Death of the flagship DSLR camera?

[shadowblade]

So, what's the conclusion? Is it dead or just moribund?

Rob

Not dead yet - just sitting around in its nursing home, waiting for the inevitable. At present, it can shoot action about as well as action-oriented mirrorless bodies, while still having a greater range of action-capable lenses (although mirrorless bodies can take just about any lens, most of them aren't exactly action-capable when using an adapter). For nonmoving subjects, mirrorless bodies (or SLRs using live view, which are functionally the same thing) are already better, due to accuracy of focus when zoomed in and the true WYSIWYG nature of through-the-lens composition.

Once 8k video hits and we have mirrorless cameras shooting at 39MP/24-30fps, it will be on life support, kept alive mostly for enthusiasts of that technology without being better than its replacement at anything.

[BernardLanguillier]

Not dead yet - just sitting around in its nursing home, waiting for the inevitable. At present, it can shoot action about as well as action-oriented mirrorless bodies, while still having a greater range of action-capable lenses (although mirrorless bodies can take just about any lens, most of them aren't exactly action-capable when using an adapter). For nonmoving subjects, mirrorless bodies (or SLRs using live view, which are functionally the same thing) are already better, due to accuracy of focus when zoomed in and the true WYSIWYG nature of through-the-lens composition.

Well... it is sometimes easy to forget that these bodies are amazing image capturing machines and will remain on top of most applications in the future.




D5 + 70-200 f2.8 E FL

Cheers,
Bernard

[shadowblade]

Well... it is sometimes easy to forget that these bodies are amazing image capturing machines and will remain on top of most applications in the future.

The fact that they're good doesn't mean they won't be replaced by something better.

Image quality is a function of the sensor (and supporting electronics) and glass, which can be the same for both mirrorless and SLR cameras. So that part's out of the equation - once you've captured it, there's no difference between an image captured on an SLR camera and that captured on a mirrorless camera. Your SLR images wouldn't look any different had they been taken on a mirrorless camera. The main difference is the process of capturing it - getting an image that's focused on the right part of the subject, taken at the right moment and correctly exposed. That's where mirrorless has been making huge progress, catching up to SLRs with the release of the A9, and where they will surpass mirrorless cameras in the near future, with faster frame rates (you can't make an SLR shoot at 25fps while using the mirror, while video demands it, and an 8k video camera would be able to shoot 39MP stills at the same speed), AI-based AF (an expansion of eye focus) and WYSIWYG exposure. SLRs are up against the limits of their technology - you can do a lot with the raw data from the same sensor as that used to capture the final image that you just can't do with a dumb mirror and prism.

[BernardLanguillier]

The fact that they're good doesn't mean they won't be replaced by something better.

Certainly, but considering the current level, we will be deep into the area of diminishing returns.

Cheers,
Bernard

[bcooter]

The fact that they're good doesn't mean they won't be replaced by something better.

snip

I don’t understand the mirrorless, vs standard dslr discussion.

I’m not against mirrorless, own 5 from 4/3 to full frame, but in reality the 1dx 1 and 2 can perform and focus as well if not better than pure mirrorless.

If your dslr has that dual pixel thing, Just clip  a magnifier on the back and shoot. 

Since I photograph moving people the only issues I have with mirrorless is battery use and except for the newest A-9 (I think that is what it’s called) is the only mirrorless camera that will track focus well.

The thought that mirrorless will eventually shoot 24,  30 or 60 fps and you pull a pristine 12 or 14 bit still that is as sharp and deep as a standard still image is a ways off. at least from a small camera. 

Have you seen the size of an 8k RED W?  It makes a 1dx II and a D5 look tiny and motion cameras that shoot a high bit rate get hot and are heavy.

Then there is the technical problem of shutter speed/angle.    To shoot smooth motion imagery your at 1/48th, 1/60th, 120th or if the camera goes to 120fps (very few do) 1/240th.

If you shoot at 24fps and set the shutter at 125th you get a stutter look on the footage, so until someone comes out with a software that will compensate for that your either shooting stills or motion, but not both together.

We just finished a series of commercials that included a still shoot.   For most of the footage we shot with the REDs, for some of the footage and all the stills we shot with the 1dxII, but the 1dx2 needed a different shutter and focus setting.  Actually it was easier just to shoot the 1dx 1 set up for stills and just recreate the scene.  One thing I noticed from the 1dxII was it shot the sharpest stills pulled from motion footage I’ve seen.   

