Equivalence

[joe mama]

That’s what I’ve been arguing all along. Joe, drop the bogus total
light arguments from you essay and I might agree with it.

The "total light arguments" are far from bogus. However, you do raise an important argument that deserves special mention in the essay, so, as is the point of this whole thread, I have complied. Here are two new paragraphs in the "Total Light" chapter of the essay:

"Of course, it's also important to note that the light is not the only source of noise -- the sensor and supporting camera hardware also contribute to the noise of the image. By comparing sensors of the same design and generation, we are able to eliminate this variable. However, even for the times we cannot, for modern DSLRs, there is remarkable little difference between the noise performance (for sensors of the same size) even between brands, certainly no more than one stop, and usually much closer. Nonetheless, it is a factor to consider.

If the lowest possible noise with the highest possible DOF is important, then, from time to time, there will situations where the larger sensor is at a slight disadvantage in terms of noise (but only noise). For example, at the time of writing this article, the Canon 1DIII has been released, but the 5DII has not. Based on sensor size alone, the 5D should have 2/3 stops better noise performance than the 1DIII (for images at the same output size). However, in practice, it is only 1/3 stop better. Of course, when the 5DII comes out, we would expect it to once again have 2/3 stops better noise performance."

I doubt that Joe would do that, since the total light argument is
central to the concept of equivalence.

Exactly so. But joe is certainly interested in getting his facts straight and addressing any arguments that need addressing. : )

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[joe mama]

Yes, I might be better off keeping the 17-55 & 20D for when I need
the versatility, and as a "TC" for my 70-200, and dedicating FF for
fast primes.

As you may, or may not, know, Lee Jay uses both the 20D and 5D. In fact, he also uses an S3 IS and G7. So, he takes advantages of four different sensor sizes!

I find your shallow DoF photography inspiring by the way.

As my photography is more "one dimensional" in that regard, I find only the need for the largest sensor, and fastest glass, I can afford. : )

By the way, thanks for the props!

I did not go FF to get equivalent images, I went FF to get "better"
(for my tastes) images. But, it's important to note that I can still
take the exact same types of pics I took with my 20D with a 5D, with
no loss in IQ, and usually much higher IQ. I do not need to suffer
more vignetting, softer edges, or more distortion -- three factors
that rarely mattered to me anyway! : )

I do find it encouraging to know that I wouldn't be giving anything
up by moving to FF, and I appreciate the effort you and Lee have
taken to explain this since I wasn't aware of it before reading some
of the equivalence posts. So while some are irritated by the "gospel"
you guys have been spreading in numerous threads, my photography
could well end up benefiting from it.

And therein lies the entire point of the equivalence essay, in case anyone wonders why I bother.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[joe mama]

First, I'd like to thank all those who have read and critiqued the essay. Your corrections and objections were exactly the help I needed, and I sincerely apologize for any combativeness on my part. You have done me a fine service and I greatly appreciate it!

That said, please reread the essay:

http://www.josephjamesphotography.com/equivalence/

I have incorporated a huge number of changes that are all directly a result of the feedback you all have given me. I hope that I have addressed your objections and concerns.

Of course, this is a work in progress, and I'm sure I'll be tweaking it even a year from now. I already have plans to have links within the essay to images and diagrams to help explain (but don't hold your breath -- that will take a while!).

So, again, please accept my gratitude and my apologies to all the help that you have all given me! I know it's a long read, and I know it's easy to skip over points I've made in it (hell, I've even misread a few posts in this thread!), but if you don't mind another once-over, I'd appreciate it!

Thanks!

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[dragon899]

Hey Joe..

...total sideline question here...noticing you had the 35 1.4 when you had your 5D...but you sold it...I know you mentioned that you were getting the 24 1.4 soon...what was it about the 35mm on the 5D you didn't like? I guess I'm asking because I was considering the 35 quite strongly with the 5D and would love to know your thoughts from experience with them...

 

[bad_doggie]

"It is important to understand that exposure is not the important
measure for an image -- total light is the important measure."

Funny how precient I was when writing it, huh? : )

prescient, perhaps, but ...

this is a philosphical statement, not a statement of fact. and one that i'm sure most would disagree with, if not plunked down in the middle of intimidating language.

i read it kind of like a master chef teaching other chefs: "its not the taste of the food that matters, but how it is prepared and presented."

in photography, as most photographers understand it, exposure is an important measure for an image. who was it who said: newbies obsess about lenses; more advanced photographer obsess about tripods; seasoned photographers obsess about lighting. something like that.

