models. There is no sensor movement in those cameras, so we are assuming that Dual IS means lens-based IS combined with electronic stabilization as the data is read from the sensor. Is that correct?
YM: Ok... just to start from the beginning, the lens has a gyro sensor that actually detects any sort of movement and the camera has an image sensor built in. In order to adjust the image so that it is sharp, the gyro sensor is used to gather the necessary motion information.
With Dual Sensing IS, the information is actually read with the gyro sensor on the lens, but also with the sensor within the camera as well. So, we have two sources of information that actually allows the information to be fed in so we can stabilize the image.
DE: And the stabilization is happening only optically in the lens with a moving element or does it also happen has data is read from the chip? Sometimes you can stabilize a video image by choosing where you start reading the data out?
YM: The dual sensing IS is only for still photography, first of all, so it doesn't actually apply to video. We don't actually have a video image that comes in, where we can kind of select where to start the recording.
DE: Yes, right. Still images only...
GT: The adjusting does happen only in the lens, but [what is actually done is that] we can actually get the [movement data] from the lens' gyro sensor, but from on the body of the camera with the main image sensor, which actually detects the movement of the subject. So, those two information are combined and fed into the lens, [which] does the adjusting accordingly.
DE: And you say the image sensor, you've mentioned subject movement, so [you're not detecting] just moving the entire scene but you can actually track the subject itself moving within the frame?
Naoya Kaneda/Canon Inc.: The lens has the actual sensing capabilities, so it can detect the movement and then it will also do the adjusting on its own, [that] works fine. But let's say the lens wasn't able to capture the delicate movements that it might have missed. Then it's the camera sensor that comes in to feed the additional information, [detects that] there's slightly more movement that we can actually do more adjusting for, and that information gets fed into the lens and the lens is… does the adjusting.
DE: It's like fine tuning.
NK: Exactly, it actually realizes higher levels of image stabilizing.
[Ed. Note: This is extremely interesting! I'd imagine it would only be useful in relatively long exposures, but what they're implying is that the camera is continuously reading data from the image sensor -- without disturbing the image's exposure -- and then processing it to detect any small displacements that might be left over after the lens has shifted the image based on what the gyro sensor is seeing. They're apparently using some sort of an optical flow algorithm to sense small shifts in the image data in near-real time, and then tweaking the movement of the lens' IS element slightly to compensate.
This is the first I've heard of a company doing this in still-image cameras. There's obviously a lot of technical and engineering depth here, as the camera has to be looking for very small displacements (sub-pixel, on average?), and has to catch any deviation quickly enough that it won't affect the bulk of the exposure for the image in question. I'm 100% speculating here, but wonder if this is somehow an offshoot of Canon's Dual Pixel AF technology? In this case, it doesn't have anything to do with focus, but the architecture that lets Canon read and process phase-detect focus data separately from main-image data could also be used as a separate channel for reading pixel data to use in an optical-flow algorithm independently of the main exposure itself. And, thinking about it, the sort of correlation functions needed to implement phase-detect focusing involve very similar processing to that needed for optical flow determination. Very interesting, and something I expect we'll see more of across Canon's entire product line, going forward.]
DE: Very interesting. Probably easier to get to longer exposure times with… [*pauses, then realizes that I'm about to go down a long rabbit hole but am almost out of time*]
We should move on. I could talk about technology all day...
DE: So, my last question is about entry-level interchangeable lens cameras. Even though we are an enthusiast-oriented site, we see a lot of traffic to entry level models like your own T6 and now T7, or the Nikon D3400. How does the entry level end of the ILC market look to you? And I'm meaning the entry-level market as a whole, regardless of whether it is mirrorless or SLR.
Do you think that even lower prices for basic interchangeable-lens models could stir up more market demand, or is the demand there more or less price-inflexible, meaning not influenced by price? Do you think the ILC market could be expanded if there were lower cost options available?
