OM Digital Solutions Q&A July 2022: Great outlook, AIAF vs Tracking AF, 150-400mm eye candy and more!
posted Friday, October 7, 2022 at 12:05 PM EST
Regular readers will by now know that I finally managed to return to Japan in late July of this year (2022 for those reading this in the future), after a 2 1/2-year absence due to the COVID crisis. It was great to catch up with so many people I know in the photo industry, not to mention getting to once again visit the country that's come to feel almost like a second home to me. This is now the third interview I've published from that trip and the fourth article overall. For easy reference, here are the previous pieces I've posted:
- Sigma's new headquarters building is GORGEOUS
- Fujifilm Q&A, July 2022: NO problem with punch-in focus, and a lot more!
- Panasonic Q&A July 2022: What’s up with L-Mount, how DR Boost works, autofocus, lenses & more!
Towards the end of my stay, I made my way out to the Hachioji area of Tokyo. It was a familiar trip out on the Chuo Line, because I'd visited Olympus Corporation's R&D headquarters there many times. This time was a little different though, as my destination was the new headquarters of OM Digital Solutions Corporation, a couple of kilometers from the Olympus facility.
As I'm sure all our readers already know, OM Digital Solutions (OMDS for short) is the name of the new company that was formed when Olympus spun off their camera business in October of 2020. This was my first chance to visit them since the spinoff, and I was very interested to see how things were going for them. As usual, I also had a lot of questions about their technology and product plans.
Many thanks to the following people who were kind enough to make room in their busy schedules to meet with me:
Kazuhiro Togashi, Vice President, Global Marketing Division (Above Right)
Hiroki Koyama, Associate Expert, Global Marketing, Marketing Strategy (Above Center)
Naohiro Yamaguchi, Director, Marketing and Product Management, OM Digital Solutions Americas, Inc (Above Left)
Yuka Iwasawa, Global Public Relations, Marketing Communications, Global Marketing (Not shown)
Our discussion lasted for over an hour, covering a wide range of topics, so there's a lot of material to go over. I've organized everything under subheads, to make it easy for you to find the parts you're most interested in. All of the people mentioned above participated in the conversation, but my untrained ear wasn't able to reliably identify each of them from the audio, so I've just shown their replies as being from "OMDS". As usual, I've added explanatory notes and analysis written after the meeting; those sections are inset and set in italic face. (The one exception is the final section about autofocus and AIAF, which I wrote myself as a summary explanation. I made that its own separate section to break up the text and avoid the tedium of so many paragraphs of italics.)
With all that said, let's dive in, Enjoy!
How has the transition to being a standalone company gone?
RDE: Thanks to COVID, I haven’t visited Japan since early 2020, so this is my first time visiting Hachioji since the creation of OM Digital Solutions. Our readers and the community as a whole are very interested to know how things are working out. It had to be a massive organizational change, not just here in Japan but with all your offices worldwide. How has it been going? Have there been unanticipated challenges? How are you overcoming them?
OMDS: About a year and a half has passed since the commencement of our new company in January 2021. At the same time, we underwent organizational reforms while simultaneously revamping our business processes. Although these business process reforms are still underway, the mindset of our employees has also changed.
OMDS: On the other hand, since we were unable to use the infrastructure that we had employed during our days as part of Olympus, building business infrastructure proved to be a significant challenge for us. Progress on some of these efforts was slower than planned, however, we are making progress on solving them with assistance from our vendor companies.
This hadn’t occurred to me previously, but of course makes sense. Modern business relies enormously on information technology. At the time of transition, OMDS was essentially a startup with a global organization to manage and support, but without the IT infrastructure needed to manage it. Not only that, but it was a global organization in the midst of undergoing a radical reorganization. I can’t begin to imagine the magnitude and complexity of creating the IT infrastructure for a global organization from scratch, while the organization itself was still in the process of being redefined. Fortunately, it appears that OMDS is getting a handle on this, thanks to herculean efforts by their IT vendor companies.
How are sales going for OMDS?
RDE: From what I’ve seen in the US, OM Digital Solutions has been doing very well sales-wise since the transition. Is that the case, and if so, can you give some idea of just how well? (I realize you may not be able to release specific numbers, but can you give any relative magnitude relative to previous years operating as the imaging division of Olympus?)
OMDS: It’s a little difficult to compare between Olympus and our business these days because there has been a dramatic change in the camera market and market environment, in part because of smartphones, but also COVID-19.
OMDS: However, consumers who love photography, the high-end amateur photographers, the demand from these consumers for cameras and lenses has not changed. So our middle- and higher-end products, their situation is better than before under Olympus. So that means our business is generally improving, and our current business is going smoothly. I obviously can’t share too many details, but our current business seems to be good, especially thanks to the OM-1.
RDE: So the OM-1 produced a big surge in sales for you?
OMDS: Yes, more new users are purchasing the OM-1, compared to the products in the past, people who maybe had used other brands before. When those people buy our OM-1, they also buy our PRO lenses as well. So the OM-1 contributed a lot to our PRO lens sales as well. So our current business is very good.