 

[vor]

Think of high-frequency noise as fine sandpaper, and of low-frequency
noise as coarse sandpaper. As one increases the spatial frequency of
the noise, the noise becomes finer--relatively smaller specks in the
final image. And yet, the signal itself remains both coarse and
fine--for example, the entire head of the person we are photographing
along with its details (eyes, nose, freckles). If the size of the
specks of noise drops below the size of the pixels, the noise
effectively goes poof.

I don't quite understand this, but something is wrong here! How could
the size of noise suddenly "drop below the size of the pixels"? No
matter how many pixels you have, noise will always have the same
size, or frequence, relative to the pixel-size!

I'm talking output pixels, as in the pixels of the final image. Since we have the same output image size, the greater resolution of the larger sensor must be downsized. As that occurs, the size of the noise drops below the size of the output pixels. Sorry that I didn't make that clear.

Victor

 

[Chez Wimpy]

indeed... and by how much?

Simple: 1 1/3 stops. Take another look at the pics I linked in this
thread.

It boils down to this. If this is what we can actually see in pictures, you will be able describe how mathematically 54% more pixels of equal sensitivity can deliver not 2/3rds a stop, but a full 1 and 1/3s a stop. I just want to see the math here.

(another point while I am thinking about it: the crop factor = teleconverter is not correct in the case of the 5D vs 30D since the central 1.6x area of the 5D sensor of the 5D records less photons than the 30D as per the pixel-noise data on this site, and John Sheehy's own analysis of camera raw data

http://forums.dpreview.com/forums/read.asp?forum=1018&message=24048229 Light from that central 1.6x area that is gathered in the case of the 20D/30D is wasted in the case of the 5D, so there is actually a gain of reach and greater DOF without a noise penalty at equal ISO settings with the 20D/30D)

--
-CW

 

[Barugon]

If you increase the size of the sensor and leave the photo-sites the
same size then you will have exactly the same amplitude of noise.
However, since you now have more photo-sites you now have more noise.
That's right, you have a larger volume of noise that is exactly the
same amplitude. If you print the image the same size as the smaller
sensor all you have done is increased the spatial frequency of the
noise.

And you have NOT increased the spatial frequency of the signal, but
instead you have increased the frequency range of the signal; as
such, both low-frequency and high-frequency signal will be present,
whereas only higher-frequency noise will be present (and
lower-frequency noise will be notably absent). If you continue
increasing the spatial frequency of the noise until you reach the
spatial resolution of the output image, you then get rid of the noise
altogether and are left with nothing but signal.

You do understand what "spatial" means don't you?

I see what you're saying now. It seemed to me earlier that you were contradicting yourself. Anyway, yes, if the noise frequency is high enough it won't be visible; just like the dot pattern from an inkjet printer isn't visible because it's higher resolution than your eye. My original argument was just that the noise is still there, it's just as strong as before and there's more of it.

--
Whoever said 'a picture is worth a thousand words' was a cheapskate.

http://www.pbase.com/dot_borg

 

[joe mama]

Simple: 1 1/3 stops. Take another look at the pics I linked in this
thread.

It boils down to this. If this is what we can actually see in
pictures, you will be able describe how mathematically 54% more
pixels of equal sensitivity can deliver not 2/3rds a stop, but a full
1 and 1/3s a stop. I just want to see the math here.

You should reread the essay. Here's the link again:

http://www.josephjamesphotography.com/equivalence/

Equivalent images are compared at the same output size . I devoted a whole chapter to it and even mention this common misunderstanding in the introduction. We do not compare images pixel for pixel unless the images have the same number of pixels.

See these two links as well:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=23372035

http://forums.dpreview.com/forums/read.asp?forum=1029&message=23350027

(another point while I am thinking about it: the crop factor =
teleconverter is not correct in the case of the 5D vs 30D since the
central 1.6x area of the 5D sensor of the 5D records less photons
than the 30D as per the pixel-noise data on this site, and John
Sheehy's own analysis of camera raw data

http://forums.dpreview.com/forums/read.asp?forum=1018&message=24048229 Light from that central 1.6x area that is gathered in the case of the 20D/30D is wasted in the case of the 5D, so there is actually a gain of reach and greater DOF without a noise penalty at equal ISO settings with the 20D/30D)

See here:

http://www.ddisoftware.com/20d-5d/

Bottom 1/3 of the page, where it says, "More detailed noise profile (updated 10/08/2005)"

Basically shows that the 5D averages around 2/3 stops better noise performance per-pixel than the 20D, which is what it "should" be. That translates to 1 1/3 stops better noise performance for images with the same output size.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[Lee Jay]

"It is important to understand that exposure is not the important
measure for an image -- total light is the important measure."

