To go mirrorless (or not)

Recently there has been the launch of the new Nikon mirrorless DSLR system, and also the new Canon EOS R mirrorless system, that gives us the opportunity to review where we’ve come and where we’re going. Obviously I have taken my own experience and camera ownership, but many of us have had a similar journey but with different marques/models.

One thing I am certain of, is that the new flagship Z7, and Canon EOS R, have much farther to go – I think there will be a significantly improved in the near future that will be more tuned to the working pro – especially in its AF solution. If there is a weakness in the new Z7 it is in its AF interface for those myriads of experienced Nikon users, despite it being a small improvement on the Sony version. But that is another story. Nikon generally uses Sony sensors (as does Olympus) but seem to make better use of the available resolution.

Indeed, one could say that all affordable digital cameras were mirrorless for many years, but the DSLR eventually won through. Mirrorless has been seen as the ‘alternative’ way – some would say is still the case.

I started with film in the 1960s. I had folding Kodak and Voigtlander folders (8 on 120 film) and often used Dad’s Yashica-Mat and Rolleiflex derived machines (12 on 120). Even today, looking in to the viewfinder of a twin lens reflex body is a wonderful experience, even if there can be left/right conundrums. I also used Russian rangefinder cameras but I wanted the lens interchangeability of an SLR.

Enter the Zenith E, then a Praktica LLC, both 42mm screw thread mounts. The LLC even had electrical contacts that gave it open aperture metering; I really wanted a Pentax ME Super but money was in short supply. I still have a Praktica BC1 – plus a bayonet mount that had adapters to mount my legacy screw thread optics.

My journey:

Film            1960s                     Kodak folder

Ensign Selfix

Yashica Mat TLR

Rangefinder Werra        

1970s                     Manual SLR

Zenit, Praktica LLC

1980s                     Praktica BC1     

 

Digital               1999                  Compact

Olympus c820l

Olympus c2000z

Olympus c5050z

2005                   DSLR

Nikon D70

Fujifilm S3 Pro (Nikon mount)

2010                       Nikon D7100

2014       Olympus OMD EM-5 then EM-1

2018                       Nikon D800

(I still own those in bold italics)

In the 1990s the mass market produced all-in-ones with inbuilt zoom lenses, but I stuck with my SLR until digital. Expressed differently, the enthusiast/pro lines of TLR merged with the rangefinder/compacts with the new generation of digital. At this time there were all sorts of innovative shapes to cameras, though consumer expectations of what a camera should look like eventually held sway.

In 1999 I had my first digital camera – Olympus c820L – single fixed lens (approx. 35mm equivalent focal length). Auto only, but had lots of flash functions, and incredibly slow playback. LiveView seemed revolutionary. But this was obviously the way we were going, though resolution was a little low – 0.8MP, and shutter lag could be irritating (though nothing like as bad as the best selling Sony Mavica of the day).

In this millennium I have not shot a single frame on film!

Enter the Olympus c2000z – everything many mirrorless DSLRs have now. Aperture/Shutter priority, auto program, zoom 35-105mm equivalent, etc. By 2005 I had the gargantuan 5MP c5050z – all the functions I craved for in film days and much more, but still felt restricted in lens choices. I wanted to go beyond 105mm and, especially, wider. I started fitting wide extenders (from film days) that seemed gargantuan on this small body and was not entirely successful.

2005 was when Canon launched the 300D, the first time DSLRs were affordable by ordinary people, costing less than £1,000. By Christmas I was salivating for one… and the Nikon D70 launched. After handling the two there was no contest, the Nikon felt like a camera that would last decades, the Canon felt cheap plastic, and the Nikon kit lens was far better.

So now I am solidly Nikon from D70, to Fujifilm S3 Pro (used), to Nikon D7100 and now Nikon D800, and lenses that do what I expect them to do.

The Olympus digital experience hasn’t been abandoned – when I want a carry about, always available camera – I use an Olympus OMD EM1 mkii and a few very compact lenses that all fit in a small handbag.

Mirrorless may well be the way we are going…or maybe not, if it becomes viable to merge the two into one compact body! But at present, and from experience with the OMD camera, I do seem to be fighting the sometimes wobbly technology (see other articles on this site) that is not yet absolutely solid. As an example, Olympus have released rather more firmware updates fixing glitches per camera than Nikon (10 for the EM5 vs 3 for the d800 !). The excuse that there is more high-tec in mirrorless is irrelevant until the user experience feels on a completely firm foundation. The underlying technology of the Olympus EM1 is superb – what else could be offered?