Going slightly off topic was how well the 1dxII’s  422, 8 bit (not 10) motion file held up in post.   Probably due to the large 800 mbs file and the older but excellent motion jpeg codec.   The downside is it burns through a 128gb cfast2 card in about 20 something minutes and it takes a good computer to grade the footage in it’s original state.   On some scenes we ran two cameras, R1 and the 1dxII and even though the 1dx II is 8 bit vs. the REDs 13 bit the dynamic range of the Canon very closely matched and held up to the REDs.   The Canon has more vibrant, canon look colour but we set the REDs closer to the Canon and the Canon as close as possible to the REDs.

Other than carrying another system and the case(s) that this requires, my feeling is there is not much advantage of having one camera shoot stills and motion, but other’s see it differently.

I could be wrong.   Maybe we’ll all be shooting like this someday.  (the camera not the wardrobe . . . maybe)




IMO

BC

[BernardLanguillier]

Exactly.

The dream to extract stills from a movie completely overlooks the issue of proper shutter speed.

And on the original point, rumors about a D5s to be announced soon are starting to surface.

Cheers,
Bernard

[shadowblade]

Exactly.

The dream to extract stills from a movie completely overlooks the issue of proper shutter speed.

And on the original point, rumors about a D5s to be announced soon are starting to surface.

Cheers,
Bernard

No-one ever suggested extracting stills from a video clip. That would obviously give you a lot of motion blur.

But what you can do is shoot stills using the same camera as that used to shoot video, without any intent to shoot it as a video clip.

A camera shooting 25fps video at 1/40s exposure can equally shoot 25fps at 1/500s or 1/1000s exposure (with correspondingly higher ISO or wider aperture. It would make for an unusable, choppy video clip, but would essentially be a series of perfectly-exposed stills taken at a fast frame rate.

Basically, you're not shooting video - you're shooting stills at 25fps. It just so happens you can do both with the same camera, although not both at the same time.

[shadowblade]

I don’t understand the mirrorless, vs standard dslr discussion.

I’m not against mirrorless, own 5 from 4/3 to full frame, but in reality the 1dx 1 and 2 can perform and focus as well if not better than pure mirrorless.

If your dslr has that dual pixel thing, Just clip  a magnifier on the back and shoot.

If you're shooting in dual pixel mode, you're using your SLR as a mirrorless camera, with the mirror and viewfinder sitting out of the way and doing nothing.

Since I photograph moving people the only issues I have with mirrorless is battery use and except for the newest A-9 (I think that is what it’s called) is the only mirrorless camera that will track focus well.

That's the point. The A9 represents the new generation of mirrorless AF, and the current state of the art. It tracks just as well as the best SLRs, while allowing for the features a through -the-sensor approach can provide which an SLR can't. It means that mirrorless cameras have caught up to SLRs AF-wise. Future mirrorless bodies are hardly going to go backwards.

The thought that mirrorless will eventually shoot 24,  30 or 60 fps and you pull a pristine 12 or 14 bit still that is as sharp and deep as a standard still image is a ways off. at least from a small camera. 

Have you seen the size of an 8k RED W?  It makes a 1dx II and a D5 look tiny and motion cameras that shoot a high bit rate get hot and are heavy.

Go back a few years and you could have said the same thing about 4k cameras. A few more years back and the same would apply for 1080p. Now they're Gopro-sized.

Technology marches on. 8k cameras in three years' time will not be the size of current REDs.

Then there is the technical problem of shutter speed/angle.    To shoot smooth motion imagery your at 1/48th, 1/60th, 120th or if the camera goes to 120fps (very few do) 1/240th.

If you shoot at 24fps and set the shutter at 125th you get a stutter look on the footage, so until someone comes out with a software that will compensate for that your either shooting stills or motion, but not both together.

We just finished a series of commercials that included a still shoot.   For most of the footage we shot with the REDs, for some of the footage and all the stills we shot with the 1dxII, but the 1dx2 needed a different shutter and focus setting.  Actually it was easier just to shoot the 1dx 1 set up for stills and just recreate the scene.  One thing I noticed from the 1dxII was it shot the sharpest stills pulled from motion footage I’ve seen.   