Funny how precient I was when writing it, huh? : )

prescient, perhaps, but ...

this is a philosphical statement, not a statement of fact. and one
that i'm sure most would disagree with, if not plunked down in the
middle of intimidating language.

i read it kind of like a master chef teaching other chefs: "its not
the taste of the food that matters, but how it is prepared and
presented."

Good analogy, too bad you used it backwards. The image is the taste, the exposure numbers are the model numbers on the pots and pans.

in photography, as most photographers understand it, exposure is an
important measure for an image.

Then why isn't the exposure prominently displayed in the center of each image sold at a gallery?

who was it who said: newbies obsess
about lenses; more advanced photographer obsess about tripods;
seasoned photographers obsess about lighting. something like that.

Not very close. Nevertheless, I'll change it further. Newbies care about exif data, photographers care about the image.

--
Lee Jay
(see profile for equipment)

 

[joe mama]

"It is important to understand that exposure is not the important
measure for an image -- total light is the important measure."

Funny how precient I was when writing it, huh? : )

prescient, perhaps, but ...

No "perhaps" about it. : )

this is a philosphical statement, not a statement of fact.

No, it is a statement of fact for the definition of equivalence I gave in the essay.

and one that i'm sure most would disagree with, if not plunked down in the
middle of intimidating language.

They may disagree all they want. Disagreement does not change what is.

i read it kind of like a master chef teaching other chefs: "its not
the taste of the food that matters, but how it is prepared and
presented."

More like a master chef teaching the other chefs: "it is the temperture in the oven, not the heat of the flame, that matters."

in photography, as most photographers understand it, exposure is an
important measure for an image.

The definition of exposure is outdated. The reason is that in the past, with film, photographers did not try to match the noise characteristics between systems.

who was it who said: newbies obsess about lenses; more advanced
photographer obsess about tripods; seasoned photographers obsess about
lighting. something like that.

Yeah, something like that. However, this is a technical dissertation, not an artistic one. Who discusses art on DPR, especially in the Canon SLR Lens Forum? Quite honestly, I'd prefer FF over 1.6x even if it didn't have a noise advantage -- shallow DOF is that important to me.

But I'm not discussing the artisitic merits between the systems. I'm talking about the technical capabilities between the systems. And, as the essay states, in terms of IQ (but not necessarily operation), anything 1.6x can do, FF can do at least as well, and FF can do things that 1.6x cannot even do.

So, for people deciding between systems, they give up nothing in terms of IQ if they go FF. However, for those that shoot long, or need to maximize DOF in low light, they may find that the advantages of FF are wholly insignificant, and the operation of FF to be expensive and cumbersome.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[joe mama]

Hey Joe..

...where you goin' with that gun in your hand?
Hey Joe, I said where you goin' with that gun in your hand?

Alright. I'm goin down to shoot my old lady, you know I caught her messin' 'round with another man. : )

...total sideline question here...noticing you had the 35 1.4 when
you had your 5D...but you sold it...I know you mentioned that you
were getting the 24 1.4 soon...what was it about the 35mm on the 5D
you didn't like? I guess I'm asking because I was considering the 35
quite strongly with the 5D and would love to know your thoughts from
experience with them...

I went FF for two reasons: even more shallow DOF and less noise. When I got the 5D, my 35 / 1.4L performed as would a 22 / 0.9L on 1.6x. I had never used something that wide and fast, and, quite honestly, didn't know what to do with it. I think I took all of five pics with it.

So I sold it within a week and got the 24-105 / 4L IS. I kept that for a couple of weeks before realizing that wasn't my pot o' tea, either. So I sold that for a 16-35 / 2.8L. Instant bliss.

Now I've come full circle and will be selling the 16-35 / 2.8L this weekend to fund a 24 / 1.4L.

So, what's wrong with the 35 / 1.4L? Absolutely nothing. I'd like to own one again, now that I know what to do with it. But, as I own and love the 50 / 1.2L, the 24 / 1.4L comes first.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[vor]

If you multiply the signal by 1.6 and the noise by 1.6 then you have
exactly the same signal to noise ratio.

