DPReview's Equivalence article

jurgen

Mu-43 Regular
Joined
Jan 27, 2014
Messages
120
I saw that. It's a pretty good read.
Agreed. And, in my opinion, for the most part completely irrelevant. It's good to be able to do the math on the fly, and you can never have a too comprehensive understanding of your tools or craft, but I've never once seen an award-winning photograph taken from the DxO laboratory. It's possible to get too hung up on absolute image quality when things like location, ergonomics, non-IQ performance, and vision are (ultimately) much more significant, especially with today's sensor technology.
 

nstelemark

Mu-43 Hall of Famer
Joined
May 28, 2013
Messages
3,345
Location
Halifax, Nova Scotia Canada
Real Name
Larry
The only issue I have is the total light concept. While it is an easy way to describe lower noise in a larger sensor it is not really correct. Light is not equivalent to rain.

To give them credit they do talk about light per unit area which is refreshing.
 

fortwodriver

Mu-43 Top Veteran
Joined
Nov 15, 2013
Messages
993
Location
Toronto, Ontario, Canada
Real Name
Frank
How incredibly frustrating it is to see T-Stops and F-Stops confused over, and over, and over again.
The cine guys must be laughing a blue-streak over the comments being posted to the article.
So while the article ultimately says "get over it", it still presents the data in a misleading way.
 

Reflector

Mu-43 Hall of Famer
Joined
Aug 31, 2013
Messages
2,165
They actually get it right unlike Tony Northrop and since DPR is one of the more heavily viewed sites I'm pretty happy about that. Moreso on the real world demonstration parts where you can visibly see the noise in the crops.
 

Djarum

Super Moderator
Joined
Dec 15, 2009
Messages
3,353
Location
Huntsville, AL, USA
Real Name
Jason
The only issue I have is the total light concept. While it is an easy way to describe lower noise in a larger sensor it is not really correct. Light is not equivalent to rain.

To give them credit they do talk about light per unit area which is refreshing.
For a larger format at the same f-stop, there is more total light hitting the sensor. This generally translates to a higher signal to noise ratio. However, the science doesn't always prove this out, especially considering the number of variables at play, such as pixel density, sensor effeciency, and sensor generation.
 

Fri13

Mu-43 Veteran
Joined
Jan 30, 2014
Messages
359
Who can claim their face seriously in neutral expression that they believe that smaller sensor size is reason to worse noise or deeper depth of field, instead just to alter the angle of view?

I want anyone who supports that DPR article or Tony Northrup to show how a KODAK Portra 160 suddenly becomes 320 or even 800 by grain quality when I cut the film frame to 1/2 or 1/4 of its full frame size on SLR and I take photo with it with SLR? ALL what I get changed is the angle of view. I don't need to adjust any exposure value no matter do I take photos in same condition but only with a different size film frames.

Same thing is with digital cameras. Sensor size doesn't affect to anything else than angle of view. Noise quality or depth of field changes are caused by totally other factors what photographer changes bases angle of view or what sensor is used. We can not compare MFT sensor from Panasonic to 35mm sensor from Canon as they are not using same digital technology!

We even get totally different results for noise quality when we compare exactly same size sensors with same pixel amount when the technology is different between them or when image processors are different when using exactly same sensor!

Sensor physical size isn't the reason for does it capture more light or not. Sensor size can be one factor for does company want to implement their latest most advances technology first to most expensive pro grade cameras or to entry-level cameras with small profit margins.

There are questions about patents and licensing issues and even manufacturing costs or challenges is there worth to bring what technology to what kind cameras.

Sensors should be compared by their technical features and differences and not by their size. But most photographers arguing that bigger sensor is better because it is better, has no knowledge to imagine sensor manufacturing or A/D conversions.

Digital technology isn't black magic, but isn't it either a open book but covered with patents, NDA's, corporation secrets, software algorithms in firmwares etc. Much more than "It is the sensor size what matters!".
 

TransientEye

Mu-43 Regular
Joined
Sep 18, 2013
Messages
159
Location
Barcelona
Real Name
Mark
While it is an easy way to describe lower noise in a larger sensor it is not really correct. Light is not equivalent to rain.
Oddly enough, rain is a pretty good analog for light, at least in this context. The sensor really is like a set of buckets (the pixels) that collect drops of light (photons).

In fact in low light, that quantisation is responsible for a lot of the noise you see in an image (possibly the majority of the noise - modern sensors are astonishingly efficient). There is a good description of shot noise here.
 

OzRay

Mu-43 Hall of Famer
Joined
Jan 29, 2010
Messages
4,991
Location
South Gippsland, Australia
Real Name
Ray, not Oz
I honestly don't understand why this is so important. Learn to use what you have and don't be concerned with what's on the other side of the fence. If it bothers you so much, climb over the fence and eat the grass on the other side.
 