Going slightly off topic was how well the 1dxII’s  422, 8 bit (not 10) motion file held up in post.   Probably due to the large 800 mbs file and the older but excellent motion jpeg codec.   The downside is it burns through a 128gb cfast2 card in about 20 something minutes and it takes a good computer to grade the footage in it’s original state.   On some scenes we ran two cameras, R1 and the 1dxII and even though the 1dx II is 8 bit vs. the REDs 13 bit the dynamic range of the Canon very closely matched and held up to the REDs.   The Canon has more vibrant, canon look colour but we set the REDs closer to the Canon and the Canon as close as possible to the REDs.

Other than carrying another system and the case(s) that this requires, my feeling is there is not much advantage of having one camera shoot stills and motion, but other’s see it differently.

I could be wrong.   Maybe we’ll all be shooting like this someday.  (the camera not the wardrobe . . .


IMO

BC

Read above post. You shoot stills using the same camera as video, not at the same time as video. 25fps stills or 25fps video - just not both at the same time. An SLR can't do that - at least not without holding the mirror out of the way, disabling all the mirror-dependent systems and functionally turning it into a mirrorlesa camera anyway.

[BJL]

No-one ever suggested extracting stills from a video clip. That would obviously give you a lot of motion blur.

But what you can do is shoot stills using the same camera as that used to shoot video, without any intent to shoot it as a video clip.

A camera shooting 25fps video at 1/40s exposure can equally shoot 25fps at 1/500s or 1/1000s exposure (with correspondingly higher ISO or wider aperture. It would make for an unusable, choppy video clip, but would essentially be a series of perfectly-exposed stills taken at a fast frame rate.

Basically, you're not shooting video - you're shooting stills at 25fps. It just so happens you can do both with the same camera, although not both at the same time.

Another possibility is a mode that records some still frames (at suitably higher shutter speed) interspersed in a video stream. The new H.265/HEVC/HEIC standard allows this sort of "mixed media files", and more generally seems designed with "still/video fusion" in mind.

[shadowblade]

Most criticisms of mirrorless technology, to the tune of 'they'll never replace SLRs' and 'SLRs are better for action', do one or more of the following:

1. Conveniently ignore the A9. 'Mirrorless cameras autofocus slowly and can't keep up with action SLRs.' You know, apart from the one mirrorless camera actually designed and sold as an action camera, and which can keep up with the 1Dx2 and D5 shooting any subject. Of course the other models can't keep up. They either weren't designed to (neither are medium-format bodies, but you never hear SLR-pushers criticising them) or were designed when mirrorless was in its infancy and the technology wasn't there yet. If you're going to compare technologies, compare the current state of the technology, not the state of technology several years ago. That's like saying cars can't keep up because the first cars was slower than the state-of-the-art, 1879-model Horse Mk III.

2. Criticise things which aren't inherent to mirrorless cameras. 'They're too small to hold' and 'They don't balance well with heavy lenses' are common ones. There's nothing that says mirrorless cameras need to be small - just that most current ones are. Besides, I don't find them too small, even when handholding the biggest lenses. That's because my left hand holds the ensemble by the lens - perfectly balanced - and my right hand is just there to push buttons and operate the controls. You can always gaffer-tape a brick to an A9 if it's too light for you. I can't exactly saw the grip off a D5 and still have a functional camera.

'Not enough lenses' or 'No long telephotos' is another common criticism. Again, there's no requirement that mirrorless cameras lack long lenses - that's a valid criticism of a system, not a valid criticism of the technology. Up until seven months ago, there wasn't a mirrorless camera with an autofocus system able to take advantage of them, so there was no point launching them. Now there is, which is why they're starting to appear (watch for the 400mm f/2.8, which should take care of most sports).

3. Appeals to authority. 'Most pro sports photographers shoot Canon/Nikon.' Well, duh. Momentum only shifts so fast. A system which became action-capable only seven months ago isn't going to become dominant overnight, particularly since its lens collection is still being built up. 15 years ago, most of them shot film. Digital was new. Individuals and companies only move systems when there is a good reason to do so, and only when gear is up for replacement. Gear lasts for years. An ideal system built from the ground up now is not going to be the same as an ideal system for someone who is already heavily invested in gear which would have to be switched out if they were to move to a new system.

