PDA

View Full Version : A nice boost for u43 from Sharp


raichea
4th April 2019, 10:03 AM
https://www.olympus-global.com/news/2019/nr01198.html

RobEW
4th April 2019, 11:18 AM
Yes, I read that. Didn't even realise Sharp were in the camera market. Di they mainly do camcorders?

pdk42
4th April 2019, 12:29 PM
8k video = 34Mp minimum. New sensor!

Internaut
5th April 2019, 10:57 PM
Let’s see what kind of sensor Sharp come up with, and how much they’re prepared to sell it for. Like most of the other manufacturers, Olympus don’t make thier own sensors. A spectacular high res sensor, that Olympus can buy, is surely THE dream for any gear headed Olyphile?

raichea
5th April 2019, 11:02 PM
Letís see what kind of sensor Sharp come up with, and how much theyíre prepared to sell it for. Like most of the other manufacturers, Olympus donít make thier own sensors. A spectacular high res sensor, that Olympus can buy, is surely THE dream for any gear headed Olyphile?
I'm quite happy with 20 megapixels - the print size possible with 20MP is plenty good enough for most work. While more pixels would allow tighter cropping, going higher resolution would necessarily increase noise unless there's some big technological breakthrough.

I can see the 8K video guys enjoying the higher resolution though - noise isn't such an issue for them.

drmarkf
6th April 2019, 08:40 AM
Let’s see what kind of sensor Sharp come up with, and how much they’re prepared to sell it for. Like most of the other manufacturers, Olympus don’t make thier own sensors. A spectacular high res sensor, that Olympus can buy, is surely THE dream for any gear headed Olyphile?

Not for this one.

Those doing mainly landscape and studio imaging might want more resolution (and those that do actually print above, say, 12 x 16” might indeed see some resulting improvement...), but those of us shooting things that move in lowish light would happily stick at 20Mp but dream of another stop of signal to noise ratio.

It depends what you shoot. Currently, care in exposure and knowledgable processing is needed especially above ISO1600, and this corresponds to the beginning of the ISO range you need to use the 300 Pro in typical U.K. cloudy conditions and at the beginning and end of the day, while keeping the shutter speed high enough to freeze subject movement. Eg to get critically sharp feather detail in birds in flight you need 1/2000 at least. The latest FF bodies are notably better in this key area, as long as you can manage the cost and weight, while in any decent light the latest m4/3 sensors are perfectly OK.

ISO3200 is about the m4/3 sports and wildlife limit for anything beyond social media, while an extra stop would give improvement where it would really help.

However, I’m not holding my breath - the interwebs are still on a marketing-led Mp kick.

pdk42
6th April 2019, 12:46 PM
Generally speaking, and contrary to many people's understanding, increasing Mp count does not lead to greater noise in the final image - in fact, evidence points in the opposite direction. Per pixel noise is not the important measure it's noise in the final image.

raichea
6th April 2019, 01:24 PM
Generally speaking, and contrary to many people's understanding, increasing Mp count does not lead to greater noise in the final image - in fact, evidence points in the opposite direction. Per pixel noise is not the important measure it's noise in the final image.
At the individual pixel level, I think it's inarguable that noise increases as the size of the cell decreases - your comment suggests you support that, Paul.

Do you have a pointer to any corroborative information for noise in the final image?

pdk42
6th April 2019, 03:38 PM
At the individual pixel level, I think it's inarguable that noise increases as the size of the cell decreases - your comment suggests you support that, Paul.

Do you have a pointer to any corroborative information for noise in the final image?

I think it's probably true that S/N deteriorates as pixel size decreases, all other things being equal. But other things are not equal since there will be complicating factors such as micro-lens efficiency, sensor technology (BSI etc), analogue electronics efficiency, software NR (even in raw files) and other things. It's hard to isolate just pixel count and keep all the others constant. What's generally true is that overall noise performance has improved over the years whilst pixel counts have generally increased too - so if you were to just measure noise performance against Mp count based on real world cameras then I think the curve would show more Mp = lower noise. For instance:

- a7Rii (42Mp) delivers lower noise than A7ii (24Mp)
- E-M5 (16Mp) delivers lower noise than E-P3 (12Mp)
- E-M1ii (20Mp) delivers lower noise than E-M1 (16Mp)

This article is interesting too. Look to the section entitled "Different Sized Pixels in the Same Sized Sensor":

http://www.clarkvision.com/articles/does.pixel.size.matter/

raichea
6th April 2019, 04:23 PM
Thanks - that looks interesting - I'll read it later.

I did add the caveat that technology improvements could mitigate the noise so I shall lean on that as a response to your counter-examples. :)

Internaut
7th April 2019, 09:53 AM
Generally speaking, and contrary to many people's understanding, increasing Mp count does not lead to greater noise in the final image - in fact, evidence points in the opposite direction. Per pixel noise is not the important measure it's noise in the final image.

There are some very good threads on DPReview on this subject. Essentially, it seems more pixels will always mean better IQ but, I think, cramming more and more pixels into the same small area also gives diminishing returns.

Things will start to get rally interesting when global shutter becomes a reality. Imagine, for example, being able to expose for both highlight and shadow by having pixels alternating between two shutter speeds.