DSLR vs Mirrorless technology

[jbcohen]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

 

[ShootMeAlready]

This is about market segment more than technology. The only diff is about a mirror and longer tube to support a mirror. The data engine is the same issue, so is all over heating issues. Canon should have stayed in DSLR space and made a D5 MKIV that competed in video with Canon A7SII but they did not, as wanted to protect the high end video market. But by not competing Sony A7III has now become the go to indy came, and 5DMIV is only for still, and looked over for video. The future is video, stills is less common video only more. I think Canon killed the DSLR market by not innovating.

 

[justmeMN]

Any technology that is put into a Canon/Nikon mirrorless camera can also be put into their DSLR Live View. As one example, both Canon mirrorless cameras and Canon DSLR Live View use Dual Pixel Autofocus.

 

[Ellis Vener]

Back to the Future (or Forward Into The Past)

The Single Lens Reflex design was an amazing engineering feat, but it was essentially a work around engineered in the 1920s and 30s to get very close to what you see is what you get photography. It enabled a long line of optical innovation making possible every thing from ultra wide angle lenses including fisheye lenses that could project onto film up to a 220-degree angle of view, out to super telephoto lenses up to 2000mm which take in an angle of view measuring less than a degree.

But it was still a work around.

The development of first digital sensors and then live view feed from the sensor to a small high resolution LCD and OLED panels make it possible to get even closer to that WYSIWYG (through the lens) dream. The elimination of the space needed for the reflex viewing system also means that commonly used lenses (especially “normal” and wide angle lenses) can now be reformulated for better performance.

in short, photographic technology is evolving and I think one result is that more people will find the practice of photography more enjoyable, especially if you include in the definition of “mirrorless cameras” the recent generations of smart phones like the iPhone and Halaxy.

Some people won’t like this development but others will embrace it.

There is also the marketing and business aspect of selling cameras and lenses but that is separate.

 

[Mackiesback]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

Yes, because that is seen as the future, and that is where the most improvement can be made.

 

[Klaus dk]

I absolutely agree. One very significant aspect you forgot: production cost is lower with mirrorless, because the whole AF assembly is now built into the sensor. Since mirrorless cameras are not cheaper than DSLRs, the profit margins goes up, so in a few years time, we'll see no more new DSLRs.

 

[73Instamatic]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

From all the replies so far, it's easy to see a pattern...

When it comes to camera bodies, most of the R&D is being boiled down to what's happening in the chips (both the sensor and the imaging SoC).

You could make the case that before autofocus was a thing, cameras had pretty much reached a definite plateau by the end of the 1970's. Metering was through the lens, and the camera needed nothing more than a battery the size of a dime. Most cameras could still function even without the tiny battery if you were using an external meter or common rules of thumb (eg Sunny 16).

But once you started putting chips in there, camera bodies moved from essentially 100% mechanical devices to consumer electronics devices. Slowly at first. But always in the same direction. Always more and more and more discreet logic.

So it isn't necessarily just about getting rid of the mirror box. It's about making the AF better, the Live View / EVF better, the DR better, and of course bumping up MP's occasionally.

Regardless of whether the market demands bodies with or without a mirror, the R&D is mostly about the chips, and the usability (ie the software side of things).

My opinion is that mirrorless isn't the only factor driving the R&D, since the chips are used in both body styles. But I believe mirrorless will be the main benefactor.

 

[techie takes pics]

Mirrorless is definitely the direction.

You'll still hear that DSLR can do the same thing as mirrorless (facedetection, video, live view, etc, etc, etc) when they just flip the mirror out of the way.

That's the whole point: The sensor can do a lot of wonderful things, if only the mirror is out of the way.

So yes, the improvements are happening in software - and they are possible when the mirror is out of the way.

 

[Jon555]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

Possibly more because the big two decided to launch MILC models so had a bunch of work to do and a deadline.

As mentioned a lot of things can be shared between camera types. Even on-sensor PDAF.

There have been some DSLR-only improvements coming along, but most seemed to be on pause. Motor driven mirrors and transparent LCDs are presumably still being improved, as are the discrete AF sensors and the exposure tracking sensors that sit beside them.

I think it's really a case of where the DSLR manufacturers go with their next Halo and Halo-ish products, i.e. the 1Dx mk III, D6, 7D mk III and D600 (or whatever they call them). If they go for another generation as DSLRs then a bunch of technology should appear. I'm not sure but unless they can run an EVF continuously for 3 hours (along with all the other stuff) and have 50% battery left they might go with DSLRs for one more round. (BTW I bought a pair of small binoculars to check for wildlife so I wouldn't need my EVF on all the time as you wait an hour plus for something to make an appearance, or walk around looking for stuff in the medium+ distance.)