4. Demonstrate a lack of understanding of technology and its implications. 'Not the same as an SLR' doesn't mean that it's worse. Some advances in technology involve a tradeoff. AF lenses generally have short throws and are more difficult to manual focus than pure MF lenses. But the AF is so useful that most lenses are designed that way anyway. Current OVFs are nowhere near as good as old ground glass OVFs for manual focusing. But they're much brighter, you can actually see an image in them when shooting an f/4 or slower lens, and, with autofocus, you don't need to MF as often anyway - and, when you do, there's always that big, bright LCD on the back of the camera, which is more accurate than any ground glass. So it is with new technologies associated with mirrorless designs.

[shadowblade]

And on the original point, rumors about a D5s to be announced soon are starting to surface.

Haven't heard any credible rumours yet. Just the usual, 'Well, there was a D4s, and a D3s, so there will probably also be a D5s.'

What reason does Nikon have to release a D5s? What aspect of the D5 would it improve on? what would it achieve for the company, other than making a (now cheaper) second-hand D5 much more attractive to buy than a brand-new D5s?

A new sensor? Hardly - on past record, that would be a D6.

A new AF system? So that it can go from nailing focus quickly every time to... nailing focus quickly every time? When you're already hitting your target just about every time, there's not much to improve, other than AI-based focus modes (which require a through-the-sensor approach for the most part, and are therefore limited in SLR cameras, which only have a low-resolution metering sensor to work from).

Better high-ISO performance? Probably not without a new sensor.

Better low-ISO performance (the D5 being famously weak here compared to its Canon and Sony equivalents)? Not likely, since this is due to A/D conversion, which, in modern sensors, is part of the sensor itself.

Higher frame rate? Not likely with a mirrored design.

Better video? Perhaps. But hardly worth a new $5500 camera, unless you primarily shoot video. And, if your primary purpose is to shoot video, you're probably not using the D5 anyway. And it would still have the same limitations on video AF (with the mirror tucked out of the way), since the sensor has neither on-sensor PDAF nor dual-pixel AF.

Extra gimmicks such as pixel shift and automatic stacking? If you're doing that, you're probably shooting a D850 or A7r3 anyway. And it's probably not worth $5500.

This time, it would be very hard for Nikon to come up with $5500 worth of improvements that would justify an upgrade from most users, without making changes that would be more in line with a D6 model (with a four-year release cycle). Any D5s is unlikely to be a material improvement on the D5, although they would probably try to spin it as such in advertising anyway, should they choose to release such a model.

[bcooter]

Read above post.

I think you basically said what I said, except someone building a small go pro sized 4k camera.  Yes Go Pro is small and 4k, but 4k at a low bit rate, low bit depth is much different than a cinema camera.  If go pro has anything going for it, they have the best auto exposure I've ever seen.  It's amazing at that.

Arri’s, REDs, Canons, Large Sony’s shoot huge bit rates.  The Canon 1dxII shoots 800mbs, where the Sony A7sII shoots around 100 mbs and there is a huge difference, because I have both.  The Sony will band in skies, or blow highlights by just turning it on.

The Sony A9 shoots the same 100 mbs as the A7sII and it’s not an inexpensive.   There are 4k cameras that shoot low bit rates, do not have adequate cooling, bit depth, higher frame rates, adjustable compression, dedicated conversion suites.

4k is not a magic number.   Up to a few years ago Arri's were shot for 70% of Hollywood feature movies and they were 2.7k open gate and nobody complained. 

In fact films like the Martian were shot with over 4k RED's, I think 6k, though they did all the post effect work in 2k because of the cost of 4k effects.  On a large cinema screen you have to be on the front row to notice any difference between a 2k and 4k dcp.

I have REDs, the 1dxII and the Sony a7sII.  The Sony we’ve only used once and it’s suppose to be the low light king, though mine isn’t.  With mine anything above 800 iso goes ragged, in fact I tested it next to the older 5d2 and it didn’t get close and would never do this.

from 5d2 screen grab from 1920 x 1080 motion file:


If any of these companies could build a cinema quality camera the size of your palm then they would do it, because there is a market for smaller cinema quality cameras.

We shoot most projects in still and motion and it’s easier to have a second camera, lens mounted set up for stills than it is to take a motion centric camera, even a dslr like the 1dxII, change focus settings, remove the nds, usually change ISO and shutter, f-stops.

