bobn2

bobn2

Lives in United Kingdom Worcestershire, United Kingdom
Joined on Aug 28, 2007

Comments

Total: 143, showing: 1 – 20
« First‹ Previous12345Next ›Last »
In reply to:

bobn2: On reverse engineering. The Nikon patent came up in forum discussions, I read through it, and it's interesting. First it specifically patent protects the lenses, not the system as a whole. Looking at the claims, I can't see how any company could make a Z compatible lens without violating the patent (in fact, the mount looks to have been designed specifically to be patentable). That would mean that reverse engineered lenses will not be a possibility, they will have to be licensed by Nikon - which might well choose to do so, but they will keep control of the mount.

Spectro: didn't sigma already win a reverse engineering lawsuit some time ago, yeah.
Yes, but not all 'reverse engineering' is the same. A patent protects new inventions. To bust a patent, you have to prove either that the invention wasn't novel, or that it was obvious. That's what Sigma did when the won the lawsuit. Each patent makes a different set of claims, and in the case of Nikon's, this lens has two separate lens-body communication channels. That's not an obvious thing to do, nor has anyone done it before. The claim is specifically for a lens with that attribute, not a whole system. I don't see myself how a lens manufacturer could break it.
So far as what the Nikon rep said, it could mean specifically that Nikon has licensed it, why not? Or it could mean that they know someones doing it and they'll take them to court.

Link | Posted on Sep 18, 2018 at 07:28 UTC

On reverse engineering. The Nikon patent came up in forum discussions, I read through it, and it's interesting. First it specifically patent protects the lenses, not the system as a whole. Looking at the claims, I can't see how any company could make a Z compatible lens without violating the patent (in fact, the mount looks to have been designed specifically to be patentable). That would mean that reverse engineered lenses will not be a possibility, they will have to be licensed by Nikon - which might well choose to do so, but they will keep control of the mount.

Link | Posted on Sep 17, 2018 at 20:29 UTC as 106th comment | 6 replies
In reply to:

Photomonkey: I will be launching a Kickstarter funded lens featuring no glass, solid metal build shortly.
User needs to supply mount of choice. Some machining needed. $12,000.

That's a great idea. It's big advantage is that it will perform just the same, whatever the light level.

Link | Posted on Sep 11, 2018 at 17:45 UTC

They should have called it the Theresoneborneveryminutar.

Link | Posted on Sep 11, 2018 at 17:43 UTC as 99th comment
In reply to:

JTWayt: On my Sony a7Rii, ISO below 100, like 50, is called fake. It is only going to affect captures to JPEG, where it will reduce the values by 1 stop. What doesn't change is the effect on RAW.

This is instructive. It means the exposure time would be doubled, and over expose the highlights, potentially blowing them. So the article is discussing the similar issue with ISOs greater than base.

When you say it's called 'fake', who is it doing the calling. Not Sony, I think.

Link | Posted on Aug 13, 2018 at 17:56 UTC
In reply to:

FrithjofA: The article addresses an old topic that has been much better an clearer presented on Roger N. Clark website, the standard reference site on the web for excellent information on digital sensors since 1998. You can read there for years the statement:

"ISO is simply a post sensor gain followed by digitization. ISO settings are needed mainly to compensate for inadequate dynamic range of down stream electronics. One could specify ISO such that downstream electronics digitize the full range of signal up to the full well capacity. "

You also find much more in depth information about the noise sources of images.

The idea that ISO is "gain" is fundamentally wrong, since "gain" implies amplification of a value, whereas what is happening in a camera is a transformation between different kinds of things. The input is exposure, the output, 'lightness', is grey scale. There is no direct connection apart from that determined by ISO. Intrinsically, there is no 'gain' required at all.

Link | Posted on Aug 10, 2018 at 20:35 UTC
In reply to:

John Koch: RB: "ISO changes the lightness of the final image but it doesn’t change the fundamental sensitivity of your sensor."

How does it raise the lightness without greater sensitivity to light, other things being equal?

RB: "Nor is it an indicator of amplification being applied: although many cameras do increase their amplification as you increase the ISO setting, this isn’t always the case."

This entails that the prior sentence may be wrong most, if not all of the time. Amplification and increased sensitivity seem complementary, or the distinction is cloudy. And can one know a camera's ISO amplifies or not? DPR's reviews don't seem to disclose one way or the other. The manufacturers are also rather tight lipped about their sensor technology. All we see are the ISO ranges and some noise comparisons.

The article does not explain why some cameras' high ISO performance is better than others. What is the difference between the Sony A7s series and prior or competing devices?

Think of what a camera does as painting by numbers. The numbers in the raw file are exposure values, pixel by pixel. The camera chooses a shade of paint according to each number in the raw file. ISO chooses which paints correspond to which numbers. Increase the ISO and you lighten the paints. No change in 'sensitivity' required.