This was great to hear: Especially in the current global environment, for a camera company executive to describe their business as “very good” seems pretty remarkable. (Although the mood I heard from everyone I spoke with on this trip was generally upbeat.) It’s particularly interesting to hear that increasing numbers of OM-1 customers are new to the OM SYSTEM/Olympus platform, having migrated from other brands. Olympus had always had excellent optics, and their bodies were generally very good as well, but it seemed like they were always just missing being able to capture pro shooters en masse. It seems that the advances in speed and AF capability in the OM-1 may have finally tipped a number of shooters over into the OMDS camp.
Will OMDS be publishing financial data publicly?
RDE: That actually touches on a more general question: Under the current business structure, will OM Digital Solutions produce quarterly and annual financial statements, or does the structure under Japan Industrial Partners (JIP) not require them? (I think under JIP, it’s more like you’re a private company, there’s no general stock, so maybe not the same reporting requirements?)
OMDS: We produce annual financial statements, under the current business structure as well.
RDE: So those are publicly available somewhere?
OMDS: In accordance with any applicable government regulations, basic financial information may be published on their platforms, however as we are not a publicly listed company, financial statements are available only to our shareholders. [Any statements about a company's finances and reporting must adhere to strict rules, even if the company in question isn't publicly traded. The above is a statement given to me by OM Digital Solutions that meets those requirements, to replace a more casual reply in the conversation itself. The bottom line is that OMDS will adhere to any reporting requirements of any relevant government agencies, but as a matter of course for a privately-held company, their financial statements are only available to their shareholders.]
I read that OMDS was created to spin off an Olympus “video system”(?)
RDE: OM Digital Solutions’ Chief Technology Officer Setsuya Kataoka - mentioned in an interview with the publication Asahi back in February of 2021 that you had developed a “video system” under Olympus, and the desire to continue to pursue that was part of the reason for the spinoff. I also note that the investment is described on JIP’s website as “Carve out of the video business of Olympus Corporation.” Can you explain what “video system” means in this context, and how that will influence ongoing product plans?
OMDS: This may be a mistranslation of the content of the Japanese article when it was translated into English. Mr. Kataoka actually mentioned the “imaging business” (not "video business"), so this should be interpreted as the “camera business.”
Is there any relationship with Olympus when it comes to R&D?
RDE: The fact that we’re meeting here in Hachioji, but not in the Olympus R&D headquarters building where we always met before touches on one of my biggest questions, namely what does OM Digital Solutions’ R&D structure looks like post-transition? As I understood it, R&D for the photography and medical sides of Olympus’ business were always separate, but there was a lot of technology sharing between the groups. How is OM Digital Solutions’ R&D structured, and how would you say the headcount or general level of resources there compares to the way things were when you were the imaging division of Olympus Corporation? Do you share or lease any facilities at the Olympus R&D building?
OMDS: We are now a new company as OM Digital Solutions, and the resources and structure of an R&D organization have been altered in line with the size of our business. However, we believe that it's most important to continue to offer attractive products supported by the market, therefore we do retain the primary development resources necessary to achieve this.
OMDS: We also utilize outsourcing for some of the elements required for product development. The R&D resources owned by Olympus are one part of this outsourcing, and we work with them in that aspect as well.
RDE: I didn’t quite catch that; there’s cooperation between you and Olympus on some things?
OMDS: We still use some of the Olympus R&D resources on the basis of outsourcing.
RDE: So you contract with them for some R&D services?
OMDS: Yes, but it’s not only with them; we have other sources as well.
RDE: So you contract with them, but it’s an arms-length relationship.
OMDS: Yes, but of course, we also have our own R&D staff within our organization as well
Olympus' R&D center in Hachioji is an impressive facility that was the venue for many meetings and the source of many articles over the years. Research originating in the camera division was often cited by executives as being an important contributor to the development of their medical products as well. Given the amount of work that was done there to develop image stabilization systems, image processing algorithms and optical designs for their lenses, I wondered how R&D would be structured for OMDS post-transition. I'm glad to hear that there's still a relationship there. We don't know the details of that (I knew better than to bother asking for deeper detail ;-) but it's interesting that OMDS mentioned that Olympus was just one R&D contractor among an unspecified number of others.
OMDS of course also has their own internal R&D staff. We don't know the detailed structure, but I'm counting it as a positive that they're still able to tap into the R&D resources at Olympus on a contract basis.
It's obviously an open question how the new R&D structure will play out over time, but to my mind, the technology embodied in the OM-1 gives OMDS an excellent starting point. They led the field with their AI-based subject recognition (which is still arguably class-leading), the OM-1's 120fps capture at full resolution (or 50fps with autofocus active) is top-tier, and OM System/Olympus-developed image stabilization technology led the field for years until open-market technology finally let others catch up. This is an incredibly strong base to start from, and even incremental advances will keep OMDS competitive for years to come.
What’s the arrangement for manufacturing OMDS products?
