Checking the landscape files from my 'new' e-m1 mk II I am letting go of my 24mpx A7 mk 1 full frame. The difference is hardly obvious even when pixel peeking and that's in standard non-hires mode. First micro 4/3 cam that can do it all for me, and if I want to go superlight I grab my Panasonic GM1. That one is to stay with me too. So downsized now completely to micro 4/3, keeping one system only.

The phase detect with e-m1 mk ii in video is surprisingly good too.

Re: "Canon (and all other large sensor brands) make higher resolution models. " If you want to talk resolution, 20Mpix M43 has the "resolution" (pixel density) of 80MPix FF, just cropped. So where are the 80Mpix Canon (and all other large sensor brands)?

the TG series did go backwards in mp count. Still the best tough camera.

There is not any big issue there, like already said, you just need to select whether to use ibis or is in lens. The systems just don't cooperate like on its own brand. They can possibly iron this out too by sharing some data you know.

@Kona Mike - It's not that big of a deal, you lose or gain a stop depending on which brand combo you use.

Looking at the GH6 I think it's perfectly OK to differentiate between video-centric and stills-centric cameras. I don't feel like I need all the lenses to work with all the bodies. I'm a sports-action guy who shoots an occasional landscape. I have everything I need in an OMS lens so I don't; care if the Panny lenses work at all with an OMD.

Kona,

You are absolutely correct. And in fact, when the Panasonic GX-8 debuted eight years ago, the DPReview was openly critical of its departure from the M43rd's Consortium standard with its new "Dual-IS" stabilization. (Since their original review, DPR -- in its typical mealy-mouthed fashion -- excised all its original critical language.)

Oh great - then OM can easily hire a two man team and beat Nikon and Sony at AF algorithms. I wonder why it didn't happen?

Smart-**s brownie with a snarky post. I spent 23 years in the software business, a lot of it with startup companies in Silicon Valley. You are entitled to your opinion and have the right to post it. You don't know anything about it.

On the contrary - I do. Researchers can make breakthroughs in small teams - might even be better. To make a competent product requires more than just lines of smart code. It takes a team to make a product that makes sense. And as I said before - it takes the right people who have experience in this field.

You said smaller companies can't compete and will be left further behind when you started this thread. I know what I read. It's up there if you want to read it. You contradicted yourself. Then you implied you can't hire a small team and compete with C&N, made a snarky remark about it. Small companies are the innovators in most every industry. That's how they compete. They have to offer something newer, better, or for less money by innovation or there is no reason to buy their products. OLY was the most innovative camera company for many years, and one of the smallest, dwarfed by C&N. We did not get new tech from Nikon and little from Canon. We got FF mirrorless from SONY. That's the way they broke the C&N FF duopoly.

If small companies are the innovators then OM should demolish Sony and Nikon with autofocus performance. Can you explain why they are not doing this?

Is beating them at AF your criteria or their criteria? You can't beat perfection. What makes you think they tried? OLY dropped PDAF with the PEN & EM5. Tracking & CAF was not their interest at the time. They have to play catchup since the EM1.1 brought a PDAF sensor back in the system. They are catching up. If you think OLY was trying to make world-class CAF all this time that's speculation. If beating S/N/C is the only measure of camera superiority, that's your opinion. Why can't they beat OLY at IBIS? Why did OLY have it years before any of them did in MILCs? Why did OLY have in-camera focus stacking first? Sonic wave sensor cleaning first? Why does it have starry night? HDR first? Tilting LCD first? Fully articulating LCD first? Mirrorless ILCs years sooner? Live bulb? In-camera NDfiltering? HiRes tripod & hand held? Focus assist? Manual focus assist? PRO-Capture. I'm may be forgetting a few. Why didn't S/N/C have all these things before OLY? Because they aren't innovators.

But - it should be easy for OM to beat scn right? They are a smaller company where you say "the best innovation occurs". Unless we are talking about video - AF systems are one of the most important things about cameras like this. If they can't excel at the most important thing - then they have their priorities all wrong.