This would be nice, except that it doesn't work that way. The coherence of signal results in a cumulative effect as you increase signal, whereas the randomness of noise results in a canceling effect, reducing it. As the sample size approaches infinity, the signal approaches its ideal value while the cumulative effect of the random noise approaches zero.

http://www.cambridgeincolour.com/tutorials/noise-reduction.htm

This article is about taking two or more images and averaging them
together. This is using a temporal variable to reduce noise and has
nothing to do with sensor size. It would work equally well for any
sensor.

But the principle is the same: you are adding both signal and noise, and yet the cumulative effect is noise reduction. It doesn't matter whether the extra data is across time or across space, the result is the same. If you want to see more relevant articles, try this one:

http://www.cambridgeincolour.com/tutorials/digital-camera-sensor-size.htm

Look at the end of the article, where it says this: "No matter what the pixel size, larger sensors unavoidably have more light-gathering area."

Victor

 

[bad_doggie]

I honestly just don't see the resistance to the term. By the way,
I've a favor to ask: could you take a look and respond to this post:

why is this so hard for you to understand? you've been told so many times.

people don't agree because it has no relation to their experience. f-stop determines exposure. offsetting equivalence OF LENSES by tweaky adjustments of ISO and noise and final image size is just (the noise argument a dubious one) is just quirky.

but more than anything else: most folks don't look at lenses and ask "what camera can i use to obtain the shallowest DOF possible?" they want adequate DOF.

yes, yes, you can make the slippery slope argument and say, well what about a P+S camera, you can't even blur the background, yadda yadda. but that's not an issue with FF vs 1.6x crop. virtually nobody needs less DOF than can be got with a 1.6x crop and an f/2.8 zoom.

they just are not equivalent lenses. maybe you can produce equivalent images USING those lenses, together with appropriate settings of the cameras (inequivalent settings). people work with equivalence in an operational sense: if i put a 30mm/1.4 lens on 20 i USE it the same way i use a 50mm/1.4 lens on a 5d. sure, i'll have to stop down the 5d a bit more, oh well. and up the ISO on the 5d, oh well. but in all the ways that i'm using features of the LENS - to choose framing and perspective and magnification and f-stop - they are equivalent.

anyway, i continue to say, this whole argument is just not interesting, unless you're one of those folks pining for the one-eye-in-focus-blurry-ears portraits on a rebel. and you can still produce such monstrostities with a rebel and an 85L.

 

[bad_doggie]

Good analogy, too bad you used it backwards. The image is the taste,
the exposure numbers are the model numbers on the pots and pans.

fine. but i said nothing of ev's.

so if the ev is the model number of the pots, then i guess all these other things - aperture, ISO, sensor size - they're what? the essence of photography? gimme a break.

Then why isn't the exposure prominently displayed in the center of
each image sold at a gallery?

ah. and total light is?

exposure is "featured" because it is self-evident in the photo, regardless of what the camera's wrote into the exif. "total light" is not. or are you claiming that you can look at a photo in a gallery and tell me what ISO was used?

Not very close. Nevertheless, I'll change it further. Newbies care
about exif data, photographers care about the image.

wow, big statement coming from you guys, whose whole thesis is comes down to an equivalence relation on exif data.

 

[joe mama]

I honestly just don't see the resistance to the term. By the way,
I've a favor to ask: could you take a look and respond to this post:

why is this so hard for you to understand? you've been told so many
times.

Unfortunately, I've been told by people who don't understand. : )

people don't agree because it has no relation to their experience.

You are totally and completely wrong there. What person has experience that the following:

30D @ 30mm, f / 2.8, 1/200, ISO 100
5D @ 50mm, f / 4.5, 1/200, ISO 400

produce different results at the same output size? What person has the experience that:

30D @ 30mm, f / 2.8, 1/200, ISO 100
5D @ 50mm, f / 2.8, 1/200, ISO 100

produce the same results are different output sizes? I'd like names, please. : )

f-stop determines exposure.

That's as misleading as saying focal length determines FOV.

offsetting equivalence OF LENSES by tweaky adjustments of ISO and noise and
final image size is just (the noise argument a dubious one) is just quirky.

Seriously? You think that using 30mm on 1.6x and 50mm on FF is quirky? You think printing images, or displaying them on the web, at the same dimensions is "just quirky"?

but more than anything else: most folks don't look at lenses and ask
"what camera can i use to obtain the shallowest DOF possible?" they
want adequate DOF.

Guess what? Most people own compact digicams, not DSLRs, too.

yes, yes, you can make the slippery slope argument and say, well what
about a P+S camera, you can't even blur the background, yadda yadda.
but that's not an issue with FF vs 1.6x crop. virtually nobody needs
less DOF than can be got with a 1.6x crop and an f/2.8 zoom.