Also less chance of messing up when you go back to motion settings. 

Most people don’t but I like the 1dxII form factor for some motion projects, as  It mounts easily in small areas and takes much smaller car mounts, steady supports and two batteries will last more than a long shoot day.

Regardless, I have mirrorless cameras, don’t dislike them, though my favorites are the gh series from panasonic.    The gh4 for stills the gh5 for motion are very underatted cameras.  The first mirrorless cameras I  bought was the gh3, loved it, found it's menu system and set up was very canon like, so no big learning curve.  The next were the em series Olympus because I just wanted them and use them some, then the a7sII.

IMO

BC

[BernardLanguillier]

This time, it would be very hard for Nikon to come up with $5500 worth of improvements that would justify an upgrade from most users, without making changes that would be more in line with a D6 model (with a four-year release cycle). Any D5s is unlikely to be a material improvement on the D5, although they would probably try to spin it as such in advertising anyway, should they choose to release such a model.

They were facing the exact same situation with the D3 and D4 that were still the best action cameras when the D3s and D4s were released.

It is always possible to improve and that is pretty much what has driven the Japanese economy till date. Companies such as Nikon understand they must change gear to stay alive, but they are not willing to given up on their essentials and kaizen is so deeply rooted in their DNA that they will not take the risk to forget about it.

I know, I work with these companies on a daily basis here in Tokyo.

The key area they could improve on the D5 is on sensor AF and, fortunately, they have the technology ready since they are going to release new mirrorless cameras soon. Why not equip a D5s with that? Coupled with their best in class AF algorithms (that are not related to the type of AF sensors used) they could further extend their technological lead in the action camera segment.

This being said, it seems that the rumors were based on false information, so there may be no D5s, who knows.

Thom Hogan seems to think Nikon should release one and I agree based on a reasonable set of infomation. That doesn't mean they will.

Cheers,
Bernard

[shadowblade]

They were facing the exact same situation with the D3 and D4 that were still the best action cameras when the D3s and D4s were released.

It is always possible to improve and that is pretty much what has driven the Japanese economy till date. Companies such as Nikon understand they must change gear to stay alive, but they are not willing to given up on their essentials and kaizen is so deeply rooted in their DNA that they will not take the risk to forget about it.

I know, I work with these companies on a daily basis here in Tokyo.

The key area they could improve on the D5 is on sensor AF and, fortunately, they have the technology ready since they are going to release new mirrorless cameras soon. Why not equip a D5s with that? Coupled with their best in class AF algorithms (that are not related to the type of AF sensors used) they could further extend their technological lead in the action camera segment.

This being said, it seems that the rumors were based on false information, so there may be no D5s, who knows.

Thom Hogan seems to think Nikon should release one and I agree based on a reasonable set of infomation. That doesn't mean they will.

Cheers,
Bernard

As I said - all the possible significant improvements are on the sensor, not elsewhere. That's a D6, not a D5s.

There's not much they can do off-sensor to improve the D5 in a meaningful way. They can't make it much faster while retaining the mirror and there's not much to improve AF-wise, apart from on-sensor AF for video and live view.

They don't have a technological lead in action cameras. They have a lead in action SLRs (it's unknown whether Canon has caught up), but it's questionable how much further the action SLR concept can be pushed. Any sensor-based improvements are even more applicable to mirrorless cameras, since they have full-time sensor data to work with, whereas SLRs only benefit from features that apply at the moment of exposure, except when working in live view mode (i.e. not functioning as an SLR). There's probably not much they can do outside of the sensor to improve performance. And it's unclear whether they can transfer their proficiency in off-sensor AF into on-sensor AF yet - it took Sony a good few years, and several generations, before they could release a model able to keep up with the 1Dx2 and D5.

Also, what's on-sensor AF going to add to the core function of the D5? Better on-sensor AF only helps in live view mode or when shooting video. But the D5 is an action stills camera, not a video camera. People who primarily shoot video do not own the D5. Hardly anyone uses it in live view mode - if the bulk of your work involves careful composition on a tripod, you're better off with a D850, A7r3 or MF body. D5 shooters mostly shoot handheld (or monopod-supported) stills, through the viewfinder, at moving subjects. On-sensor AF doesn't come into that at all. And very few people are going to pay for a new $5500 body that doesn't perform its core function any better, but can focus a bit better when shooting the odd video clip.