RDE: Pre-transition, Olympus had made a huge move with their manufacturing, withdrawing from China and consolidating operations at their factory in Vietnam. (A move I’m sure they’re especially happy about now, given the current world situation :-) How is OM Digital Solutions’ manufacturing currently handled? Are your cameras and lenses still being made in that same facility? Did ownership of the production equipment transfer to the new company? Are you leasing space in the factory from Olympus Corporation? Is Olympus Corporation doing the manufacturing under contract to OM Digital Solutions? Or none of the above?
OMDS: Part of the Olympus factory in Vietnam has been transferred to us as an asset, and serves as a manufacturing site for OM Digital Solutions. Therefore, we are essentially manufacturing our own products in our own factory.
RDE: Wow, that’s a huge asset. I think Olympus manufactured other things besides cameras and lenses in that same factory, I think some of the medical manufacturing was done at that location also. So it’s your asset; are you leasing some of it back to Olympus, or working under contract with Olympus to do their manufacturing as well?
OMDS: We do have an imaging business part within that factory that is our asset, so it’s not leasing back.
RDE: Ah, so you have your part of the factory, and other parts belong to Olympus?
OMDS: Yes, other parts of the factory are still for Olympus. There’s one factory (building), but it has been divided into two parts, and we own our part of it.
RDE: Ah, I understand. As I said, that’s critical for you to have, and it was a huge move that Olympus made.
Has having your factory in Vietnam helped with supply-chain issues?
RDE: With your main factory in Vietnam, you’re shielded somewhat from the current disruption and uncertainty caused by the lockdowns in China, but I imagine that some of your supply chain is still there. To what extent has the upheaval in China affected your production ability? Are you experiencing supply-chain disruptions with parts coming from other parts of the world as well? What has been your strategy for managing supply-chain disruptions? (Changed suppliers? Carrying more parts inventory? Shifting production from lower-demand products? All of the above?)
OMDS: A part of our supply chain still remains in China. Although we have been affected to a certain extent by parts supplied from other regions of the world, we have responded by working with additional suppliers and adjusting order quantities to secure parts, as required by the situation.
RDE: Have you been able to meet consumer demand? I guess the demand increased with the OM-1, but from the consumer’s side, have they seen any product availability issues?
OMDS: We (have to say) we are very sorry to the users who have been waiting to buy the OM-1. It’s still continuing in short supply to the market. But I think starting in two months we will be able to catch up. [That would be two months from the end of July, when I spoke with OMDS, so it sounds like they could catch up as early as the end of September.]
RDE: So to be clear, some of that isn’t just supply chain issues, it sounds like demand for the OM-1 has been much higher than you projected, right?
OMDS: Yes, but we’re very sorry for the customers.
Warranty repair in Canada goes beyond what Olympus offered. What's the story there?
RDE: OM Digital recently shared news of a new warranty/repair service in Canada, which is something that I think Olympus didn’t offer. Is this new service the result of increased customer activity in Canada, a result of OMDS’s different business approach versus that of Olympus, or something else entirely?
OMDS: With the growing demand for Olympus/OM SYSTEM products, the importance of the Canadian market, and our commitment to offer outstanding service to our customers, we felt it was time to add another layer of support service for our customers in Canada.
OMDS: Therefore, we have partnered with Sun Camera Ltd. to offer a Canadian Authorized Repair Center. Sun Camera Ltd. has an excellent reputation, having nearly 50 years of experience and professionalism in the field of photo, video and digital product maintenance and repair. The service center will provide full-service and warranty repairs for Olympus and OM SYSTEM branded products, allowing us to service the needs of our Canadian-based customers quickly and seamlessly.
How will products be branded going forward?
RDE: Can you speak to the overall brand transition from Olympus to OM System? For example, the recent OM-1 camera still includes ‘Olympus’ branding on the front. It’s been stated that this is a celebration of the 50th anniversary of the original OM-1 35mm camera, but some might say it further muddies the overall brand transition. Is the OM-1 the last OM System product we’ll see with any Olympus branding on it? (Will the branding change on existing lenses that were released under Olympus?) What will the brand transition look like from here on out, and how will you manage communications to preserve the awareness and brand loyalty that Olympus enjoyed for so many years?
OMDS: The Olympus logo featured on the OM-1 was not for the 50th anniversary of the OM-1 film camera, but to symbolize the fusion of the innovative new technology offered by the OM SYSTEM with the DNA of the brand and development philosophy inherited from Olympus.
The OM-1 will be the last camera with the Olympus brand logo. Although we are currently considering brand changes for existing products such as lenses, we cannot provide any details at this moment.
OMDS: The Olympus brand name has been built through years of manufacturing and communication in the market, therefore we feel that from the standpoint of name recognition a simple shift to a new name would be difficult. However, as you know that the OM SYSTEM brand does not deny its Olympus origins, and actually it continues the Olympus philosophy and passion for craftsmanship, while seeking to offer more new values. That's why, in terms of maintaining and growing brand loyalty, we believe that the most important method of communication is to reliably continue to provide attractive products and CRM measures on an ongoing basis.
