Søk i denne bloggen og Lenker

onsdag 3. august 2016

Will We Ever Get Full Frame Camera Performance in Small Compact Cameras, Even Mobile Phone Form?

Will we ever be able to slip some gizmo out of our shirt pocket and take images akin to those on a Canon 1D or a Nikon 5D ?

The answer is a resounding yes yet an indifatigable no.

No of course not, you cannae defy the laws of optics jim-there are two or so big arguements as to why. In terms of 'real' shallow depth of field, optically this just cannot be delivered by current lens technology on very much smaller cameras and sensors. Secondly the low density of pixel elements allowed for on full frame sensor means a far higher signal to noise ratio per pixel - area increase is on the square so the difference over APS-C, mFT and 1" is bigger than first meets the eye so to speak. This means better dynamic range, better low light performance and vis a vis better high ISO image quality.

However this is all given, precaveated, disclaimer with simple, single lens, esatblished sensor technologies. What can we expect then from technologies in smaller packages?

We also have to accept on the one side how important the "waf-ferr thin" DOF and ultra high ISO performance are to image representation and dessemination today. ISOs of 6400 negate the need for flash photography per se. Beyond that, the mega tens of thousands of ISO points are getting to freezing a frantic blackncat having a fit in a dark room. Outside some special applications and cameras with super fast shutter speeds, these features are my one is bigger and better than yours. They are academic, the returns on investment are at the thinner end of the diminishing.

I would say that what we actually 'need' for taking dslr quality photography from full frame are roughly using some bench marks-

* f2.8 or even f3.5  type of speed and depth of field
*D7000 dynamic range
*  ISO 6400 with little notable noise and only mild smoothing
* 12 - 20 mpx depending on the above vs size of sensor
* tonal depth and colour capture like the D3 and 5dII

Is any of this achievable in a compact camera or even a mobile ? We are really in the realms of not doing the same things but in miniature for the large part but rather emulating the final results with technologies.

Of course mobiles are stealing a march on compact cameras. For years they have been eroding the necessity for shirt pocket compacts and have hit trad' camera brand-sales, which often carried a nonsense premium price mark, out of the ball park. Now they are poised to go up a gear and not only challenge 'enthusiast compacts' but rival mirrorless and even DSLR territory.

Mobile phone technology has also driven some of the innovation we see in user interface and file handling but more than that, the always on immediacey of social media has changed how we view and  "consume" images, with some pro photo journalists and artists shunning their full frames to take people and street scenes where a dslr or even qaulity compact would create a barrier to the subject and perhaps invite a mugging.

The recency and availability is often the wow factor in social media, and often the professional photographer is out of the picture if you pardon my pun. The famous and daring now whip out a mobile and take a selfie a-top Everest, or traipsing up the red carpet to the Oscars. The image then has a different and perhaps higher value in the subject owning the moment, and most often sharing their emotions in a way which as third person behind another lens may make look contrived as they ask them to turn, or wait for theh flash to charge.

We consume images most often at 72dpi and a maximum size of 1080 wide pixels now. I dare say most images now viewed are no bigger than indeed the screen on the iPhone 6s. Gone are the days of the art litho printed coffee table photo extravaganza books. Image technical quality is reduced to the average smart phone's output, with image impact being far more important in the new, nano attention span on line world of image and sound.

Many say that the best camera you can get, is the one you can take with you and have therefore, all the time when you place yourself in a landscspe, street or event. Using live view screens instead of viewfinders has a kind of immersion in the scene and a causualness brought on by the mobile smart phone revolution.

Not only this of course, but smart phones offer not just picture capture but significant post processing ability. They are image systems and they are integrated to this new God, digital social media. An image has not just recency, but immediacey is a large part of its value. Being able to take multiple shots, inspect and select the best, crop and enhance them at the scene and instantly disseminate them to millions is a kind of modern tearing down of the old Rome - the antiqauted print media controlled by the political demigogues.

The way we take and view images then has changed much of the relevancy of full frame and even mid sized sensor photography to the mass market, in which more images have been made and shared in the last five years than in all of preceeding history. There are though some few superior tehcnical features which pro's and IQ snobs alike desire whcih are only to be fulfilled in their eyes, in FF.

