Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
01-06-2014, 08:35 AM   #61
bxf
Site Supporter
bxf's Avatar

Join Date: Jul 2010
Location: Lisbon area
Posts: 1,121
Original Poster
QuoteOriginally posted by jsherman999 Quote
...
Basically: it doesn't matter if the light falls on one big pixel or four smaller pixels in the same area - the same amount of light is being collected per area. To increase the total amount of light captured, you need to increase the area.
.
OK, thanks for that easy-to-understand explanation. BUT, we all know (we do, don't we?) that, other things being equal, increasing the number of pixels in all these P&S over the years often resulted in inferior noise performance. Assuming that that can be taken at face value, the only explanation that comes to mind is that the more pixels you divide a given area into, the more wasted space you inevitably end up with, because, I assume, you can pack the pixels only so much - there has to be some space around them. Would that about sum it up, or is there some other explanation?

01-06-2014, 08:44 AM - 1 Like   #62
Veteran Member




Join Date: Feb 2011
Posts: 4,893
QuoteOriginally posted by bxf Quote
OK, thanks for that easy-to-understand explanation. BUT, we all know (we do, don't we?) that, other things being equal, increasing the number of pixels in all these P&S over the years often resulted in inferior noise performance.
But it didn't, really. Maybe with JPG's it's tougher to get the noise OUT, but there's better SNR overall with the more recent sensors, whether they're binned into 2 million spots or 10 million spots.

QuoteOriginally posted by bxf Quote
the more wasted space you inevitably end up with, because, I assume, you can pack the pixels only so much - there has to be some space around them.
That's absolutely true, but it's fairly marginal. There are tiny lenses on top of the photo sites that attempt to direct the light to the photo sensors. Alternatively (or rather, in conjunction with that) some sensors, called 'back illuminated' use almost all the sensor area for light gathering, and put the wires onto the 'back' (actually the 'front' just turned around backwards).

If you're seeing better noise performance out of 2 MP rather than 10, that's great, but if a newer sensor was downsampled to 2 MP then it would kick the pajamas off of the old sensor.
01-06-2014, 08:45 AM - 1 Like   #63
Veteran Member
jsherman999's Avatar

Join Date: Dec 2007
Photos: Gallery
Posts: 8,228
QuoteOriginally posted by Reltih Quote

If more signals mean more noise, how do we explain why the Df has very good ISO performance with very big pixels?
The DF/D4 has very high QE, the performance is not necessarily directly attributable to it's pixel size. The guess is that the # of pixels there was chosen mainly to assure a high FPS rate, which is important to sports shooters and the like. (higher MP = slower frame rate) and to differentiate the D4 from the D800 as the 'sports' camera.

If you look at the pixel-level performance vs. the 36MP D800, the Df looks better (btw below simulates what you'd see 'pixel peeping' ) :



But when you look at the image, sampled to the same size, the noise per pixel averages out, and the real performance of the sensor as a whole shows itself:


,

That ^^ shows exactly what this article describes.


.

Last edited by jsherman999; 01-06-2014 at 08:50 AM.
01-06-2014, 09:32 AM - 1 Like   #64
Pentaxian
audiobomber's Avatar

Join Date: May 2008
Location: Sudbury, Ontario
Photos: Albums
Posts: 6,683
I am a stickler for aperture equivalence, because it is the only way to understand differences due to sensor size. If you say a 4/3 100mm f/2.8 lens has a FF equivalent focal length of 200mm, you will get no argument from me. If you say it is equivalent to a 200mm f/2.8 lens, I will definitely call you on it; 100mm f/2.8 on a 4/3 camera is 200mm f5.6 equivalent on FF.