Olympus should be praised for improving functionality of the EM-1 through firmware upgrades.

It does seem that there is a new merging of lines, the traditional established SLRs (Nikon, Canon, Pentax etc) to what we currently term mirrorless (Olympus, Sony, Panasonic, Fuji etc). Incidentally, I don’t understand why this is the tenth anniversary of mirrorless, it strikes me that Olympus have made digital ‘SLRs’ (albeit fixed lens with converter lenses such as the c1400L of 1997, or the four thirds E series from 2003) long before.

The digital chart that follows does not illustrate the importance of sensor sizes in each generation. Early digital cameras sometimes had very small sensors, but then the APS-C (DX) became prevalent, with the four thirds and now micro four thirds (MFT) joining in. Full frame was initially out of reach, but is now the sensor of choice that is capable of giving us approaching medium frame levels of resolution.

For example:

Olympus c820l:                  0.81 MP               4.8 x 3.6mm                                       4.67

Olympus c2000z:               1.9 MP                  6.4 x 4.8mm                                       6.16

Olympus c5050z:               4.9 MP                  7.11 x 5.33mm                                   12.88

Olympus OMD EM1i:       16.1 MP:              17.3 x 13mm (MFT)                         7.15

Nikon d70:                           6.1 MP                  23.7 x 15.6mm (DX)                        1.65

Fujifilm s3 pro:                 6.1 MP x2            23.7 x 15.6mm                                   1.67

Nikon d7100:                      24.1 MP               23.7 x 15.6mm                                   6.59

Nikon d800:                        36.3 MP               35.9 x 24.0mm (FX)                         4.22

The progression as shown is thus not just in resolution in megapixels, but also in size of sensor. Pixel density is shown on the right. There are two big jumps; firstly from the small sensor sizes prevalent from the beginning od digital cameras until many of today’s compact systems; then to the popular DX or APS-C sensors of earlier and most popular DSLRs of today; and thirdly the FX (full frame) that are today’s advanced and professional cameras. Today’s d800 has a huge sensor compared to the earlier Olympus cameras, but also a lower pixel density – no wonder it performs so well!

Newer lens models are not only cheaper to produce by using moulded glass rather than cut glass, but new technologies create different trade-offs. With mirrorless mounts, the flange to sensor distance shrinks and has benefits in design, but the smaller lens bodies seem to be more prone to distortion. One might say that modern camera electronics correct this, but this at once negates the benefits of those ultra high resolution sensors.

The new Nikon Z seems to be a further trade off- a little bit smaller and a little bit more resolution, but a slight worsening of pixel density.

Conclusions from my own (brief) handling of the Nikon Z6 was that it is lighter (but only if using Z lenses, otherwise existing gear requires an adapter, negating this advantage. The viewfinder seemed very good, but migration of buttons was confusing. To express some caution, it is noteworthy that the recent Z series 50mm lens is both bigger and heavier than its predecessors… where is the gain?

In the future, we may see the demise of APS-C or MFT… or not, because they are both cheaper to produce when in a price sensitive market, and are each little brother to the up-market full frame gear whether mirrorless or otherwise.

Recently, I am coming to the conclusion that MFT will be the format of choice, because smaller cameras/lenses are de rigeur. The AI technology underpinning modern smart phone cameras will migrate, and then we won’t be needing high resolution tanks!

For me, I don’t want to jump completely just yet. From what I have read so far, I would be very wary of changing the superb AF of current cameras to the less well developed implementation of the new Nikon Z,  even though it is said to be superior to that of Sony. There is also a need to carry more batteries and more recharging, and the presence of only one memory card socket seems to be a huge discrepancy for the semi and pro user. As I said at the start of this article, for these reasons alone I think the Z7 seems to be precursor of another pro camera that is still in the wings.

Nikon has always been innovative, and has rarely produced a product less than well made or technically sound (Nikon 1 excepted).

I am awaiting the reviews, a look and feel of the real thing, and see how it stands a few months from now.

Just to add to the proposition, my most recent shoot was at Maiden Castle near Dorchester when I used my D800 in rather poor light in dull conditions. I was confident of my photography and I did not use the rear screen to view shots at any time until I got back to base. (I have a sunshade fitted, that I left closed all the time.) On this basis I could easily go mirrorless.

But I do not have the need to change my gear yet… I think in 10 years time I might wish to carry something lighter (I will have upgraded the Olympus EM-1 by then) and will no longer aspire to selling shots. Probably. But for now, I am happy.