Priorities for which customer? Not for landscape, street or portrait photographers. On further discussion with the semi-expert, sensor read-out is fixed. Some posters said they think it reads out at half the speed they think it can/should. It seems like this impacts AF. The faster the sensor the better the camera can track accurately. Focus has to change quickly to stay sharp & on the subject. M43 has always been hardware handicapped because it's thought to be inferior to larger sensors. It attracted buyers who expect to pay ~ $1,200 for an OMD, $700 for a PEN. People will pay up to $6,500 for a C/N/S body. With $2,400-$6,500 retail, you can use a faster sensor & memory & the AF system can be faster & more accurate. I'm speculating on top of speculators & I'm no semiconductor engineer. Seems to me customers not the number of OMS engineers holds OMS back because they won't pay enough for hardware. Not a software resource problem. OLY beats them at software.

Then why doesn't OM just innovate there way out of the problem? They are a small company - the best kind according to you. They should have a big advantage over SCN. Yet - we see big white lenses on NFL sidelines - not Olympus bodies. Somehow SCN continuously beats them in almost every way. The one thing OM does better is stabilization. That is pretty much it. I am sure SCN could do sensor shift if they wanted to - but who needs a 200Mb image? OM does it because their sensors are, relatively - low resolution. If small companies are the best innovators - OM (formerly Olympus) must not be a very good one.

A matter of opinion. Canon and now SONY throw money at the professional sports camera business. It's supported by funds from their marketing department. They are buying the business. They have support programs, instant on-site loaner gear at big events, funded by large consumer camera businesses. The consumer gear buyers pay for it with their purchases. The AP switched from C to S. Because C has an inferior product? NO WAY. Canon's cameras and lenses are GREAT. Can pros make good photos with M43 sensor cameras? Of course they can. Look at the photos OMS pros take with them. As good as any. You keep saying SNC is better. I don't think they are. I can buy any gear I want. I have zero desire for a larger sensor camera or better photos. I need better eyes. I'm of the belief that all cameras will get so much better, the worst camera you can buy will be so good anything better is wasted. It will be beyond your eyesight. For me - it's already here.

Please don't make me use more resolution/bigger files. They are already painful and the IQ is good enough. I'll use sensor shift when I need to. 16MP is enough. Even 12MP is when you don't have to crop.

EZGritz - you are deflecting. You said that small companies are where the best innovation occurs in every industry. I don't think so. SCN are killing it in AF performance. OM isn't. The OM-1 is just (from the reviews I have read - and here on this site) slightly better than the OM-D 1 iii. How could this be - if SCN have made huge advances during that same time period. And - I agree - sensor size isn't that big of a deal any more. If OM could bring an AF system like SNC in this small body - photographers would flock to the OM system. But - OM didn't bring that AF system. It is literally the only thing they had to do to sell their system. Bring that fantastic SNC AF system and BOOM - done. They win.

AF is only one measurement and it is not innovation. Watch OMS improve AF+Tracking. It may take a faster sensor and faster and more memory. Parts cost will be higher. Customers may have to pay more but they have improved steadily and they can get there. SNC had a big head start when OLY dropped PDAF for a while. They have to catch back up. There is no technical reason I can find from talking to associates in the semiconductor business that prevents them from doing it.

I'll throw out a theory to consider. SNC are AF perfection. You can't get better than perfection. You can improve on imperfection. O can improve their AF but SNC cannot. Therefore O can catch SNC and when they do they will as you said, "Flock to M43". It's just a matter of time when parts that are used in an M43 camera will be inexpensive enough, small enough, and fast enough to match the AF of SNC because the camera needs to be able to adjust AF to erratic subjects as fast as SNC does to track subjects without losing them. It isn't quite there yet. If a software update doesn't do it the next-generation hardware or maybe an M2X with dual image processors will, but SNC will not improve because you can't improve on perfection. I think maybe, the hardware isn't quite there yet in a low volume M43 camera in the $2K price range but the architecture is. O isn't all that far back from what I see in the videos. It's gaining on SNC. It will get there. Then when the gap is erased...