There are two or perhaps three major hurdles for small cameras in equalling the final image output of Full Frame. Firstly optics, and shallow depth of field. Secondly there are two sensor pixel density related features - dynamic range/tonal depth and signal to noise ratio. Optics or rather emulating their shallow dof and telefoto images, are something we will return to and are natural physical barriers to miniaturisation actually working.

However in terms of sensor technology, I believe we will see major improvements in the smaller sensors in terms of dynamic range, colour tonal rendition and noise. HTC have already recognised that masses of dense megapixels is not the way to go, and reduced their mpx chip,  for a camera which produces more realistic images with better DR and better low light performance by lower noise levels.

We have a distance to go too with lens and hardware but that in terms od the single lens is limited unless light-field -lenticular capture becomes practical in pocket-thin devices. Small, flat lenses van be very bright, even sub f1.0 , but are limited to short hyperfocal legnths and very short possibilities for zooming within the form factor of consumer mobiles. Yes we have seen 'Phameras" or "Phonecams" and android on compacts which then offer zoom. However digital crop zooming is also now a better prospwct with the advent of shift-sensor multiple capture quadrupling image size. We also have powerful interpolation which 'repixelates" zoomed areas to recreate a higher mpx count, thus enabling crop zooms via purely software. The stregnth in actually acheiving more flexibility in 'apparent' focal legnth is likely IMHO to be from a combination of these three technologies - small zoom lenses, sensor shift and interpolation. These could offer social media ready results in the range of 22mm to say 300mm with small screen acceptable IQ.

Shallow depth of field is something which can also be achieved with software, as any Photoshop fan au fais with mask-and-blur will tell you. Currently 'fake blur' is still in its automated infancy, with equipment or off camera software selecting foreground or faces based on contrast and preprogrommed recognition. One physical answer is to take two images, one with usual focus and the other with a very short hyperfocal distamce near the lens, thus giving a naturalistic blurred background and the software a base point to work from. A more innovative and potentially successful way is the twin camera approach seen in the HTC One M8 and current Huwaei P9.

Dual lens cameras are almost as old as film photography with both stereo and parallax views being used in 3D imaging and range-finder focusing for more than a century. They have a particular appeal in mobiles when one camera captures something different, such as when they are BW and capture more light information, allowing for both better DR and tonal range in the final, interpolated image output from the colour sourced camera, and also delivering more depth queus to the bokeh software. So far the real world results from the HTC M8 and the Huw'P9 are mixed though. Perhaps the real value will be in a second lenticular camera which captures light field information which can then be used create very much more accurate and aesthetic shallow and long DOF images.

Mobile phones by in large are given user interfaces which are to a low common denomintor, yet i the M9 i owned before and the desire Z i hold onto for the keyboard, there are quite a few settings to play with. They are hidden way in menus and awkward to get to or use multiple times. User interface will be another area which will need to improve if mobiles are to ever rival big old cameras, because the photographer needs control over input and output at their fingertips to get the best captures at least.

When we start talking about combinations of technologies we then also hit another barrier which is a function of the mass market naturw of the mobile smart phone- price. It may be great to appeal to enthusiasts, pros and those wanting a great camera in a shirt pocket, but doing some of the combo's i describe above, costs. Top end mobiles are around €~$ 750. That is around the same as enthusiast compacts, mid range "super zooms" and more than entry level DSLRs and MILCs - which are very good indeed these days.

Software though need not be part of that price barrier equation due to firstly the App' phenomenon and secondly, cloud computing. Image processing does not need to happen in camera, nor at time of capture. This opens the field to very advanced computing which is held propriety and 'micro licensed' for each image proceassed in the cloud or App download. Simple computing tasks which require pure power and memory are ideal for cloud computing,  such as face outline matching which would then lead to near perfect blurred bokeh i  portraits within milli seconds.

I have kind of made my conclusions under way, but to summarise we perhaps won't see physical camera features which add to the cost of top end phones, although the follower manufacturers, LG, Huwaei, HTC and most likely Nokia now, may well explore this as a way of capturing more top-end consumers from Apple and Samsung. Software is a different matter economically speaking, either as installed in camera system, as an App' or in the cloud. The cost is inversely proportionate - 'disponential' in that the initial high cost has miniscule unit costs as volume sales increase.

The converse of my proposition is also going to be true of course, that mobile digital telephony becomes more integrated to full frame digital SLRs.


Ingen kommentarer:

Legg inn en kommentar