I'm one of those "nobodies". But, that's even besides the point. Are you saying also that virtually nobody needs the greater MP of FF? Are you saying that virtually nobody needs the better noise performance of FF?

Brace yourself: I TOTALLY AGREE.

I think FF is way overkill for most people. I do not recommend it to most people. I think most people are better off with a quality compact digicam.

they just are not equivalent lenses. maybe you can produce equivalent
images USING those lenses,

Guess what: that's the definition of equivalent lenses.

together with appropriate settings of the cameras (inequivalent settings).

You confuse "equivalent" with "equal". The sensors have unequal sizes, therefore the settings are unequal to get equivalent images.

people work with equivalence in an operational sense: if i put a 30mm/1.4
lens on 20 i USE it the same way i use a 50mm/1.4 lens on a 5d. sure, i'll have
to stop down the 5d a bit more, oh well. and up the ISO on the 5d, oh well.

Exactly right and they will be "rougly" equivalent images.

but in all the ways that i'm using features of the LENS - to choose framing
and perspective and magnification and f-stop - they are equivalent.

I don't know what you mean.

anyway, i continue to say, this whole argument is just not
interesting, unless you're one of those folks pining for the
one-eye-in-focus-blurry-ears portraits on a rebel. and you can still
produce such monstrostities with a rebel and an 85L.

So are you saying that no one but shallow DOF whores like myself have a need for FF? Go test that theory in the 1D/5D Forum.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[Lee Jay]

anyway, i continue to say, this whole argument is just not
interesting, unless you're one of those folks pining for the
one-eye-in-focus-blurry-ears portraits on a rebel. and you can still
produce such monstrostities with a rebel and an 85L.

You think it's all about shallow DOF?

Try to reproduce this shot on a 1.6-crop camera.

5D, 15mm fisheye, f2.8, ISO 1600, 1/125th to freeze the fish. Yes, that means you'll need a 9.5mm, f1.8 fisheye and ISO 640 on a 1.6-crop camera to get this shot. Got one of those 9.5/1.8 fisheyes handy? Fortunately I do, and it only cost me $370.



Or try this one with your 15mm, f0.9 lens on 1.6 crop (you have one, right?). This is from Daniel Bayer with a 5D, 24/1.4L at ISO 1600 for 30 seconds. Think the DOF is shallow on this shot? Before you ask, shooting a longer shutter speed with a 10-22 would cause visible star trails, thus dramatically altering the shot. (Posting a link since it's not my shot):

http://photo.net/photodb/photo?photo_id=3799192

--
Lee Jay
(see profile for equipment)

 

[Lee Jay]

wow, big statement coming from you guys, whose whole thesis is comes
down to an equivalence relation on exif data.

You really need to pass 5th grade this time (third time's a charm, don't you know). As usual, the entire point of this equivalence has gone right over your head.

--
Lee Jay
(see profile for equipment)

 

[joe mama]

exposure is "featured" because it is self-evident in the photo,
regardless of what the camera's wrote into the exif. "total light" is
not.

Total light is featured as well, in combination with the ISO setting. The total light determines the image, the ISO determines its presentation.

or are you claiming that you can look at a photo in a gallery
and tell me what ISO was used?

I can't even tell how far away the subject was from the camera, what FL was used, what f-ratio was used, or what shutter speed was used. But all those things are important to me in creating the image.

Not very close. Nevertheless, I'll change it further. Newbies care
about exif data, photographers care about the image.

wow, big statement coming from you guys, whose whole thesis is comes
down to an equivalence relation on exif data.

Sure. This is a technical forum for a technical discussion. Check out this thread:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=24395241

The guy wants an 18-50 / 1.8 for his 1.6x DSLR. No such thing exists. But he could get the same results with a 5D and 28-75 / 2.8, but he doesn't understand that.

Now check out further down in the thread:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=24406082

This guy, due to all the myths spouted by people who don't understand equivalence, is scared away from a system that will suit his needs.

That's why people need to understand. That's the reason for all the technical talk. If croppers didn't spout all the misinformation in the first place, I'd not have to be setting the record straight.

--
--joe

http://www.josephjamesphotography.com
http://www.pbase.com/joemama/

Please feel free to criticize, make suggestions, and edit my photos. If you wish to use any of my photos for any purpose other than editing in these forums, please ask.

 

[thw]

-------------------------------------------
See the colors of my world in:
thw.smugmug.com