Well - as you said - AF isn't everything. SNC are riding very high right now on some very good bodies and very good AF systems. If OM had the 'perfection' in AF right now - they might be able to steal some business from SNC. They don't. And the longer they take to get there - the more entrenched SNC will get with people buying lenses and accessories for those other mounts. We will see if OM ever gets there, if they can take any business from SNC. Because all of that time that OM is throwing resources at perfecting AF - SNC will be throwing an even larger amount of resources at some other aspect of the camera that will leave OM even further behind.

It's just a theory. OMS has many things SNC doesn't & photographers who use OMS bodies think are valuable. Who is further behind? What aspect of camera will leave OM further behind? Name them. OMS is a little behind in AF in a pre-production system but it's perfectly usable for photographers who don't need tracking AND can't use the subject ID they have now. Chris's video demonstrates it's pretty darn good if you are tracking a plane, car, small animal, human, bird, and it's a pre-production camera test. How far behind was OLY with the EM-1.1? EM-1.2? They are catching up generation by generation and do they need to? For people who buy spec sheets, not for photographers and not everybody pans moving subjects. Few of the current generation do. Continued...

I for one, never had a problem with photos I took with a PM2 or an EM5, even panning moving subjects with CDAF. When OMS AF catches up they will truly have the best system for most photographers. I'm encouraged by what I see. It's already the best for me and has been since I bought an EM5. I'm perfectly satisfied now. I'd just like them to keep making cameras and lenses. I don't need a camera that's better than SNC. I only need a camera that works for me. The EM1.2 already does that. The EM1.3 has two things I'd like. ND filtering and HHHR but I haven't jumped on it because neither is critical. OM-1 does some things better and faster but the EM1.2 already does them well enough. I might just stick with the EM1.2 and go for more f/4 and f/1.4 PRO lenses. I have no interest in a camera with a larger sensor. I don't feel like I'm better off with the one I have.

For me - aps-c is the right balance of equipment size to IQ ratio. Smaller gets noisy fast. Larger and the lenses get big, heavy and expensive.

For lots of people but it doesn't seem to be as well supported as FF for vendors that have both. Fuji is doing a good job. But not behind SNC? I borrowed a Fuji kit for a few days. I loved the body except for placement of a button, forgot which one it was. I was constantly ruining shots by accidentally hitting it. Probably a different body would fix it. I tried very hard but could not make a better photo with it. The ASP-C sensor is a stop faster but I don't shoot much in low light & I have lots of F/1.8 & 1.2 primes & f/2.8 zooms for that. The Fuji lenses are good but they don't have enough of them & they are bigger & heavier. You are reading out of an old book on noise. I use ISO3200 without a thought & can get away with ISO6400 when I need to before using & third-party NR software. Look at the OM-1 photos. ISO6400 looks great & ISO12,500 usable. A few posts above you said you don't need a big senser. Do you or don't you? Not much difference between M43 & ASP-C.

Lenses. Fuji needs more of them. NOBODY has anything like the OLY 40-150 f/2/8 PRO and the TCs. FF Range = 80-300+ 420f/4 and 600/f.5.6. IQ holds up great with the TCs. This lens alone is enough to keep me in the system. Covers the range of two lenses in any other system and a 2lbs, not real heavy. Can't wait to buy the 40-150 f/4 PRO. F/4 is a sweet spot for me. DOF is thin enough to isolate the subject by positioning. I'm usually stopping down to f/5.6 or f/8 to get everything I want in focus anyway. Thin DOF is overrated unless it's a shooting style you use a lot. I don't.

Yes - lenses is obviously where OM has an advantage on most other systems. It is really the only reason to prefer OM over other systems. I admit OM has better zoom lenses than Fuji. Better in the sense that they pack more FL into a smaller package. And don't get me wrong - that is very important. I thought long and hard about the OM1iii before commiting to Fuji system. In the end the noise difference does matter to me.