Amin Sabet

Administrator
Joined
Apr 10, 2009
Messages
10,905
Location
Boston, MA (USA)
I'm sorry to have to move this to the back room, but I can't be the only one who is nauseated by seeing this over and again. After being beaten over the head with it at DPR for years before starting this site, I've had my fill.
 

Fri13

Mu-43 Veteran
Joined
Jan 30, 2014
Messages
359
Oddly enough, rain is a pretty good analog for light, at least in this context. The sensor really is like a set of buckets (the pixels) that collect drops of light (photons).

In fact in low light, that quantisation is responsible for a lot of the noise you see in an image (possibly the majority of the noise - modern sensors are astonishingly efficient). There is a good description of shot noise here.
The photons are like rain, more light = more droplets.
The photons are hitting given area randomly, what makes a challenge to gather them. Just like trying to cover a whole ground with water when it rains only a 1 seconds. Compare that to whole day length raining and you have huge difference how many droplets does hit the given ground area.

If we draw a 5x5 meter area to ground and set a garden springer to cover that area with same pressure, we get different amount of droplets in different time periods. Like compare 1 second to 10 minutes.

The photon collectors are same as the buckets surface area.
We can have a funnels what collects more droplets from larger area and directs them to smaller bucket below them to fill small bucket faster (microlenses) or we can have a buckets size of the funnels but shallower by their depth (technological differences A) so we can get them filled faster. Or we can have same size buckets as those two by surface area but deeper what requires more droplets to get filled (technological differences B).
We can have a buckets what are different shapes (like X-trans sensor from Fuji, technological differences C) or we can have a buckets with different filters in them to allow different size droplets pass trough to another below them (like Sigma layered sensor, technological differences D).
Or we can fill buckets with a other material like pellets, oil, paper, iron etc and do calculations differently than metering water depth like how well does electricity flow trough them (technological differences E) or we can even meter the weight of the buckets (technological differences F) or we can even set a microphones bottom of the buckets to listen droplets causing noise and meter amount of water by that (technological differences G) or we can set sensors sensing vibrations....

The bucket surface area or the amount of the buckets or the total area of laid with buckets doesn't matter much when we can have totally different ways to meter the amount of water droplets falling from garden sprinkler randomly.
Let's say that garden sprinkler is a cloud what is causing rain to whole city. It doesn't matter is the covered area 5x5 meters or 50x50 meters as each 1x1 meters area inside those areas gets same amount of droplets randomly.

If we now claim the metered area is reason for more accurate droplets counting for given square meter, then it would mean that we claim that the technology used to actually meter the droplets doesn't matter and we only need to improve the metered area size with any metering technology (any sized, shapes or sensor in them) in buckets.

Those who like to argue the sensor size matters, are exactly arguing that the metering technology doesn't matter at all.
There can be so huge differences between metering methods even when the metering surface or individual bucket surface is exactly same, causing totally different results amount of droplets in same time period. We could use very small tea cups or whiskey glasses instead buckets or even cover whole area with large paper sheet and we can get totally different results for same metering area in same time from same amount of water randomly spread over the same area.

Each camera manufacture use different metering methods, different metering surfaces and even between sensor manufacturers they can use same or different technology to same or different size of sensors.

We can not just compare metering surface area and say larger is better if we don't compare the technologies used to sense droplets and how they are counted and processes and how data is fianlly presented and stored.
 

nstelemark

Mu-43 Hall of Famer
Joined
May 28, 2013
Messages
3,345
Location
Halifax, Nova Scotia Canada
Real Name
Larry
Photons - rain and the dual nature of light
If the sensor was a bathtub and we were draining it to make an image then the total light hitting the sensor would make a difference, but fortunately for the electronics in our cameras we aren't :smile: .

Seriously though light is a wave, until it interacts with something else.* At that point at the individual sensor sites light behaves like a particle - ie a photon. Assuming equivalent sensor site performance the key is light per unit area not total light. It is generally accepted that larger sensors have large photosites and these sites are more efficient than smaller ones assuming you are inside their performance thresholds. Ie is there enough light that one does not have to increase the signal amplification which also amplifies the noise floor. But, assuming you can expose the image you want without amplifying the noise the smaller sensor will perform in roughly an equivalent manner to the larger sensor. (DOF/diffraction issues aside which are related to the overall geometry of the sensor/lens opening).

So there is a corollary that larger sensors exhibit lower noise characteristics however the cause is not the "volume" of light in the cavity, but the electrical performance of the larger photosites when you have to amplify the signal. If they said the larger photosites was the cause of the improved noise performance I'd be happy. But this rain hitting the sensor analogy implies that more light in the cavity will simply be better, and that is a fallacy - ie corollary not causality.

* the f-stop vs t-stop issue is very relevant here. Really they needed to compare lenses with equivalent t-stops and equivalent aperture to say anything useful about the light hitting the sensor.
 