Link | Posted on Aug 9, 2018 at 07:02 UTC
In reply to:

Tom Holly: I like the ISO speed concept, it’s a useful reference, but one thing I agree with is ditching ISO nomenclature when it doesn’t actually reflect a genuine increase in base sensitivity. Eg If ISO 800-12800 are all deriving from the same initial base sensitivity/amplification then just call it 800 and tag it with subsequent “push”. Same as with film, where you denote +1, +2 etc to denote a post-exposure push.

I would much rather know sensor is exposing at 800 speed and subsequently pushing things by 3 stops than be told it is exposing at 6400

One problem is that in digital there is nothing similar to the film 'speed point' which could be used to define an absolute ISO for a sensor.

Link | Posted on Aug 8, 2018 at 11:16 UTC
In reply to:

TFD: SHUTTER SPEED.
One topic that I think deserves coverage is Shutter speed. Film relied on exposure to an amount of light (photons) which was determined by the aperture & shutter speed in which there is a reciprocal relationship

CCD & CMOS are different CCD charge a capacitor integrating time & qty. CMOS use a photodiode.

Exposing film results in changing the sliver halide crystals, the larger number of photons hitting the film (time & qty) the greater the change in the crystals. Digital sensors I don't think work the same way. The electrical signal (voltage) generated by a CMOS pixel in response to be exposed to light is the same regardless if you expose it for 1 sec or 1/1000 of a second. Changing the aperture will vary the number of photons & the voltage generated but changing the shutter speed just lets the camera read the signal for a longer period of time - If you generate the same voltage during a 1/1000 as you do in a 1 sec exposure - why the longer exposure.

"Digital sensors I don't think work the same way. The electrical signal (voltage) generated by a CMOS pixel in response to be exposed to light is the same regardless if you expose it for 1 sec or 1/1000 of a second."
Not the case, the charge accumulated on depends on the number of photos, it doesn't matter what he that number was produced by a long exposure time or a wide aperture. The equivalence of exposure time and aperture with respect to exposure is a basic tenet of sensitometry, and applies to all photographic media.

Link | Posted on Aug 8, 2018 at 11:14 UTC
In reply to:

TFD: SHUTTER SPEED.
One topic that I think deserves coverage is Shutter speed. Film relied on exposure to an amount of light (photons) which was determined by the aperture & shutter speed in which there is a reciprocal relationship

CCD & CMOS are different CCD charge a capacitor integrating time & qty. CMOS use a photodiode.

Exposing film results in changing the sliver halide crystals, the larger number of photons hitting the film (time & qty) the greater the change in the crystals. Digital sensors I don't think work the same way. The electrical signal (voltage) generated by a CMOS pixel in response to be exposed to light is the same regardless if you expose it for 1 sec or 1/1000 of a second. Changing the aperture will vary the number of photons & the voltage generated but changing the shutter speed just lets the camera read the signal for a longer period of time - If you generate the same voltage during a 1/1000 as you do in a 1 sec exposure - why the longer exposure.

"Exposing film results in changing the sliver halide crystals, the larger number of photons hitting the film (time & qty) the greater the change in the crystals."
Film uses the photoelectric effect just like digital. The released photoelectrons reduce silver halide to silver, and the silver atoms form a catalyst which allows the developer to reduce the whole grain. It's an all-or-nothing digital style effect, if there are enough silver atoms, the whole grain gets reduced, if there aren't, it doesn't. The density reflects the proportion of the total number of grains that have been reduced to silver, not the proportion of each grain that has. In that sense film is more 'digital' than digital.

Link | Posted on Aug 8, 2018 at 11:14 UTC
In reply to:

TFD: SHUTTER SPEED.
One topic that I think deserves coverage is Shutter speed. Film relied on exposure to an amount of light (photons) which was determined by the aperture & shutter speed in which there is a reciprocal relationship

CCD & CMOS are different CCD charge a capacitor integrating time & qty. CMOS use a photodiode.

Exposing film results in changing the sliver halide crystals, the larger number of photons hitting the film (time & qty) the greater the change in the crystals. Digital sensors I don't think work the same way. The electrical signal (voltage) generated by a CMOS pixel in response to be exposed to light is the same regardless if you expose it for 1 sec or 1/1000 of a second. Changing the aperture will vary the number of photons & the voltage generated but changing the shutter speed just lets the camera read the signal for a longer period of time - If you generate the same voltage during a 1/1000 as you do in a 1 sec exposure - why the longer exposure.