Of course, if they do improve on-sensor AF significantly - to the point where it becomes competitive with off-sensor AF (i.e. A9-level capability), the question must then be asked - what's the point of the mirror?  If you can focus just as well without the mirror as with it, why bother having a mechanical part that adds complexity, vibration and lag, while slowing down the maximum frame rate? Essentially, you've then gotten yourself a mirrorless action camera.

Essentially, improving the D5 means turning it into more of a mirrorless camera, directing its development along a path that makes greater and greater use of sensor-based capabilities and eventually does away with the mirror entirely.

[Rob C]

I appreciate that you cats enjoy these forum games, but thing is, at the end of the day, cameras are not key: photographer is, always has been and always will be unless the game is reinvented as a hands-off exercise in sterility. That may please some more than somewhat, then LuLa would need a new section to cater for it.

I'd also guess that over 99% of photographers already own cameras beyond their ability to exploit to the maximum - I know that I do.

Rob

[BernardLanguillier]

I appreciate that you cats enjoy these forum games, but thing is, at the end of the day, cameras are not key: photographer is, always has been and always will be unless the game is reinvented as a hands-off exercise in sterility. That may please some more than somewhat, then LuLa would need a new section to cater for it.

I'd also guess that over 99% of photographers already own cameras beyond their ability to exploit to the maximum - I know that I do.

Com'on Rob, let us play.

You are not telling us that cameras are about taking pictures, are you? I thought they were just about showing off and being right in forums?

Cheers,
Bernard

[uaiomex]

Both sides of the coin, this thread has been incredibly informative and most amusing.
Sincerely, thanks

Com'on Rob, let us play.

You are not telling us that cameras are about taking pictures, are you? I thought they were just about showing off and being right in forums?

Cheers,
Bernard

[ErikKaffehr]

Hi Rob,

I would agree with you to some extent. But, the TopGear stuff may actully make a sports shooter a more successfull sports shooter. Twelve or twenty frames a second will catch more peak action five frames a second. An advanced AF system will deliver more sharp images than a less advanced AF system.

Of course, sports photographers were able to get great images before the 12 FPS with continuous AF era.

Personally, I was shooting a lot of show jumping shooting single shot manual focus. That worked, too, but it was a lot of experience and image quality standards may have been less demanding.

Best regards
Erik

I appreciate that you cats enjoy these forum games, but thing is, at the end of the day, cameras are not key: photographer is, always has been and always will be unless the game is reinvented as a hands-off exercise in sterility. That may please some more than somewhat, then LuLa would need a new section to cater for it.

I'd also guess that over 99% of photographers already own cameras beyond their ability to exploit to the maximum - I know that I do.

Rob

[Rob C]

Com'on Rob, let us play.

You are not telling us that cameras are about taking pictures, are you? I thought they were just about showing off and being right in forums?

Cheers,
Bernard

Bernard, I'm not your parent! Of course you can go out and play; just be careful not to talk to strangers and never, never tell them where you live!

Other, naughty boys may be showing off, but you know why they feel compelled to do that, so just ignore them and be happy. Never know what Santa may have in his sack for you this year!

:-)

Rob

[Rob C]

Hi Rob,

I would agree with you to some extent. But, the TopGear stuff may actully make a sports shooter a more successfull sports shooter. Twelve or twenty frames a second will catch more peak action five frames a second. An advanced AF system will deliver more sharp images than a less advanced AF system.

Of course, sports photographers were able to get great images before the 12 FPS with continuous AF era.

Personally, I was shooting a lot of show jumping shooting single shot manual focus. That worked, too, but it was a lot of experience and image quality standards may have been less demanding.

Best regards
Erik

Yes, as you sort of suggest, old equipment didn't feel old in its day, and the newspapers and magazines were flourishing better than they are today. I really believe, all teasing aside, that we have passed a critical point in photography where the images simply do not improve, there are just millions more of them. Something magical has vanished. It's the same with fashion and advertising: there's now an ennui in the zeitgeist; nothing surprises anymore. What might have been thought technically amazing twenty years ago now is just the daily norm. Nobody gives a shit about that "special" shot today because almost everybody can make it.

That, for me, sort of sums up where we are at in this business today. When everything is possible, nothing is special.