[Leica] Leica: 100 Jobs lost

[Howard Ritter]

I hope you’re right about recognition of the merits of tailoring a camera's firmware to the lens(es) used, and of high pixel counts as engineering and algorithmic approaches to reducing small-pixel noise march on.

A few years ago I conducted an informal experiment where I imaged the same scene with cameras of several different sensor sizes, compact pocket 11 MPx P&S to 36 Mpx FF 35s, with the zooms all set to the same actual (not equivalent) 21mm FL as the 21mm WA lenses on the FF bodies. This caused any given detail in the scene to have the same physical extent on each of the sensors. In effect, this made each of the smaller sensors a section of a FF sensor behind a 21mm lens. 

By looking up the physical dimensions of each of the smaller sensors, I was determined what fraction of FF each one was equivalent to. Dividing a sensor’s pixel count by the fraction of FF that it represented gave the pixel count of a FF sensor with the same pixel size. The 11-MPx P&S, with a sensor approximately one-third the linear size of FF (i.e., 12 x 9 mm vs. 36 x 24), had about 1/9 the area of the FF sensor and therefore produced an image that was equivalent to a one-third crop of an image from a 100 MPx FF sensor.

I compared the images at the same image scale in Photoshop and noted that resolution improved steadily from the 24 MPx FF through the smaller sensors to the very smallest, which clearly had the best resolution. My conclusion was that FF sensors will continue to improve in resolution at least until they reach 100 MPx, and possibly beyond. I have not been able to find any discussions of what the practical maximum pixel count for a FF sensor and modern lenses might be. I think it’s going to take a while before a 100-MPx FF sensor with acceptable low-light performance and an acceptable cost is available, but if the Leica Q5 or whatever has one, I’ll get it.

The notional 100-MPx FF sensor would have pixel dimensions of about 12000 x 8000. An exhibition-quality 11 x 17 print, with 300 dpi, needs a 3:2 file size of about 5000 x 3300 pixels, which is 40% of the linear dimensions of a 100-MPx FF sensor. The Q’s 28-mm FL divided by 40% is 70mm. So a Q with such a sensor would produce fine large prints at the equivalent of any FL from 28 to 70 mm, or a very acceptable 200-dpi laptop or tablet screen image, or 8 x 12 print, at the equivalent of up to about 135 mm. It would be truly the equivalent of a FF camera with a 5:1 zoom.

Of course, the Q2, with a 47-MPx sensor, has pixels that are only about 40% bigger than a 100’s would be, so its 28mm lens would be good for 50 or 70-mm equivalents under the above conditions. There are times when I’ve wished for a wider angle than my Sony RX1’s 35mm, more times than I’ve wished I could make a smaller tele-equivalent crop, so I’ve been thinking…


—howard

> On 19Jun, 2019, at 1135, Don Dory via LUG <lug at leica-users.org> wrote:
> 
> Leica is not missing this revolution probably most widely done by Olympus.
> Olympus runs a very tight integration of lens information and camera
> processor manipulating the image to correct lens flaws even in "RAW"
> formats.  This lets them produce smaller less expensive lenses optimized
> for correctable aberrations.  If you look at the Q2 it is doing similar
> things to the image.
> 
> Going off the plantation, I believe that the Q2 is a prototype for future
> Leica products.  Higher pixel counts and tighter integration of lens
> processor.  I think that Leica will profit greatly from this.  I have been
> seduced by the optical quality of the Sony G Master lenses and very
> disappointed by the size required to do this.  Also, the Sigma Art series
> has outstanding image quality at the expense of huge volumes and weight.
> So, if Leica can integrate digital aberration control especially if they do
> the work to integrate older designs then we will pay to code out lenses,
> and pay to acquire new lenses to save our shoulders and backs.
> 
> As an aside, the new sensors coming out with their very impressive high ISO
> performance really eliminate the need for very fast lenses unless you want
> the effect of F1 on your image.  So, the 50 APO Summicron is also a way
> forward for Leica as pixel counts go up and superb lens performance becomes
> very important.