OMDS: We believe that our camera systems provide unique values, by offering easy handheld shooting and computational photography that allow users to expand their photographic abilities. We hope to convey the fact that our systems are optimal for a broad range of shooting scenarios both indoors and outdoors.
This was as expected; OMDS is a new company, distinct from Olympus, so of course they need to change the branding of the products. I wasn't sure whether it might be a more gradual transition or if the OM-1 would indeed be the last camera to bear the Olympus name, but it's no surprise that that is in fact the case. It's a little sad to see the Olympus brand disappear from the photo market after so many decades, and it probably presents at least somewhat of a challenge for OMDS to build mind-share for their new brand in the marketplace.
That said, as OMDS themselves alluded to, the Olympus name and its connection to OMDS are very well known: It's hard to imagine any prospective customers for the system not already being familiar with the backstory, the long history of the Olympus brand and the connection between it and the new "OM System" branding. It's often challenging for more generic consumer-goods companies to make major branding chances, but I don't see OMDS's target customers (enthusiast and professional photographers) having any difficulty making the connection between new OMDS products and the long Olympus heritage.
Of course, this is a completely moot point in the present moment, since the OM-1 is dual-branded with both the Olympus and OM System logos, and all the other cameras and lenses currently in the market still bear the Olympus logo. I did ask whether or not any of the existing products (especially lenses, which typically have a much longer life cycle than camera bodies) might have the Olympus name removed and be rebranded exclusively as OM System products, but the answer was noncommittal. Reading between the lines, it sounds like there aren't any current plans to do so, but it may or may not happen in the future.
What sort of lenses will the coming tele zooms be?
RDE: On the last lens roadmap from September 2021, there are two telephoto zoom lenses listed that I don’t believe we’ve seen yet. Can you talk about these lenses at all? Will they be aimed at novices or pros?
OMDS: Unfortunately we cannot go into any detail on these lenses at this moment, but we believe that these lenses will please the photography enthusiast.
RDE: Of course you can’t say anything specific, but I can at least tell our readers that the coming tele zooms will be enthusiast-focused. As you said before, the enthusiast market has continued throughout all the ups and downs of the industry. New people come and use cell phones or they’re vloggers, but that core community of people who are passionate about photography is still there and continues, so this is aimed at that group. I tend to agree that that’s where OMDS’s true core customers are.
OMDS: According to our company's surveys, it shows that the market volume of pro photographers and high-end amateurs has not changed. We believe that in the future, this market volume will not change significantly. So we are focusing on the mid- to high-range. That is our strategy.
This was a recurring theme in my conversation with OMDS: They recognize that their core user base are enthusiast photographers, and that’s where their focus will be going forward. I take this as very good news indeed.
Are there any more lenses coming in the f/1.2 line?
RDE: Since announcing a trio of F1.2 prime lenses in 2017, there haven’t been any new F1.2 PRO primes. Is this because the trio of 17, 25 and 45mm lenses satisfies the needs of your users, or are there particular challenges in developing more F1.2 prime lenses for Micro Four Thirds?
OMDS: As far as the F1.2 series goes, we understand that there have been requests for F1.2 lenses with other focal lengths. So we have studied what lenses we can or can’t provide. If the focal length were wider or longer than the current ones, the lenses would be bigger, and consumers might not be as willing to accept them. Consumers have told us that they want a compact-sized lightweight camera system. So expanding that series isn’t as high a priority, because longer or wider lenses would be larger than the current ones.
RDE: Ah, so they wouldn’t fit with the concept of the system. The consumers are saying “I want an 85/1.2 but only this big. <guesturing, showing a tiny lens> <laughter>
What’s behind the backlog for the 150-400mm?
(See below for some additional eye-candy from the display shown above :-)
RDE: The M.ZUIKO ED 150-400mm F4.5 TC1.25x IS PRO is an amazing lens; we were pretty blown away by its image quality, capability and overall performance when we reviewed it back in late 2020. It’s a very complex design, with no less than 28 elements in 18 groups and a *lot* of very high-spec glass in it, including an aspherically-molded ED optic (you refer to it as an “EDA” lens). While I don’t think there’s been such an announcement in the US market, I saw that back on June 23, you announced that you were pausing the acceptance of new orders in the Japanese domestic market. Has the complex design proven difficult to manufacture, or are there other factors at play? How do you think you’ll overcome the issues, and do you have an approximate idea of when production might match demand?
OMDS: As you say, we have recently announced that we have temporarily paused the acceptance of new orders for the 150-400mm PRO lens in Japan. There are two reasons for this.
OMDS: Firstly, production is managed to strict quality control standards, using advanced production and assembly technologies, and only a limited number can be made each day.
OMDS: Secondly, since the release of the OM-1, we have seen an increase in the number of new customers wanting to use our systems, in addition to existing customers. This can be attributed to the increase in orders for this product, particularly among enthusiasts in wild bird photography. Especially for wild bird photography, users have a very strong interest in the OM-1, and many of those people also want the 150-400, so the order quantity is higher than our expectations. Please note that this temporary pause in acceptance of new orders is specific to Japan only. It is not the case in the North American region.