GT: I think there is a great potential for growth in the ILC market, given that the customers are diversifying. There are various types of customers who are looking to take photography and also the instances of taking photos and taking videos is exponentially expanding given the current culture of where we are.
And one piece of data that we can give is for example people who do take photos mainly with their smart phones, many of them are actually not happy with the level [of image quality] that smart phones provide. So that's why I think there will be a stream of customers who would come in who would want a proper camera to take a proper picture at a satisfactory level.
DE: If you can get the price lowered.
GT: We don't think it's about the price really, because if the camera does provide that excellent performance I think customers will be willing to pay the price. So that's why we're always evolving our features, evolving our performances, so we don't have to lower the price but give them more at the price point.
DE: More features and smaller or whatever to make it more attractive, right.
GT: iPhones are quite expensive as well.
DE: No kidding. Oh, we ran over time, I'm sorry! Thanks so much for taking the time to meet with me!
[Ed. Note: This last little bit was interesting, and seems somewhat at odds with Canon's introduction of the EOS 4000D in some parts of the world, a very stripped-down SLR model using older sensor and AF technology and a much lower-resolution LCD display, at a very low price point. That seems to be exactly the approach that Canon said they weren't pursuing in the interview above.
The EOS 4000D isn't being introduced in the US (at least not presently), and it's not clear whether Canon intends it for some very specific sub-market (education?), but it's odd to see it announced so closely on the heels of an interview in which they said they weren't going to pursue a strategy of building lower-featured models to hit lower price points. I personally think that the EOS 4000D is meant for something other than mainstream sales channels, so in fact doesn't represent a mainline strategy for them, but I guess time will tell...]
This was an interesting interview on several fronts. Here are the main takeaways I saw:
SLR/Mirrorless strategy: Canon's strategy for SLRs vs mirrorless is to have full product lines in both areas, from entry-level to professional. They obviously have a long ways to go on the mirrorless front, particularly at the high end, but we've seen considerable evolution in their APS-C mirrorless models over the last few years. Personally, I'll be very surprised if we don't see some sort of a higher-end mirrorless camera model from Canon before the end of this year.
Next up, more speed: A key focus for Canon in upcoming models will be on shooting speed, as that's where they said they're seeing the strongest demands from professional shooters. (The question phrased in terms of market demands as a way of avoiding questions they wouldn't answer about specific model features ;-) It's also an area where traditional SLR technology is at a disadvantage relative to mirrorless, due to limitations imposed by the moving mirror. My prediction here for Canon's DSLRs is that they're going to provide live-view shooting modes with faster frame rates and enhanced Dual Pixel autofocus, essentially making mirrorless cameras out of their SLRs. Short of that, I don't see any way they can match the kinds of frame rates we're seeing from Sony and Olympus, while still having to flap a mirror up and down for each shot. Of course, if they can deliver the needed frame rate and AF performance in a mirrorless mode on an SLR, that would mean they could also make a pro-level mirrorless camera directly. I think the next year or so is going to be a very interesting time for Canon shooters :-)
Dual Sensing IS(!) I found the explanation of Canon's Dual Sensing IS extremely interesting. I'd very much like to have a more detailed discussion with a Canon engineer about it at some point, but frankly don't expect them to reveal too many details regardless. It's a great example, though, of how ever-advancing processor technology can couple with sensor design to yield new capabilities and performance enhancements. Optical-flow algorithms require quite a bit of processor power to implement, perhaps explaining why it's only now that we're seeing this used as an IS tweak by Canon, even though they've had Dual Pixel CMOS sensor technology for several years now. I predict we're going to see this tech expand across Canon's camera line in the not too-distant future.
EOS 4000D?(??) As mentioned above, the EOS 4000D seems completely at odds with the general strategy espoused by the execs in this interview. Personally, I think the EOS 4000D will turn out to be a narrow niche model aimed at very specific, non-traditional sales channels, vs representing a mainstream strategy. But we'll see...
What do you think? Please share your thoughts in the comments below!