 

[Christoph Stephan]

I absolutely agree. One very significant aspect you forgot: production cost is lower with mirrorless, because the whole AF assembly is now built into the sensor. Since mirrorless cameras are not cheaper than DSLRs, the profit margins goes up, so in a few years time, we'll see no more new DSLRs.

ON the other hand, the EVF will always need extra battery power, whereas its free solar power with an OVF. The advantage of the DLSR in battery power will remain. And there are photo situations where this is significant, read the other poster below (1).

--
Chris
-----
http://www.redbubble.com/people/christopher363

 

[Ellis Vener]

I absolutely agree. One very significant aspect you forgot: production cost is lower with mirrorless, because the whole AF assembly is now built into the sensor. Since mirrorless cameras are not cheaper than DSLRs, the profit margins goes up, so in a few years time, we'll see no more new DSLRs.

ON the other hand, the EVF will always need extra battery power, whereas its free solar power with an OVF. The advantage of the DLSR in battery power will remain. And there are photo situations where this is significant, read the other poster below (1).

Which mirrorless and which DSLRs have you photographed with?

 

[Michael Fryd]

Back to the Future (or Forward Into The Past)

The Single Lens Reflex design was an amazing engineering feat, but it was essentially a work around engineered in the 1920s and 30s to get very close to what you see is what you get photography. It enabled a long line of optical innovation making possible every thing from ultra wide angle lenses including fisheye lenses that could project onto film up to a 220-degree angle of view, out to super telephoto lenses up to 2000mm which take in an angle of view measuring less than a degree.

...

The SLR design gets around the issue that film doesn't provide the immediate feedback necessary for aiming, framing, auto-focus, focus-tracking, metering, etc.

The mirror initially allowed the use of a separate sensor (the human eye) to aim and frame. Metering was soon added. Eventually the mirroring system allowed for a number of special purpose sensors designed for their specific purpose. (Focus, subject tracking, etc.)

With mirrorless, you have to do everything with a single sensor. The engineering challenge with mirrorless is how to get a single general purpose sensor to do a great job at a variety of very different tasks.

For instance, my DSLR looks at RGB+IR light when tracking a subject. It's tough to get the mirrorless sensor to track IR light, without sacrificing some of the ability to capture visible light for the final image.

Similarly, the purpose-built phase detect focus sensors in my DSLR have sensors for both horizontal and vertical details. One set of regular precision for regular lenses, and a set of high precision sensors that are used with fast lenses.

Obviously, modern mirrorless cameras are more than good enough for the vast majority of photography, however overloading so much functionality into a single sensor presents challenges to making it good at everything.

To be fair, DSLRs have their own challenges. Using separate sensors for each function requires careful alignment. The mirroring mechanism is complicated, and the mirror prevents the rear lens element from getting too close to the sensor.

====

Of course none of this matters. They're both more than good enough. History tells us that technical issues probably won't be a major factor in which ends up dominating the market.

The real issue will be marketing.

My prediction is that it's going to boil down to whether or not mirrorless can convince consumers that it offers something better than a smart phone. With a DSLR, it's easy to show differences (there's all that complicated mechanics). From a consumer point of view, a mirrorless is like a smart phone, except the smart phone has multiple cameras at a variety of different angles of view.

 

[Jon555]

Mirrorless is definitely the direction.

You'll still hear that DSLR can do the same thing as mirrorless (facedetection, video, live view, etc, etc, etc) when they just flip the mirror out of the way.

BTW a number of DSLRs do face-detection with the mirror down by using the exposure sensor as an image sensor.

That's the whole point: The sensor can do a lot of wonderful things, if only the mirror is out of the way.

So yes, the improvements are happening in software - and they are possible when the mirror is out of the way.

 

[Ellis Vener]

Back to the Future (or Forward Into The Past)

The Single Lens Reflex design was an amazing engineering feat, but it was essentially a work around engineered in the 1920s and 30s to get very close to what you see is what you get photography. It enabled a long line of optical innovation making possible every thing from ultra wide angle lenses including fisheye lenses that could project onto film up to a 220-degree angle of view, out to super telephoto lenses up to 2000mm which take in an angle of view measuring less than a degree.

...

The SLR design gets around the issue that film doesn't provide the immediate feedback necessary for aiming, framing, auto-focus, focus-tracking, metering, etc.

The mirror initially allowed the use of a separate sensor (the human eye) to aim and frame. Metering was soon added. Eventually the mirroring system allowed for a number of special purpose sensors designed for their specific purpose. (Focus, subject tracking, etc.)