RDE: Well, I guess that’s a good problem to have, that the demand has been so high. And that’s not an inexpensive lens, it’s a major investment, so it speaks well that there are so many people who want to get it. (I can really see that especially for bird or wildlife photography, that lens paired with the OM-1 would be an incredible setup.)
Here again, it seems that the OM-1 and possibly the specific combo of OM-1 and the 150-400mm are drawing new users to the OM System platform. OM System's combination of compact size and weight, especially at longer focal lengths, great glass, great IS and excellent weather resistance had already made birders a strong market for them. The OM-1’s high frame rate and capable AI-based autofocus for birds in flight, and the exceptional reach of the 150-400mm (including its built-in teleconverter, it has a 35mm-equivalent focal length range of 300 - 1,000mm in a very compact package) made the platform all the more compelling, bringing a significant number of new shooters into the OM System fold.
How are you using the stacked-CMOS technology in the OM-1?
DE: The OM-1 has the first stacked-CMOS sensor in a Micro Four Thirds camera, contributing to its blazing speed. I’m curious how the readout processing and storage are distributed across the layers of silicon, so have several questions related to this:
I asked three sub-questions here:
Does the stacked chip have any buffer memory on it?
As the camera OEM, can you choose what logic or memory is on the stacked chip?
Is there image-processing circuitry on the stacked chip?
Can some low-level parts of focus determination be done on the stacked chip? (eg, cross-correlation processing) How about video resampling?
Unfortunately, I got the following answer to all of these questions:
OMDS: Unfortunately, we cannot provide any answers regarding the processing details of our Stacked BSI Live MOS sensor. Thank you for your understanding. (Although they did confirm that they’re able to choose what exactly to put on the stacked chip.)
This was all of course a little disappointing, but I guess I shouldn’t have been surprised by the lack of answers. I asked a similar question of Fujifilm when I spoke with them, and learned that in their case, they’re just using the second chip to hold the A/D converters and high-speed data channels to carry the image data to the processor.
The big new learning for me in this trip was that the sensor customers can indeed put any kind of circuitry on the stacked chip that they want to.
One obvious thing to do is what Fuji did, namely using the second chip to get data off the sensor chip and into the main processor much faster. Another possible use might be to add buffer memory directly behind the sensor, allowing some number of frames to be captured even more quickly than if the data had to be sent to the processor first. Or, as I suggested, how about resampling video on the fly? What if you didn’t have to clock the data out to the processor? Could that permit even higher video frame rates than we’re seeing currently (especially for things like 4K or FHD from higher-resolution sensors)?
Given that the system designers can put anything they want on the logic chip makes me wonder if it could lead to even more impressive autofocus operation in the future. Could it make a difference if some low-level elements of both autofocus and AI subject detection could be executed directly on the sensor stack?
It might turn out not to be practical for a variety of reasons, but both conventional and AI-based AF involve a lot of low-level processing as input to higher-level algorithms. In the case of phase-detect autofocus, a process called correlation takes sets of data from two sets of focus pixels and “slides” them past each other, to find how much they’re offset from each other. (The amount of offset is basically the focus error.)
AI subject-recognition algorithms take as their input “features” extracted from the raw image data. Typical features are things that look like edges or lines in the subject, and their orientation. Other features could be things like small dark objects against lighter backgrounds (think: eyes), oblong objects of various sizes against different-colored backgrounds (think: faces), etc. These features are identified by so-called convolution operators. There’s not the time or space here to discuss this more deeply, but convolution involves millions of multiply-add calculations, operating on data all across the image frame. Could we implement more sophisticated AI algorithms by offloading some of this preliminary calculation to processing circuitry right behind the sensor?
All of this is pure blue-sky speculation on my part, and there may be very good reasons why you can’t or wouldn’t want to do any of the above. (As just one example, the semiconductor fabrication processes you’d use to make highly efficient processor chips would be prohibitively expensive to use for chips the size of image sensors. You’d probably need to use a lower-spec process for any logic on the stacked chip, which might entirely erase any advantage gained by putting the processing closer to the sensor.) Still, it’s interesting to speculate what the tech might bring us in the future. I’d give my right arm (or at least a finger or two) to know exactly how the camera manufacturers are using stacked-CMOS technology now, and what they’re working on that we’ll see 3-5 years from now…
With cross-quad AF technology, are all of the OM-1’s pixels focus pixels?
RDE: With the OM-1’s cross-quad autofocus technology, all pixels on the sensor were cross-point AF pixels, meaning they are sensitive to distance information both horizontally and vertically. I wanted to confirm first that indeed *all* the sensor pixels are PDAF pixels, or is it just some of them?
OMDS: Yes, your understanding is correct. Phase detection information can be acquired at all pixels, so AF is possible regardless of where the subject is in the image.
RDE: So that means you can basically form phase-detect AF points with their centers anywhere in the frame, or with any baseline you’d like?