It seems most people regard resolution (ie number of pixels on the digital sensor) to determine the quality of the digital camera. But, provided the image is not cropped significantly most sensors provide enough resolution in today’s world. Indeed, the size of the sensor itself can have a huge impact on overall picture quality.

Resolution is perhaps easiest to understand – a simple function of the number of pixels on the camera sensor.

It is the dynamic range capable of being resolved by the sensor that has an enormous bearing on the quality of an image – thus, pixel density has huge importance in conjunction with actual resolution.

Related to this is the signal to noise ratio the camera generates. This can be represented by the hum old audio amplifiers produce when turned up loud. The hum is hardly appreciable at normal and low sound levels so that we hear the music clearly, but as we increase the volume/amplitude we may hear increasing hum and/or background noise. The music is likely to be clearer when there is less hum interfering – it is the difference between the reproduction of the music and the background hum/noise that we can call a low signal to noise ratio.

With a higher density of pixels on a camera sensor there is likely to be a consequent reduction in signal to noise degrading the image. Additionally, If pixels are very close together it might also be the case that light ‘bleeds’ between pixels, further confusing the electrical signals that is desirable in a ‘clean’ image. It is these parameters that create what we perceive as ‘noise’ in photographs, impacting on final image quality.

Less appreciated is the improvements to image processing power both within cameras as well as developments in software development.

This article looks at several cameras owned over the last 20 years, and considers their relative resolution, pixel density and signal to noise ratios.

Cameras to be compared are:

Camera                                   Diagonal         Surface           Pixel   

                                                                        Area                pitch   area     density

                                                 mm                  mm²                µm       µm²     MP/cm²       

1) Olympus C-820L             6.00                 17.3                 4.63     21.44   4.67

2) Olympus c2000z            8.00                30.7                4.03    16.24  6.16

3) Olympus c5050z              8.89                 37.9                 2.79     7.78     12.88

4) Nikon d70                          28.37               369.7               7.78     60.53   1.65

5) Fujifilm S3pro                  28.21               366.6               7.74     59.91   1.67

6) Nikon d7100                      28.21               366.6               3.9       15.21   6.59

7) Olympus OM-D E-M5      21.64               224.9               3.74     13.99   7.15

8) Nikon d800                        43.18               861.6               4.87     23.72   4.22

9) Olympus OM-D E-M1      21.64               224.9               3.72     13.84   7.24

Generally, the bigger (and newer) the sensor, pixel pitch and photosite area, and the smaller the pixel density, the ‘better’ the camera.

In the early years of digital cameras, the optical crop factor had much greater importance.

Some further observations:

  1. The D800 has the largest sensor and one of the largest in terms of pixel pitch. The density is still one of the smallest, and hence has a deserved reputation for image quality (IQ).
  2. The E-M1 has a similar density to the d7100, but with a slightly smaller pixel pitch creates excellent images for the sensor size.
  3. The S3 pro has an extremely low density, though its novel use of a two pixels per photo site has meant that it is capable of excellent dynamic range.
  4. The biggest weakness of the early cameras is now shown to be the pixel density so the mass market introduction of DSLRs in 2004 increased IQ significantly.

The graphs produced by dxomark illustrate the various strengths of cameras, but measurements largely reflect sensor sizes.

It seems likely that the move to mirrorless full frame models will not result in a significant change in IQ. It may well be that full frame pixel counts can increase a little more and processing will make further improvements, but possibly not at the rate experienced in the last 20 years. Improvements in signal processing will result in lifting the performance of micro four thirds and APS-C/DX sensors since resolution has less importance if prints of 15.3×11.7 inches can already be produced by a 16mp micro four thirds camera (A3 with a small border).

The DXOmark graphs do show what may seem obvious – the camera with the biggest sensor performs best between the d800 and EM-1 – both these cameras, and especially the d800, display invariance characteristics. The surprise is the performance of the 15 year old S3 pro that even beats the d800 in terms of dynamic range up to an ISO of 800. (A more modern APS-C/DX camera such as the d7100 lies in between the EM-1 and d800 graphs.) Full frame seems to be roughly two stops ahead of MFT.

When comparing the Nikon Z7 mirrorless with the d800 there is little difference in signal to noise ratio at all, and dynamic range is little different – though the d800 seems better at iso 100 and 200. The d800 also beats the new Canon EOS R up to cISO 800.

26/01/2021