Quote: "Seems a little odd that they give it a good initial review then do a second video talking about how they failed." ... Pressure from unhappy sponsor(s) seeing this as potentially affecting their market share comes to mind. Who knows, I don't know if they are sponsored by the big 3, but I agree that this is rather odd, hence why I questionned earlier what their motivation was. Nobody does anything for no reason.

So both of you are 2% ers. You cannot sell cameras based on what 2% of the markets demands.

You have a better one?

Any guess is better than your 'statistics'.

I guess that is an admission of having nothing.

I have what we could call an educated guess. But contrary to what you want to suggest there, even 'nothing' is better than 'something' which is obviously wrong and could also be described as 'lying with statistics', something propagandists of any provenience have been making an art of since the invention of numbers.

How do u know my educated guess is not better then yours? Do you have some sort of certification? Let’s face it. I have data. You do not.

I explained that already, I have nothing to add there. The characteristics of the numbers your "educated guess" is based on is similar as a presidential election prognosis would be that was based on the answers of a sample comprised of the family members of one of the candidates. That also would be "data". And it would be worse than no data, just like yours.

Except that I actually have a poll sample.

They do say its tracking AF is behind competition if you don't select a specific subject, but they're also saying is that the camera should be easier to use, that it lacks the Auto subject recognition the Nikon (Z9) has. Nikon doesn't have it on their $2000 cameras either.

Generally speaking, I imagine it's only a matter of time before it comes to every camera. 10-15 years ago we had the same issue with compact cameras, were we had to manually select a scene, or else it would try and shoot at 1/200s in low light in Auto mode.

[cont'd]

Then again, there are photographers and applications (and, potentially, customers) which demand more, and the stupidest thing OMDS could do would be saying "we don't want them as customers" if they could easily and perfectly well serve them, too.

But, maybe most importantly, by denying its customers further megapixel increases, OMDS would forgo the only real advantage of the smaller sensor format beside its compactness potential, which is its effective advantage for both macro magnification and telephoto reach. With APS-C at 33 MP and FF at 60 MP, no-one needs to increase focal length by full 50% (APS-C) or 100% (FF) anymore to get the same magnification or reach as MFT. And better still—for any other application that doesn't need cropping the APS-C or FF camera would retain its full format and megapixel advantage. [cont'd]

[cont'd]

And with OMDS gear more and more being (rightly so) advertised for natur/wildlife/outdoor photography, that would be fatal. OMDS needs to continue to increase megapixel counts to keep its native advantage for those applications, lest more and more people rather go to larger formats even for those.

@Hubertus Bigend:

Yes, valid points!

But are OMDS not depending on sensor manufacturers to actually offer them a higher pixel count sensor in the first place? Even if it can be said OMDS (Olympus before them) did at least some parts of the sensor design and had it manufactured, I think it is such a big thing, overall, that OMDS cannot, at this stage, already offer a sensor that BOTH offer less noise AND higher pixel count.

This time they went for (or, "managed to offer") less noise, and from early reports, a notable step in that direction is now made. Next time we may hope they will at least keep this new low level noise and offer a notable step upwards regarding resolution.

When the E-M5 came it was both less noisy and had higher resolution, the same for E-M1.2, albeit a bit smaller step up than that from 12>16 mpix.

This time it simply was too early(?) to enable BOTH higher res. and less noise.

In that context I think "less noise" was absolutely the best choice!!
;-)

Others are having much better experiences using the generic C-AF + Tracking mode, so don't be too quick to make those assumptions. See for instance:

https://www.dpreview.com/forums/post/65945009

Those that feel the autofocus is More then enough for their needs. I shoot portraits and only need one af point with no tracking so the af for this camera would be more then fine for my needs

Probably should have watched the review video then. The AF tests showed it to be much improved, but there is one test where it did not do so well unless used in combination with another AF feature.

"Which popular moving subjects are still missing?"

The obvious one that's missing is human detecting in scenarios where the face is not detectable by a camera. Panasonic has it on their cameras and it's great.

"The E-M1 III and E-M1X prioritize near distance instead and perform much better."