Phase-detect AF involves looking at two lines of pixels separated from each other either horizontally or vertically along their common axis. At very wide apertures, you can get better AF accuracy by having a longer string of focus pixels with a wider span. On the other hand, small apertures need greater resolution across a shorter baseline. The fact that *every* pixel on the array is a focus pixel means that the processor can create an AF point anywhere in the image just by choosing which pixels it wants to use to focus with. In the old days of SLRs, the AF pixels were on a separate sensor, with a specific, fixed pattern of pixels on it, so the camera could only measure the distance to a subject that happened to be on top of an AF point. AF tracking algorithms are another whole (and deep) subject on their own, but a mirrorless system with every-pixel AF like the OM-1 has at least the theoretical flexibility to create AF points that follow a subject around the frame, always knowing its exact distance without having to predict where it will end up as it moves from one fixed AF point to another.
Can the OM-1’s AF system run faster than 120fps?
RDE: The OM-1 can shoot full-resolution images at 120 frames/second, and the EVF operates at 120fps as well. Is that also the rate at which the AF system “looks” at the subject, or can the AF processing run even faster?
OMDS: We cannot give you specific numbers, but the AF is as fast or faster than the EVF display.
Can the full quad-pixel PDAF data be read out for other uses?
RDE: Could it be possible to capture the sub-pixel information that makes up the quad-pixel AF system in the RAW format? Is the sub-pixel data digitized, or are the sub-pixels only read out on a selective basis during AF cycles? - If it could, it might allow post-capture processing for DOF enhancement, with very high spatial resolution…
OMDS: Individual sub-pixel data is only utilized in AF processing, therefore data is not provided as raw data.
OMDS: We are evaluating ways to put sub-pixel data to use, such as achieving photographic expression that was not possible to realize before.
RDE: So you’re studying or considering how to save that information to allow that processing?
OMDS: Yes, we are studying how we can utilize it in various ways.
RDE: Yeah, I think that information would be useful for depth-of-field processing and other things.
How has your AI AF progressed, beyond the addition of new subject types?
RDE: A hallmark of the OM-1 is its AI-based autofocus technology, inherited from the E-M1X and expanded to include more subject types. (Originally birds, trains, cars and motorcycles, now with dogs and cats added) In addition to adding the new subject types, in what ways has the performance of the AI autofocus technology improved?
OMDS: A new image processing engine allows faster subject detection. Combined with the new image sensor, this has improved AF tracking performance.
RDE: So you’re able to initially find the subject much quicker than you used to be able to?
OMDS: Yes.
RDE: And of course, now you’re reading the data out so fast, you can track it better.
Does the OM-1 have dedicated neural net hardware?
RDE: When the E-M1X was announced, you used a general-purpose processor for the AF processing because the algorithms were changing so much that you didn’t want to lock-in to custom neural-net processing hardware at that time. Has your neural-net architecture continued to evolve since then?
OMDS: Although the E-M1X featured a DSP (Digital Signal Processor) or a CPU within the engine to achieve what was deemed to be optimal neural-net processing at the time, the OM-1 utilizes a custom processor for more advanced neural-net processing based on the achievements of the E-M1X, delivering high-performance, high-speed processing, as well as lower power consumption.
Will it be possible to update the AIAF algorithms via firmware updates?
RDE: This wasn’t in my original questions, so no problem if you’re not able to answer it, but it occurred to me to wonder about the ability to update the AI algorithms. At the time of the E-M1X’s introduction, I remember talking with an autofocus engineer, asking if it would be possible to update the AIAF algorithms via firmware to add new subject types or that sort of thing. With the current OM-1 architecture, would it be possible in the future to increase the number of subjects, the types of subjects it recognizes, or to update the algorithms, or are they more at the hardware than the firmware level?
OMDS: The update could be done at the firmware level. We cannot provide any details, but we are considering the evolution of AIAF based on market conditions and customer feedback.
What’s behind the OM-1’s improved AF performance?
RDE: How much of the improved AF performance is a result of improved hardware, and how much is the result of changes to the algorithms?
OMDS: It is difficult to offer a percentage because it changes depending on the situation in question, but improvements in both the hardware and algorithms used have helped improve AF performance.
What’s coming next with AI?
RDE: What’s next for AI in your cameras? How about enhanced shallow DOF, a la the iPhone’s portrait mode? That could make a lot of sense, to help Micro Four Thirds compete more with full-frame cameras in this area.
OMDS: We hope to leverage further advances in computational photography to allow high-quality images and greater photographic expression without relying on sensor size, so that customers who own full-frame cameras can also enjoy our camera systems.
RDE: So if I’m interpreting that right, you’re saying that one of your hopes is that by further developing computational photography, your cameras will appear as an attractive option even to people who are currently full-frame users.
OMDS: Yes
What’s the story with AI Detection AF (AIAF), Tracking AF (TR-AF), Continuous AF, Target AF, etc?
RDE: Finally, a rather detailed question about autofocus operation: What’s up with AF tracking and continuous AF mode, vs using Target AF in continuous AF mode. It seems that the combination of AF tracking and continuous AF doesn’t work well, to the point that you advise against it. Could you explain a bit about how the various modes work, and what it is about continuous AF mode that interferes with AF tracking?
