From: Ray Fischer on 20 Jul 2008 14:28
David J Taylor <david-taylor(a)blueyonder.neither-this-bit.nor-this-bit.co.uk> wrote:
>Ray Fischer wrote:
>> David J Taylor
>> <david-taylor(a)blueyonder.neither-this-bit.nor-this-bit.co.uk> wrote:
>>> Ray Fischer wrote:
>>>> Who cares? People don't care about "pixel densities". People care
>>>> abut noise per pixel, not noise per sensor area.
>>> So why buy a 12MP APS-C DSLR over a 6MP APS-C DSLR if people don't
>>> care about pixel densities?
>Which suggests that you /do/ care about the pixel density....
No, it says that they care about the number of pixels. Notice that you
didn't ask about pixel density? You asked about number of pixels?
That should have been a clue.
From: Bob Newman on 20 Jul 2008 16:37
On 20 Jul, 18:10, ejmartin <ejm_60...(a)yahoo.com> wrote:
> On Jul 20, 11:30 am, Bob Newman <bob.csx...(a)gmail.com> wrote:
> > I think currently realised is the issue at the moment, both for sensor
> > technology and signal processing chain. In the best of all possible
> > worlds many things might be possible on both fronts. EF certainly
> > seems to have something up his sleeve, and from his slide show, it
> > seems to include really tiny pixels and really big DR. In the limit,
> > when you get to a true digital sensor, when each pixel has a FWC of
> > 1e, read noise ceases to be an issue. Near that limit, if each pixel
> > has a FWC of 2e, it's not much of an issue. This is another thing that
> > makes me think that you and Roger are not right, fundamentally, on
> > this. Somewhere between here and there, there would need to be a
> > turning point when the read noise issue stopped getting worse and
> > started getting better. In fact, I'm beginning to think I could mount
> > an inductive proof that you are wrong.> There may be in some hoped-for future a means of lowering the small
> > > pixel read noise to about 1 electron (input referred), which is not
> > > simultaneously available for bigger pixels; perhaps the reason will be
> > > the sort of capacitance arguments you have put forth. At that point,
> > > small pixel DR on a per area basis will equal that of the 1D3's fully
> > > realized sensor DR, and small pixels will be competitive on SNR and
> > > DR. But there is no such pixel like that among current examples.
> > John and I would say, because no-one has bothered to develop it,
> > because it lies so far off the accepted orthodoxy of camera design. I
> > think that's the way Eric's going, though.
> I just remembered, while we are waiting for production small-pixel
> CMOS sensors from Canon, there is one further data point: that 52MP,
> APS-H sized sensor that they made a prototype of:
> 3.3µ pixels, with 5.5 electrons of read noise.
With a column gain of 3, look at figure 5. Who knows how that compares
with the figures obtained using the various amateur testing methods.
> Still not getting
> smaller in proportion to pixel pitch (in fact, a bit worse; granted,
> it's preproduction, but if it were easy to beat down the read noise by
> making the pixels smaller, shouldn't they have been able to at least
> match the performance of pixels with 4 or more times the area?).
Who knows also what the design goals were? There are reasons to keep
the cell capacitance high, which will increase electron referred read
noise but will also increase FWC and therefore DR. Read noise is a
parameter the designer can trade off against other things, unless you
know what those tradeoffs were, you can't draw any hard conclusions.
From: Roger N. Clark (change username to rnclark) on 20 Jul 2008 16:43
John Sheehy wrote:
> There is nothing intrinsically
> significant about a 1:1 SNR,
I agree. It is a convenient standard.
Steve: see figure 5 on at:
for some examples of sub-photon per pixel imaging.
Multiple pixels beat provide multiple sampling that allows
one to see detail below S/N = 1.
> The more the so-called noise floor is, and/or the finer its grain
> relative to subject detail, the better you can see through it.
A simpler way to put it is more sampling with more pixels.