I can't say I agree with this at all. I've owned the E-M1ii/iii/X, even all at the same time for about a month, and I never saw any difference in the general tracking performance at all.

No, I didn't. I already bought the GX9 and also the E-m5 III, as I prepare for Olympus' exit. I believe the GX9 is a better bet than E=P7 for my needs for that type of camera, and it also allows me to use Panasonic lenses with fully functionality, and helps to bridge over to Panasonic M43 products, just in case OMDS never bothers with compact lenses again or it vanishes entirely. I agree that the E-P7 and E-M5 III and IV are the types of cameras OMDS should focus on, compact, affordable and more purchaseable (Why buy very expensive products from companies that may disappear?)

Theoretical indeed. Went a bit further and researched the current lenses to see what they can resolve in line pairs/millimeter, and it dampens this quite a bit.

Found line pair/millimeter tests from LensTips.

Olympus 17mm f/1.2 Pro @2.8 98.6lp/mm is the sharpest = 37.8MP Resolved
Olympus 45mm f/1.2 Pro 96lp/mm = 35.8MP Resolved
Olympus 300mm f/4 Pro only 73lp/mm = 20.7MP Resolved (Bad Sample? Don't know)
Olympus 75mm f/1.8 82lp/mm = 26.1MP

So, we really need outstanding lenses to get to 35MP or so. Don't think we will see any improvement at 40MP and beyond without some kind of super lenses near 100lp/mm or higher. You would need a lens resolving 110lp/mm to resolve 47MP in a 4/3 sensor.

I wonder if some computational approach could be used with the 4 subpixels on this new sensor to reduce diffraction. Many cameras come out with firmware updates a year or two later that extend the functionality and features.

This has me thinking. At 400mm f4.5 has a diffraction limit of 44.6MP. Suppose you want that level of DOF achieved.

A 800mm FF lens would have to be stopped down to f9 to achieve the same DOF at the same focus distance. So if you NEEDED that level of DOF, your diffraction drops the lens below the resolution of a current OM-1?

tmascolo: how did you compute those 'MP resolved'? Those tests by Lenstip are done at 16Mpx or even 12Mpx, so it seems quite nonsensical to obtain such high numbers. At 100lp/mm I get:

4/3 x (2x100x13)^2 = 9 'Mpx'

GXAlan: if you use equivalent settings, the effects of diffraction are the same across formats.

@kolyy Hope you are well. Example computation Line pairs per millimeter for the 17mm f/1.2 shot at f/2.8 was 98.6lp/mm. Formula is (1000/(98.6*2))/2) = 2.54 microns of pixel pitch resolved. Sensor size of 4/3 is 18mm x 13.5mm. 18mm / .00254 = 7,099.2 , 13.5mm / .00254 = 5,324. 7,099 x 5,324 = 37,795,076 or 37.8MP.

tmascolo: Thanks, I think you might have an error there, what is the reason for the additional factor of 2 in that 'pixel pitch resolved' calculation? Also, the sensor size is 17.3x13mm, but that is a minor factor.

@GXAlan on a 35mm sensor pixel pitch is going to give you better numbers. No worries at f/9 on 35mm sensor, theoretical diffraction limited megapixels, if we average red/green/blue light, would be 39.33MP. Sensor size does matter. Also, easier to design lenses for 35mm since the pixel density won't be as high at the same megapixels.

At f/9 on a 35mm sensor of 35.9mm x 23.9mm the Diffraction Limited Resolution in Megapixels are:

Green light = 36.8MP
Red Light = 26.3MP
Blue Light = 54.9MP

Later,

Tony

tmascolo: yes, that is one factor of 2. But you have another factor of 2 in your calculation.

@kolyy Hope you are well. The last division by 2 is for adjusting for a color sensor, if it was a monochrome sensor you would divide by 1.41.

See previous reference page: https://www.optowiki.info/faq/how-to-convert-line-pairs-per-millimeter-to-pixel-size/

Thinking about lens resolution more now on 4/3 vs 35mm sensor. 4/3 really needs outstanding lenses, and 35mm can get away with lesser lenses for a given megapixel count.