OMDS: We recommend using Target AF in C-AF when shooting with AI Detection AF. Movements of the subject detected are predicted, allowing AF with high level of tracking performance to be achieved.
OMDS: When tracking is used, tracking the primary subject with the integration of data such as color and total screen vector detection other than subject detection is given priority, so subject detection is handled as part of the information to be tracked, and thus other detection results may be given higher priority than detection of the subject alone.
When tracking, if the subject is lost, AF stops, and is restarted when the subject is detected again, or the button is half-pushed once more. When operating with subject detection alone, operation switches to AF operation in zone settings when a subject is not detected.
OMDS: If you try to shoot the subject, for example a cat or a bird, we recommend Target AF and CAF -- not tracking AF -- with AI subject detection. On the other hand, if you tried some other subject [not an AI subject, just a moving object], we recommend C-AF + TR-AF. Because the AF process is different between C-AF and Tracking AF, so if you use AI Detection AF or AIAF, it’s working only with the information from AI detection, but if you use Tracking AF, the camera AF works not only detection AF, but other information, for example, color information or movement direction and so on. Sometimes, only the AI detection process is better than several AF processes with several pieces of information. This is the reason if you turn on the AI Detection AF or AIAF, we recommend C-AF, but if you tried to shoot a moving subject and its movement is unpredictable, we recommend Tracking AF.
RDE: So I think I understand. The AI detection, the neural network is looking at the image, and all of the colors, the shapes, where the eyes are, that sort of thing. But it’s not relying on the PDAF information to find the subject, it’s just looking at the picture like a human would. On the other hand, when you’re Tracking AF, it’s looking at distance over time and how it’s moving, but not at what the information is like. Although you said that Tracking AF also uses color and shape. So… let me try to boil that down.
RDE: So AIAF is really looking to find the subject just based on what the image looks like to it. And once it identifies the subject, then it will use PDAF just in that area (I think) to focus. Whereas tracking AF, it’s looking for general moving shapes, it’s a blob of red and it’s moving based on distance information and so on. So really they’re just two different things, and trying to combine them doesn’t work well. [Actually, they don’t combine at all, they’re two entirely different systems.]
My summary of how AIAF and Tracking AF work…
There was a lot of back and forth in the above, so I've made this summary a separate subhead here. This is my take on the key points about how AIAF and Tracking AF work, and why you can't combine the two:
-
AIAF and Tracking AF are two fundamentally different and separate systems.
-
AIAF looks at the scene in front of the camera the way a human would, identifying subjects by their appearance. Once a subject has been identified, the camera uses the PDAF pixels in that specific area to set focus.
-
Tracking AF is the conventional AF that we’ve long been familiar with: It uses a combination of distance (from the PDAF pixels), color and shape to identify the subject (or to follow one that you’ve told it to via the user interface), then uses its movement over time to predict its likely future position. The color and shape help it avoid being confused by other objects in the scene as the subject moves around, but there isn’t the sort of AI-based “intelligence” to recognize an object as a specific type of subject.
- If you’re doing continuous shooting with AIAF, the camera is basically re-identifying the subject in each frame and then focusing on it. It doesn’t make any predictions about the subject's future position based on past behavior.
The two processes are quite distinct. Here's a narrative description of how I think they each proceed:
-
AIAF’s subject tracking frame by frame is like:
- “Ah, there’s a bird, focus, click
- (next frame) “Ah, there’s a bird. Focus, click”
-
etc, etc...
-
Tracking AF is like:
- Ok, the human told me to focus on this thing here. It’s a pink, roundish blob about so big, and it’s this far away from me right now. Focus, click.
- Ok, that blob is a bit more to the right now and 2 feet closer to me; Focus, click.”
- All right, between the previous two frames, the blob moved closer to me by two feet and a bit to the right, so I’m expecting it to show up a bit more to the right and another 2 feet closer to me.
- Shift focus by 2 feet. (This can happen ahead of time, before it's time to grab the next frame)
- Look at the scene. Ah, sure enough, there’s that blob right where I expected it to be, but it’s only about 1.8 feet closer to me this time. I’ll make a note of that for next time...
- Focus, click.”
- and so on…
Hopefully that makes it all a bit more clear. AIAF and Tracking AF are two fundamentally different processes, and the camera is either in one mode or the other. In the future, we can hope for the two to be more closely integrated, but it's understandably complex to combine the two separate flows of information. It will require the addition of a whole new higher level of processing, looking at and evaluating the AI-based and conventional distance/shape/color-based tracking information. Ultimately (probably in some distant future), AI could handle the whole job, but the challenge will be coming up with the human-interpreted raw data to use to train the algorithms. It's one thing to manually label tens of thousands of still images to train the AI system with, but creating labeled video streams that also incorporate full PDAF data will be another matter entirely.