> I was not demonstrating the cameras; I was BORROWING their
> pixels for a demonstration of the effects of pixel density. It has
> *NOTHING*, I repeat *NOTHING* to do with the cameras, except for the read
> and shot noise of their individual pixels and how they image per unit of
But that's the problem. Your demonstration is limited by the electronics
of THOSE cameras, and in particular a limitation of the low ISO
of the DSLR which has little to do with the sensor which you
exploit to drive an agenda. You then use that conclusion
as a general property. Instead you conclusion should be
much more qualified. If that limitation of the post sensor electronics
is fixed in future generations of cameras, you will find your general
conclusions invalid. That is what people have problems with.
Also, if you looked at the bigger picture, you would see there are
several other factors against your sweeping conclusions. For example,
many may not want a gigapixel APS-C sensor because they can't afford
the lenses to take advantage of it, and they don't want to wait
20 minutes for the readout!
From: Roger N. Clark (change username to rnclark) on 20 Jul 2008 17:23
John Sheehy wrote:
> Why don't you just admit it; a DSLR full of Panasonic nMOS could
> outperform your beloved big pixels at everything but ISO 1600 read noise
> (which would still have the lower resolution in the DSLR, of course).
Try looking at the big picture. For example, high dynamic range
is not needed in every photo, and often high dynamic range means
poor light, and unnatural pictures when you try and compress a
high dynamic range into something that can be printed. From a
beautiful light and photographic composition viewpoint, I would
argue that current large pixel sensors, even at ISO100 with the
electronics limitations is more than adequate for outstanding
images. Then there is a need for fast readout speed of the
Images such as these require fast readout, fast cycle time.
If a panasonic class pixel 100+ megapixel, 1.3x crop
camera came out with the current readout rates,
I would not choose it as it would be too slow.
I would choose a ~20 megapixel full frame sensor if it
had 8+ frames per second.
Resolution and dynamic range is not everything. Everything is a compromise,
and all compromises must be considered in designing and choosing a camera.
So admit it, your sweeping statement above is false.
From: Roger N. Clark (change username to rnclark) on 20 Jul 2008 17:35
> Note also, that if one decreases pixel pitch at fixed aperture,
> resolution does not completely saturate at the Airy disk size, as I
> pointed out in a couple of recent posts elsewhere:
> There are still resolution gains in the diffraction limited regime,
> they are just somewhat less than proportional to the reduction in
> pixel pitch. Again, when figuring SNR at the resolution scale, one
> should bin the pixels within that scale size in order to determine
I agree and there are several places on my web site where I discuss that
very issue. And it is not just in resolving line pairs. Stars, for example
look better when there are several samples across the diffraction disk.
> Yes, I was the one who pointed that out to you :)
I didn't recognize your email and you never signed you name.
>> So let's see a real analysis with all the data presented before concluding
>> the FZ50 is out of the box. Also, be aware that not all "raw" data from
>> cameras is actually raw data from the sensor. Perhaps the raw FZ50
>> data files have had processing.
> I agree. It would be nice if John would publish in one place the
> results of all the tests he has done. I should say that I get similar
> numbers for my LX1: about 4100 electrons full well at ISO 100, with
> 2.2� pixels.
That would be an electron density of 4100/2.2^2 = 847 e/sq micron.
If we take off some area for inter-pixel dead space, we get something
like 4100/1.2^2 = 2800 e/sq micron, both numbers bounding typical
ranges. But 4800 e full well in a 1.97 micron pixel (John's FZ50)
is outside the range of number I've seen from manufacturers. Maybe
it can be done, but I've not seen real numbers that high.
> I haven't seen any evidence of NR of the raw data that
> would contaminate the results, but then again I haven't done as
> exhaustive analysis of that as I have on DSLR's such as the D300 and
There are only a handful of people who actually publish camera
sensor analyses, including you and I. I have never seen John publish
a full analysis.