For Example:

35mm Sensor
73lp/mm Lens = 73.2MP resolved
60lp/mm Lens = 49.4MP resolved
45lp/mm Lens = 27.8MP resolved

4/3 Sensor
73lp/mm Lens = 20.7MP resolved
60lp/mm Lens = 14MP resolved
45lp/mm Lens = 7.9MP resolved

Much easier to design lenses for 35mm sensor.

This makes no sense to me. We know the pixel size. For example, it's 3.8um for the E-M5 II used in the test of the 17mm/1.2. That's 3456 pixels per picture height. The line widths per picture height measured by Lenstip (in the center) is 2x99x13 = 2574 lines (at MTF50). That's reasonable, though substantially lower compared to what Opticallimits have found with same camera (around 3100 lw/ph).

Then you apply a completely arbitrary factor of 2, which is then squared to 4, to arrive to some 'MP resolved' metric, which has many times the real pixel count of the sensor and has no basis in the number of lines resolved.

I did not make the formula. Did you read the reference site I listed that explains the formula? https://www.optowiki.info/faq/how-to-convert-line-pairs-per-millimeter-to-pixel-size/

I will use the 35mm sensor size of 35.9mm x 23.9mm and a lens that measures 73lp/mm.

73lp/mm = 146 lines per millimeter (multiply by 2 to get lines per millimeter)

Next step convert from lines per millimeter to microns

1000/146 = 6.85 microns

Next step: What is your application Monochrome or Color?

If Monochrome divide by 1.41, so 6.85/1.41= 4.86 micron pixel pitch

If Color Bayer divide by 2, so 6.85/2 = 3.43 micron pixel pitch

Will use Color sensor of 35.9mm x 23.9mm and divide the horizontal and vertical by 3.43 micron pixel pitch. Have to convert from microns to millimeters first.

35.9mm / .00343mm = 10,466 horizontal pixels

23.9mm / .00343mm = 6,968 vertical pixels

10,466 x 6,968 = 72.9MP max resolution resolved on a 35mm sensor with a lens measuring 73lp/mm.

All steps shown.

It's not enough to just use the formula, it's quite crucial to understand what it means. Looking around that site I think I understand where the confusion comes from. What they are talking about is whether the lens, presumably tested on a purely optical bench, "supports" a sensor with a particular pixel density. What they mean, I think, is that the lens is then good enough for such a sensor, as they explain here:

https://www.optowiki.info/faq/why-can-color-cameras-use-lower-resolution-lenses-than-monochrome-cameras/

A few points:
- the site is very outdated. An ideal lens in their thinking is such that it does not induce moire, because the alternative is "an (expensive!) so called “OLPF” (=”Optical Low Pass Filter”)". But OLPF was a standard part of sensors for decades, so a too sharp lens has not been a problem for a very long time. Furthermore, OLPFs has been recently ommited as the users are demanding more perceptual sharpness, false detail being damned. A good lens is often identified by its ability to induce moire and thus outresolve the sensor. In other words, the standard they are applying is pretty low and most tests would evaluate a lens which can only resolve 1/2 of the pixel pitch as soft (e.g. less than 2000lw/ph on a 20Mpx M43 sensor with 3900 pixels per ph).
- looking at the performance of a lens on a low Mpx sensor and concluding it can "outresolve" a much higher Mpx one is circular reasoning. One would need an optical bench (no sensor) to evaluate that.

Exactly. Therefore if your pixel density of your sensor is denser than the calculated value then the lens can not resolve the full resolution of the sensor.

Therefore in my 17mm example, if you go to a 90MP 35mm sensor than the lens will still only resolve 72.9MP not the full 90MP of the sensor. The lens becomes the limiting factor of the system. You will, however, get a bigger version of that 72.9MP image for printing purposes obviously.

And one more point - we are only talking about center sharpness on the 17mm/1.2. The outer parts of the image are much softer, so concluding it "outresolves" a very high Mpx sensor based only on the performance in the center is not quite correct.