All true, and as I wrote, the SLR design was a work around. Specifically and optical- mechanical work around. Mirrorless camera designs are a work around solution as well, but are more efficient at solving the problem.

With mirrorless, you have to do everything with a single sensor. The engineering challenge with mirrorless is how to get a single general purpose sensor to do a great job at a variety of very different tasks.

But it is a solvable problem.

For instance, my DSLR looks at RGB+IR light when tracking a subject. It's tough to get the mirrorless sensor to track IR light, without sacrificing some of the ability to capture visible light for the final image.

Similarly, the purpose-built phase detect focus sensors in my DSLR have sensors for both horizontal and vertical details. One set of regular precision for regular lenses, and a set of high precision sensors that are used with fast lenses.

But with mirrorless designs all the AF sensors are high precision ones.

Obviously, modern mirrorless cameras are more than good enough for the vast majority of photography, however overloading so much functionality into a single sensor presents challenges to making it good at everything.

It is not as large a hurdle as you present it to be, and the sensor and processing technology continues to develop.

To be fair, DSLRs have their own challenges.

Such as the physical and temporal limits imposed by bhaving to move the reflex mirror iout of and back into the light path.

Using separate sensors for each function requires careful alignment. The mirroring mechanism is complicated, and the mirror prevents the rear lens element from getting too close to the sensor.

Exactly.

====

Of course none of this matters. They're both more than good enough. History tells us that technical issues probably won't be a major factor in which ends up dominating the market.

And we can trust history on this.

The real issue will be marketing.

Forget the forum discussions. The marketing Reality is that the more people who put their hands on a mirrorless camera that performs on par with whatever DSLR camera they are considering the more they will see the advantages of the mirrorless approach.

My prediction is that it's going to boil down to whether or not mirrorless can convince consumers that it offers something better than a smart phone.

YES! That is the real challenge, not a DSLR vs mirrorless ILC.

With a DSLR, it's easy to show differences (there's all that complicated mechanics). From a consumer point of view, a mirrorless is like a smart phone, except the smart phone has multiple cameras at a variety of different angles of view.

I disagree with you there. For most people’s needs, the photographic system in even a middle of the pack a smart phone is likely deemed good enough. The real advantage a camera equipped smartphone possesses for these people is that you already have it with you and it is easy to use. It fulfills the marketing mantra of the original mass marketed Kodak camera: “You push the button and we do the rest.”

 

[Michael Fryd]

...

With mirrorless, you have to do everything with a single sensor. The engineering challenge with mirrorless is how to get a single general purpose sensor to do a great job at a variety of very different tasks.

But it is a solvable problem.

I don't see how you would add IR subject tracking to a typical sensor, without impinging on the ability to capture visible light.

This suggests that there is a disadvantage to the mirrorless design.

I think the "solution" is to make the the sensor performance good enough, that it isn't a problem to lose some of it to IR. The DSLR sensor has the potential to be better, but that "better" may not be important.

Once a camera performs "more than good enough," there may be little advantage to further improvements.

Both mirrorless and DSLR are both certainly more than "good enough".

For instance, my DSLR looks at RGB+IR light when tracking a subject. It's tough to get the mirrorless sensor to track IR light, without sacrificing some of the ability to capture visible light for the final image.

Similarly, the purpose-built phase detect focus sensors in my DSLR have sensors for both horizontal and vertical details. One set of regular precision for regular lenses, and a set of high precision sensors that are used with fast lenses.

But with mirrorless designs all the AF sensors are high precision ones.

Yes and no. The phase detect functionality of the DSLR is better, as it can work with both horizontal and vertical details.

However, the contrast detect focus sensors of the mirrorless is obviously perfectly matched to the image capture sensor.

Phase detect has a reputation for faster performance (as it not only detects "out of focus", but also how much, and in which direction.

As to which is more important - that depends on what you are shooting.

Obviously, modern mirrorless cameras are more than good enough for the vast majority of photography, however overloading so much functionality into a single sensor presents challenges to making it good at everything.

It is not as large a hurdle as you present it to be, and the sensor and processing technology continues to develop.

Absolutely. Sensor technology continues to develop. But that benefits both mirrorless and DSLR.

At the moment, that are both more than good enough for the vast majority of photography.

To be fair, DSLRs have their own challenges.

Such as the physical and temporal limits imposed by bhaving to move the reflex mirror iout of and back into the light path.

Yes. It's a challenge. To be fair, the engineers have had over 50 years to refine the engineering that makes this possible.

Using separate sensors for each function requires careful alignment. The mirroring mechanism is complicated, and the mirror prevents the rear lens element from getting too close to the sensor.

Exactly.