Personally, I suspect the integration of distance information into AI-based AF algorithms is something that the R&D departments of multiple camera companies are probably looking at right now. As incredibly capable as modern AF systems are, this is still the area where there's the most room for future improvement. That said though, the speed, accuracy and intelligence of current AF systems have reached levels film-era photographers could only have dreamt of.
RDE: I think that’s actually all the questions I had. You answered them very quickly! Thank you!
OMDS: Thank you as well!
Summary
This was a great meeting, and it was good to see OMDS doing so well following the spinoff. I can't imagine the sheer magnitude of the task executing the spinoff must have been, involving a complete restructuring of operations not only of the engineering and manufacturing components (in Japan and Vietnam, respectively), but of the many national and regional sales offices around the world. From their comments, this was further complicated by the difficulty of bringing up a brand new IT infrastructure from scratch. Fortunately, it seems that the most difficult parts of the transition have been completed, and sales are apparently stronger than ever.
Like many of us, I'd viewed the spinoff with cautious optimism, but also a good bit of concern, especially as the world plunged into the COVID era. Would OMDS be able to maintain the necessary levels of customer service around the world? Would they be able to efficiently manufacture and ship product? What would their technology and R&D look like going forward?
This meeting (as well as having seen for myself the ongoing strength of the Olympus/OMDS brand in the US) allayed a lot of my concerns; they give every appearance of having successfully navigated the changeover, and their sales are better than they've ever been.
Especially noteworthy is that OMDS is seeing a lot of pros switching to their platform from others, thanks to the exceptional performance of the OM-1 and the compactness of their entire system. (Some full-frame bodies are very compact these days, but when you assemble a full kit of lenses of equivalent focal lengths on both platforms, the size and weight difference is dramatic.) Also supporting this transition is the incredible 150-400mm f/4.5 supertele zoom with its built-in 1.25x teleconverter. It's an exceptionally sharp, well-performing optic, and the combination of it and the OM-1 is an unbeatable solution for bird and wildlife photographers, not to mention sports shooters looking for extra reach without breaking their arms.
- And that touches on another important point: OMDS understands who their customer base is (namely enthusiasts and pros, especially those for whom compact, lightweight and very weather-resistant gear is important) and is laser-focused on them. I'm sure they'll always have lower-end models like the E-M10 to serve as inexpensive entry points to their system, but they're not going to waste resources trying to compete at the bottom end of the market.
The long-term question of technology and R&D is one that will take many years to fully answer, but I was encouraged that some of the technical prowess of Olympus Corporation are still available to the new company via an arms-length contractual relationship, in addition to their own internal R&D staff.
Thanks to the tech embodied in the OM-1, the new company is also starting out from a very strong position. While the rest of the world has finally caught up to their IS capability, OMDS's IS is second to none, and arguably as good as 99.9% of people need it to be. At 7.5 stops of stabilization, you're literally getting to the point where the system is sensitive to the rotation of the Earth. I honestly don't think we're going to see 8.5 or 9 stop IS, at least not in any conventional sense. (I do think there's some room for advances through computational photography, but we're also talking about a small segment of the market who'd need to handhold 5 second exposures.)
On the AF front, Olympus was a strong, early leader in AI-based subject detection; it's hard to believe that the E-M1X was introduced over 3 1/2 years ago now. Its AI-based subject detection was revolutionary at the time for the range of different subjects it not only could identify, but also pick the best part of the subject to focus on. The OM-1has built on that base with more subject types and higher performance. I see OMDS as being very well-equipped to innovate and compete deep learning AI technology becomes ever more important.
The backside-illuminated stacked-CMOS sensor in the OM-1 also bodes well for future development, in that OMDS now has direct experience with stacked-CMOS technology and the very high data transfer rates it provides.
Overall, I'm feeling more confident about OMDS's position in the market now than I was about Olympus's shortly before the spinoff. Their technology is even better, their sales are strong, and they have a much leaner, more cost-efficient organization than Olympus ever did. I obviously don't have any crystal ball, but OMDS looks to me to be very well set up for the future.
What do you think? Leave your comments and questions below, and I'll try to respond in at least a semi-timely fashion :-)
Extra Bonus Content: 150-400mm Eye Candy!
After my interview with the others, we were joined by Kyoji Murayama (Marketing Strategy, Global Marketing), who walked me through some of the components of the 150-400mm lens, arrayed in a display case outside the conference room. The OMDS 150-400mm is a veritable tour de force of optical technology, delivering really exceptional performance in a surprisingly compact and lightweight form factor. (There's a *lot* of technology packed into this lens, read my Deep Dive on the OM Digital Solutions 150-400mm Super-Tele Zoom article for lots of inner details.)
I know a lot of our readers are every bit as geeky as I am, so I snapped some photos of the display to include here as "bonus content", along with comments on parts I found interesting. Enjoy!
The level of precision in the die-castings is pretty incredible. Judging from the surface finish, I'm pretty sure that the fine threads above are as-cast, not having needed any subsequent machining to achieve the needed tolerances.
That's it, I hope you enjoyed this little eye-candy bonus! Feel free to ask any questions you might have below and I'll try to answer them.