Everyone understands the DOF ramifications of a crop sensor, but few understand aperture equivalence. The reason people get confused is that they focus on the exposure triangle; aperture, shutter speed and ISO. Shutter speeds should be the same, (1/(actual FL multiplied by crop factor). Aperture is multiplied by crop factor for equivalence. ISO cannot be made equivalent, and is therefore irrelevant. To understand equivalence, you need to focus on signal-to-noise ratio, and forget ISO. The following parameters will provide functionally identical images, assuming similar sensor technologies, equally capable lenses and the same subject distance:

4/3 camera, 100mm lens, f2.8, 1/250s, ISO 200
FF camera, 200mm lens, f5.6, 1/250s, ISO 800

Multiplying the aperture by crop factor allows you to accurately compare performance across formats.


Last edited by audiobomber; 01-06-2014 at 10:21 AM.
01-06-2014, 09:41 AM   #65
Veteran Member




Join Date: Feb 2011
Posts: 4,893
^^correct.
01-06-2014, 11:04 AM   #66
Senior Member




Join Date: Nov 2011
Location: Los Angeles
Photos: Albums
Posts: 167
QuoteOriginally posted by Class A Quote
The Sigma 8-16/4.5-5.6 for APS-C (available for K-mount) includes all the focal lengths you mention. Its FF equivalent is a 12-24/7-8. Pretty slow, but wider than even your 14mm on FF.
I stand corrected and will slink back to a film SLR forum where I belong...
01-06-2014, 07:10 PM - 1 Like   #67
Pentaxian
Class A's Avatar

Join Date: Aug 2008
Location: Wellington, New Zealand
Posts: 9,193
QuoteOriginally posted by Reltih Quote
By more total light do you mean on the same picture frame or not?
I usually assume the same image, i.e., same distance to subject, same FOV, etc.

QuoteOriginally posted by Reltih Quote
Giving the same lens, FF receives more total light, but the additional light is spread on the additional area around the cropped area. Where is the advantage?
With respect to noise, excluding extreme scenarios, it does not matter what size the sensor had once you print all images to the same size (let's assume both small and large sensor have the same pixel count and that equivalent parameters were chosen during capture).

The enlargement factors will be different for the small and large sensor respectively -- and that definitely has impact on factors other than noise -- but if one of the sensors received more total amount of light, this will reflect positively in the print. Whether that amount of light was more spread out or more concentrated in the intermediate sensor recording stage does not matter.

Please note that a lot of noise is caused by so-called "shot noise", i.e., the stochastic nature of light. When light levels become lower it isn't the case that all intensities at all sensels of the sensor become lower, rather, less photons hit overall which means that at some places the intensity may even stay the same but at other places it becomes a lot lower. The less photons there are, the more inhomogeneous the intensity pattern becomes, in other words, the noise level rises (compared to the signal).

QuoteOriginally posted by Reltih Quote
If more signals mean more noise, how do we explain why the Df has very good ISO performance with very big pixels?
I'm not sure what you mean by "more signals mean more noise".
It is correct that noise levels rise with light levels, but what counts is the signal to noise ratio (SNR). Because the noise only increases with the square root of the signal, more signal means better SNR (i.e., less relative / visible noise).

The Df has very good ISO performance because the Renesas produced NC81366W sensor has low read-out noise, etc. Smaller pixels increase the challenge of maintaining the same fill factor (the ratio between light sensitive and "blind" area on the chip) but in particular micro lenses help to redirect light from the "blind" (circuit related) areas to the light sensitive sensels.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
24x36mm, aps-c, f1.8, ff, full-frame, image, lens, pentax, question
Thread Tools Search this Thread
Search this Thread:

Advanced Search


Similar Threads
Thread Thread Starter Forum Replies Last Post
I Just Want to Whine..... Rupert General Talk 27 04-29-2013 03:15 PM
Do You Ever Just Want To.... magkelly General Talk 22 02-22-2012 11:23 AM
Just want to say Hello... SpicyLizard Welcomes and Introductions 2 01-29-2012 09:55 AM
People Just want to share dycz Post Your Photos! 5 10-23-2009 06:38 AM



All times are GMT -7. The time now is 11:35 AM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top