====

Of course none of this matters. They're both more than good enough. History tells us that technical issues probably won't be a major factor in which ends up dominating the market.

And we can trust history on this.

The real issue will be marketing.

Forget the forum discussions. The marketing Reality is that the more people who put their hands on a mirrorless camera that performs on par with whatever DSLR camera they are considering the more they will see the advantages of the mirrorless approach.

Perhaps.

My prediction is that it's going to boil down to whether or not mirrorless can convince consumers that it offers something better than a smart phone.

YES! That is the real challenge, not a DSLR vs mirrorless ILC.

The dominant camera is the smartphone. In order for a different type of camera to survive in the marketplace, it has to differentiate itself from the smartphone.

With a DSLR, it's easy to show differences (there's all that complicated mechanics). From a consumer point of view, a mirrorless is like a smart phone, except the smart phone has multiple cameras at a variety of different angles of view.

I disagree with you there. For most people’s needs, the photographic system in even a middle of the pack a smart phone is likely deemed good enough. The real advantage a camera equipped smartphone possesses for these people is that you already have it with you and it is easy to use. It fulfills the marketing mantra of the original mass marketed Kodak camera: “You push the button and we do the rest.”

I think we see this differently. It will be interesting to see if either of us turns out to be close with our predictions.

The issue is not whether or not something is better than a smartphone, but whether the market sees it as being worth buying in addition to the smartphone.

Smartphones are certainly evolving. Multiple cameras with varying focal lengths give the camera a lot of data to work with. Combining the data from these cameras offer noise reductions for the center of the image. The 3D distance map allows shallow depth of field effects. With all those cameras, an additional one could be included for wide band IR subject tracking.

As smartphones become better cameras, mirrorless will be in danger of being perceived as the larger, less capable, older technology.

DSLRs also may suffer this fate. by DSLRs may be able to maintain enough differences that they maintain their niche.

====

Many years ago there was an entire industry called "typesetting".

Then Pagemaker came along. It wasn't nearly as good as professionally typeset material, but it was more than "good enough", more convenient, and less expensive.

There is no longer a typesetting industry. They no longer make dedicated typesetting machines.

There are lots of details that PageMaker got wrong. Professional typesetters could easily see the inferior quality. The market didn't care. Customers abandoned higher quality and went with lower price and convenience.

Both mirrorless and DSLR have a struggle ahead of them.

 

[justmeMN]

Ellis Vener wrote:.
Forget the forum discussions. The marketing Reality is that...

most people in the USA don't even know what a mirrorless camera is.

 

[dmanthree]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

Because that's the future. Very obvious, really.

 

[DiffractionLtd]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

About time. BTW, most of the innovations have already been done by Panasonic and Olympus. You're welcome, Canon and Nikon.

 

[007peter]

I get the feeling that most of the research and development in photography is going on in the mirrorless cameras. What is your opinion?

EASY. The reason is 100% because of video. If your HONEST about DSLR, most of Still Photography development has tapped out sometimes ago. Majority of "INNOVATION" has been video related

2007 Canon 40D introduce the 1st LIVE VIEW in DSLR, which brings about

2008 Nikon D90 - 1st (aps-c) DSLR with 720p video

2008 Canon 5Dmk2 1st (fullframe) DSLR with 1080p video

• ***************************************
• DSLR VIDEO RACE BEGIN → MIRRORLESS push
• ***************************************

DSLR video is done by (flipping up the mirror) to expose the sensor for continuous video. In short, MIRROR got in the way of video shoot. As video advances, so is the push toward mirrorless camera. Plain & Simple.

 

[Jon555]

Back to the Future (or Forward Into The Past)

The Single Lens Reflex design was an amazing engineering feat, but it was essentially a work around engineered in the 1920s and 30s to get very close to what you see is what you get photography. It enabled a long line of optical innovation making possible every thing from ultra wide angle lenses including fisheye lenses that could project onto film up to a 220-degree angle of view, out to super telephoto lenses up to 2000mm which take in an angle of view measuring less than a degree.

...

The SLR design gets around the issue that film doesn't provide the immediate feedback necessary for aiming, framing, auto-focus, focus-tracking, metering, etc.

The mirror initially allowed the use of a separate sensor (the human eye) to aim and frame. Metering was soon added. Eventually the mirroring system allowed for a number of special purpose sensors designed for their specific purpose. (Focus, subject tracking, etc.)

All true, and as I wrote, the SLR design was a work around. Specifically and optical- mechanical work around. Mirrorless camera designs are a work around solution as well, but are more efficient at solving the problem.