Listener

Mu-43 Regular
Joined
Feb 20, 2014
Messages
155
Location
SF Bay area in California, USA
Agreed. And, in my opinion, for the most part completely irrelevant. It's good to be able to do the math on the fly, and you can never have a too comprehensive understanding of your tools or craft, but I've never once seen an award-winning photograph taken from the DxO laboratory. It's possible to get too hung up on absolute image quality when things like location, ergonomics, non-IQ performance, and vision are (ultimately) much more significant, especially with today's sensor technology.
Amen to everything you said.

On the other hand, it provides endless fodder to armchair photographers on DPR and elsewhere.
 

DoofClenas

Who needs a Mirror!
Joined
Nov 9, 2012
Messages
1,280
Location
Traverse City, MI
Real Name
Clint
I'm sorry to have to move this to the back room, but I can't be the only one who is nauseated by seeing this over and again. After being beaten over the head with it at DPR for years before starting this site, I've had my fill.
Perhaps you should create a dungeon with a fire breathing dragon...I don't think the back room is deep enough to bury this :).
 

BAXTING

Mu-43 Top Veteran
Joined
Aug 5, 2012
Messages
805
Location
Los Angeles SFV, CA
Real Name
Bradley
A photon walks into a hotel and books a room for the night. The clerk at the desk asks if they need help with their luggage. The photon replies, "No, I'm traveling light".
 

Art

Mu-43 All-Pro
Joined
Mar 13, 2011
Messages
1,385
Location
San Francisco, CA
Great article, finally clearly explaining why the equivalent aperture is the f-number multiplied by crop factor. It's the physical aperture which matters not the f-number. I hope folks will stop comparing 45mm f1.8 on m43 to 85mm f1.8 on FF when they should be comparing it to 85mm f3.6 especially when comparing size.


Sent from my iPhone using Mu-43
 

Fri13

Mu-43 Veteran
Joined
Jan 30, 2014
Messages
359
Great article, finally clearly explaining why the equivalent aperture is the f-number multiplied by crop factor. It's the physical aperture which matters not the f-number. I hope folks will stop comparing 45mm f1.8 on m43 to 85mm f1.8 on FF when they should be comparing it to 85mm f3.6 especially when comparing size.


Sent from my iPhone using Mu-43
45mm f/1.8 is a 45mm f/1.8 lens.... Giving only a equivalent angle of view of 85mm.

The aperture physical size doesn't matter. You don't get more light because that, as the physical ratio is still exactly the same between both lenses.

Because you have cropped view (angle of view is equivalent to 2x longer focal length. You can not multiply the angle of view by 2x, like 50mm gives 46° angle of view for 35mm format but same 50mm on MFT gives you 20° degree angle of view, what requires a 100mm focal length for 35mm) you use smaller portion of the imagine circle while getting same amount of light.

We can mount with adapters objectives designed for larger format to smaller one, getting as much light to each pixel no matter the physical aperture diameter.

Because you only get a cropped view (angle of view) that means you don't have extra space around subject to frame. Meaning you need to either a) step further b) frame tightly and correctly c) use a shorter focal length.

A changes perspective and lens optical magnification causing depth of field getting deeper
B can be difficult if you always compose tightly and you don't leave possibilities to crop in post
C forces again to use a wider angle lens what has a smaller optical magnification what leads to narrower depth of field

Light traveling trough is same, only angle of view gets changed.

That's why it is called a "crop factor" and not "multiplier".

Place your 35mm camera to tripod and frame subject with 50mm lens. Switch body to MFT camera and mount same lens with adapter on it and use same aperture. You get exactly same exposure values for shutter speed and ISO and same depth of field but your framing is very tight because camera sensor use only 1/4 of the same angle of view (2x).

Now if you would have same sensor and digital technology in use on both cameras, you would even get exactly same noise quality. But you don't so you have all digital parts different what cause different noise quality, not the sensor size.

Sensor size doesn't affect to anything else than change to angle of field, because it is only relative what portion of the imagine circle is being captured.

If you still have hard time to understand it, I say it is your time to go and start using a film camera and learn the basics of photography, then learn differences in digital technology about electrons, software etc.

It was so easy to teach people the film size only affecting to angle of field at film era, because there was no digital technology causing assumptions. The digital part is a total blackbox, mystical thing for most photographers and they do not understand that there are different technologies to get similar things.

Just like on film era we had dozens of film rated same ISO speed, each giving a different grain look, capturing colors differently and allowing different pushing possibilities or different exposure range.

Now we have digital cameras where all characteristics from different films are from sensor, ADC processor and software algorithms.

Three factors instead one, different technologies for each three and countless possibilities to implement same technology differently...

And people falls to old myths that bigger sensor size means quality....
 
Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2009-2019 Amin Forums, LLC
Top Bottom