"CCD & CMOS are different CCD charge a capacitor integrating time & qty. CMOS use a photodiode".
Not really. Both CCD and CMOS use a 'photodiode' to collect charge resulting from the photoelectric effect and both use a MOSFET source follower to read that charge. The difference is how the charge gets to the source follower. In a CCD, it is shuffled along a row of pixels by moving the potential well in which it's trapped. In a CMOS there is a source follower in every pixel (or sometimes shared by a group.

Link | Posted on Aug 8, 2018 at 11:13 UTC
In reply to:

Bill Ferris: Thank you for this article. What I have found most challenging about the function of ISO if that it sets the lightness of an image but not through amplification. As a non-technical person, amplification was a handy way to think about ISO...until I learned that not how ISO sets lightness.

Oh well, rather than trying to find a new but effective layperson's term to describe ISO's function, I focus now on the consequences of using a shutter speed and f-stop combination the captures so little light as to require a high ISO to achieve acceptable image lightness. This back door approach is easier for me to grow.

@xPhoenix
"How is the image made lighter if nothing is boosted or amplified?"
The image is never 'made lighter'. The raw image has no lightness, it is simply a set of exposure values. How much light comes from a white surface? You don't know, it depends on the lighting. 'Lightness' is purely perceptual. So, there is never a case that an exposure value needs to be 'lightened'. All that needs to be done is decide which exposure value means 'white', which means 'black' and what are the values in between. It's like painting by numbers, and you don't need 'amplification' to paint by numbers.

Link | Posted on Aug 7, 2018 at 19:27 UTC
In reply to:

fyngyrz: TO the author:

Turning up the ISO and getting more noise from the camera's gain stages is like turning up your stereo and getting more hiss from the audio gain stages. Not hum from the power supply or external inductive pickup.

Please, if you're going to fling comparative metaphors around, make sure you have some idea what you're talking about first.

Also, certainly you can get more noise when there are amplifier stages that generate noise, and gain is applied to their output. Any remarks to the contrary are outright wrong.

These days, a well designed amplifier has an SNR of 80dB or more. You won't hear hiss from the amp. The hiss you hear is very similar to what you see in a high ISO photo - shot noise, examined closely enough to be perceptible (whether audibly or visually).

Link | Posted on Aug 7, 2018 at 19:11 UTC
In reply to:

jonas ar: Thanks bobn2!!!!

I'm glad some do! There are others who distinctly don't like their icons being clasted.

Link | Posted on Aug 6, 2018 at 16:05 UTC
In reply to:

jonas ar: Thanks bobn2!!!!

The article was all Richard. I just had a look over (Richard probably thought that if I didn't, I'd be posting quibbles here!)

Link | Posted on Aug 6, 2018 at 14:49 UTC
In reply to:

kodacolor: The higher the ISO number the more noise appears in a photo, same as film.

@Mdayonline
"ISO has to do with some of the processing that is done (amount of light amplification) to the signals from the sensor once image\data capture is performed"
There is no 'light amplification' performed in any camera. The misconception is common, and fuelled by a faulty model that what comes out of a camera (or what is specified in the JPEG file) is 'light', it isn't.

Link | Posted on Aug 6, 2018 at 14:15 UTC

Presumably, and reading between the lines of the press release, this is a feature of the Fujifilm CFA film (which most sensor manufacturers use) - as well as the Bayer array, they can print a barrier, thus no longer requiring the metal that Samsung was previously used. I suspect that this should be read more as a press release from Fujifilm the CFA provider, not necessarily as any pre-announcement that Fujifilm the camera company will start using Samsung sensors (though it also doesn't rule that out).

Link | Posted on Jun 27, 2018 at 23:40 UTC as 22nd comment | 1 reply

I wonder how they persuade all those photons to turn the corners to take those nice wiggly paths.

Link | Posted on Jun 27, 2018 at 23:35 UTC as 23rd comment | 3 replies
On article Nikon D850 sensor confirmed as Sony-made (579 comments in total)
In reply to:

Mortal Lion: Hardly a surprise. Sony is the leading sensor making company. Where else would they go?

Your list is a bit old. Sure there are plenty of sensor manufacturers, very few (one, so far as I know) with the ability to make a commercial volume/price BSI FF sensor.
From your list, Aptina is now part of On semiconductor, which seems no longer to be pursuing high volume sensors with any vigour, Toshiba's sensor division was bought by Sony (as Mortal Lion said)

Link | Posted on Jun 16, 2018 at 11:55 UTC
In reply to:

matthew saville: I'm gonna keep calling it compression, thank you very much.

Just please stop calling different ISOs "different sensitivities", OK? It's gain, people...

:-P

It's not even gain...

Link | Posted on May 24, 2018 at 07:15 UTC
Total: 143, showing: 1 – 20
« First‹ Previous12345Next ›Last »