EVFs effect on battery life and some focusing issues are still needing work. Also will the big two throw a lot more tech at one more round of DSLRs or not? (A much cleverer transparent LCD in the viewfinder would be interesting, the current ones don't do that much. Also Exposure sensors with more pixels for better face/eye detection.)

With mirrorless, you have to do everything with a single sensor. The engineering challenge with mirrorless is how to get a single general purpose sensor to do a great job at a variety of very different tasks.

But it is a solvable problem.

There's still an issue that the better you make the sensor the more heat you're likely to get if you are running it all the time. Plus on-sensor PDAF needs some serious tech input. I suspect they will get ahead of DSLRs when the DSLR manufacturers stop throwing so much development money in their direction, whenever that might be.

For instance, my DSLR looks at RGB+IR light when tracking a subject. It's tough to get the mirrorless sensor to track IR light, without sacrificing some of the ability to capture visible light for the final image.

Similarly, the purpose-built phase detect focus sensors in my DSLR have sensors for both horizontal and vertical details. One set of regular precision for regular lenses, and a set of high precision sensors that are used with fast lenses.

But with mirrorless designs all the AF sensors are high precision ones.

I think there is still some way to go with on-sensor PDAF and contrast detect will continue to have issues with not knowing which way to move. DSLR AF sensors have a range of measuring baselines from small to huge, which can really help with long lenses, also you can get them sensitive to lines that are vertical, horizontal and both diagonals, rather than just the one of those you get with current mirrorless PDAF. Plus non-Canon mirrorless PDAF has some banding issues in low light or high glare. A fair bit of dev is required but they're fine for most use cases now.
That said my mirrorless cameras are usually great for single-shot AF and track stuff reasonable well (in the case of the on-sensor PDAF one anyway), they do lose out for BIF though where the DSLR is just magic at locking on and staying on.

Obviously, modern mirrorless cameras are more than good enough for the vast majority of photography, however overloading so much functionality into a single sensor presents challenges to making it good at everything.

It is not as large a hurdle as you present it to be, and the sensor and processing technology continues to develop.

That is an interesting question, as phones tend to have a lot more processing available than cameras. Camera volumes being low (compared to phones) I doubt most manufacturers could develop a phone-level SoC as they wouldn't get the money back. I wonder if a SoC manufacturer might do a cut-down one for cameras at some point, but suspect the camera people wouldn't want to do that much software work.

To be fair, DSLRs have their own challenges.

Such as the physical and temporal limits imposed by bhaving to move the reflex mirror iout of and back into the light path.

Using separate sensors for each function requires careful alignment. The mirroring mechanism is complicated, and the mirror prevents the rear lens element from getting too close to the sensor.

Exactly.

A big cost advantage for mirrorless, but with most manufacturers at under 500k mirrorless units per year (except Sony, Canon) it doesn't seem to affect prices as spreading out all the other costs (plus a desire for decent profits) nukes it.

====

Of course none of this matters. They're both more than good enough. History tells us that technical issues probably won't be a major factor in which ends up dominating the market.

And we can trust history on this.

The real issue will be marketing.

Forget the forum discussions. The marketing Reality is that the more people who put their hands on a mirrorless camera that performs on par with whatever DSLR camera they are considering the more they will see the advantages of the mirrorless approach.

I don't quite see that as while DSLR sales have fallen they are still +90% on MILC and I think brand recognition for Canon/Nikon is a significant factor. Also a lot of people see DSLRs as "real" cameras and mirrorless as just large compacts. This will only change slowly (but it has changed somewhat).

My prediction is that it's going to boil down to whether or not mirrorless can convince consumers that it offers something better than a smart phone.

YES! That is the real challenge, not a DSLR vs mirrorless ILC.

I think they are somewhat screwed on this (along with DSLRs) as no-one is making the sort of changes that would achieve that. I'd like to see a version of Apple's Car Play for phones, so the camera can link seamlessly to the phone (when in range) and you get a simplified phone interface on the camera to send images to it, plus direct to social media or e-mail. (Also a bunch of user-interface stuff for me, but that's a longer post.)

With a DSLR, it's easy to show differences (there's all that complicated mechanics). From a consumer point of view, a mirrorless is like a smart phone, except the smart phone has multiple cameras at a variety of different angles of view.

I disagree with you there. For most people’s needs, the photographic system in even a middle of the pack a smart phone is likely deemed good enough. The real advantage a camera equipped smartphone possesses for these people is that you already have it with you and it is easy to use. It fulfills the marketing mantra of the original mass marketed Kodak camera: “You push the button and we do the rest.”

I suspect a lot of cameras are still sold over events. Have a big holiday, have a kid, etc. Where people want images to keep rather than for their 15 minutes of fame on social media.