@kolyy Alright, fair enough. You disagree with the reference site, and how I applied their formula to create data points.

Can we agree that it is easier to design a lens that can resolve 25MP on a 35mm sensor than a 4/3 size sensor of the same MP count? Is that point at least valid?

If we agree on that, then the theoretical numbers do indeed support that hypothesis.

Did you have any issue with my diffraction limited resolution data?

Later,

Tony

@kolyy You are indeed correct, with the 17mm @f/2.8 in the center was the highest lp/mm measured. Once you change the aperture, those numbers decrease. Also, at the edges the lp/mm measures worse. Almost no lens is perfect across the range. You would need to design a perfect flat field optic, and it would only be optimal at one aperture most likely.

Well we can see from the OM-1 that the AF system is actually not the issue. When it comes to speed and focusing. It’s the image or target recognition part of the software that seems to be lacking. Nikon/canon/sony has addressed this by machine learning/training. The same sort of technique would need to be used.

Not saying I am a programmer, but I think they would be able to pull it off if they make it a priority.

Oly had subject specific AF before anyone else, this was at their height though, fortunes change but there’s also the matter of hardware limitation, AI assisted AF still needs light.

On the topic of computational photography trickling down from smartphones, I think this might actually be an advantage for Sony.
We all know they have negligible mobile phone market share, and are chasing niches such as using Xperia Phones as OLED displays for their cameras (why not a recorder too, like Ninja?). But their expertise in mobile phones can pay off really well if they can implement the related mobile processing and interface developments to their cameras. If I were making decisions at Sony I’d try to keep one feet in the mobile industry for as long as possible.

@noirdesir

Sorry, but you also can’t say anything about apple’s investment in silicon. After all that terms is extremely broad and generic. Apple heavily develops it own chips for several products. Think A and M processors. That is a whole different ball game. When it comes to camera silicon, those things are dirt cheap, and they are likely bought from places like sony.

When it comes to AI learning, once you have the algorithm there, if the issue is that the pool if images you use is small, then you simply need to increase that pool At lot of the work involved is automated.

And my suggestion about them improving this is about low hanging fruit. Canon’s, and Sony’s investment in that segment. You don’t think it pays dividends in the long run? The subject tracking algorithm was developed once (for the R6 for example, and basically ends up in the R5, R3, and eventually the R1 and any other camera they release.

Sure they will improve on it over time, but the major portion of the R&D investment is for the first version.

If OM wants to stay relevant with there cameras, they they pretty much have no choice but to make improvements here. It is pony up or fade aware. They promote this thing like a sports/adventure/wild life camera. One would expect the AF to do well. Especially when comparing it to other cameras.

I would consider getting an oly due to the potential of a smaller/lighter setup and most importantly a really well weather sealed camera. But for the cost… why would I not just get an R6 instead? I rather put a rain coat on and R6 and have great/awesome AF, than simply have the oly for its weather sealing. After all… I am already on the RF system.

@lawny13:

"But hey… this is a problem with sandboxing in companies. The camera division likely doesn’t know what is available in the medical decision"

Olympus, during many years, literally spoke out about their camera division vs medical division both sharing technical advancements from each other. At least they said the camera division contributed, and thus was not going to be tossed out of the company. Maybe that was pure marketing BS , or it may have been completely or partially true.
Until, finally, they had to realize the camera division did not contribute enough to remain "in house".

So, at long last, that didn´t save the camera division into producing black numbers in the books, sorrily.

The point remains that camera companies cannot have low profits or no profits at all and at the time lots of money in spare to spend on software and hardware. Yes, the SOCs of cameras are pretty cheap (just look what processing size they use). But there are cheap because the camera companies are struggling to make profits and thus have to save money wherever they can.

And all that ignores the vastly larger economies of scale of smartphone components.

these two are all over DPR... is there anything the won't do?

I feel like Chris and Jordan just like M43 as a system and wanted to chat about it. /shrug, not everything is made for you or I.