No "Bad" Olympus digital lenses...

boggis the cat · Jul 7, 2010

Rikke Rask said:
Yes, the f-number is just a mathematical convenience and when it is used to calculate DOF, it is always used in combination with the focal length, simply because these are printed on the lens barrel and they can be used to calculate the size of the entrance pupil.

The DOF is dependent on the size of the entrance pupil, the distance to the subject, and the circle of confusion. Nothing else. These are the are the parameters and they can be measured physically with a (suitably fine) ruler or tape measure.

Yes, I agree.

Do you still claim that a given lens has a different f-number when mounted on different sensor-size bodies, though? That was the actual debated point.

You seemed to be saying that the crop-factor for APS changes the f-number of a lens mounted on by the same multiplier used to calculate the change in apparent DOF. This makes me think that you are confusing DOF effects with f-number, and this is my main complaint against the very loose usage of "equivalence theory": it can cause confusion about what the primary purpose of the f-number convenience is.

boggis the cat · Jul 7, 2010

Rikke Rask said:
boggis the cat said:

So the more you enlarge something, the shallower the apparent DOF. The opposite holds as well: the more you shrink something the deeper the perceived DOF.

Click to expand...

Yes. In fact, there really is no reason why we should call the DOF "perceived". It is after all the only kind we have.

My meaning is that if you change perspective (move closer or further away) then the DOF changes. DOF isn't a "real" thing, it is just a perception due to physical limitations (and the perception varies with visual acuity).

Josefin said:
Josefin said:

A 4/3 sensor has twice the angular shrink (from real scene to sensor) as a 135 sensor, so yields twice the apparent DOF.

Click to expand...

You can phrase it that, yes. We still agree.

Josefin said:

You can think of this as: the 135 sensor "enlarges" the same image compared to the 4/3 sensor, if this makes more sense.

Click to expand...

Another way of saying the same thing.

Josefin said:

That's the complexity level that moggies work at.

Click to expand...

:-D

Now for the final leap.

When we are looking at same size prints produced from different size crops (17.3x13 and 12.2x9.2, for instance), taken with the same f-number setting, the same focal length at the same subject distance, they will appear to have different DOF. So far we agree.

The one produced from the crop will look like it was produced with a 1.4x TC and one stop slower f-number. Say, a 70/2.8 instead of a 50/2. What really changed was of course the CoC because of the magnification, but it appears as if both the focal length and the f-number were changed by a stop each. None of them actually did, of course, we can check that with our ruler, but the effect is as if they did. The effect is so awfully close to what the TC does that it in an ideal world would be impossible to distinguish the two ways of achieving the result. They are, for lack of a better word, equivalent.

Yes, and that is an alternative explanation for "equivalence theory".

It appears that you are mixing the DOF effect (the basis of "equivalence theory") up with the actual f-number. A lens' parameters do not change depending on how much of the image circle you use -- only the parameters of the final image (the apparent DOF) does, and this is due to magnification.

All of which is far away from the original debate about whether 300 mm f/2.8 prime lenses were equivalent, and if there was any justification for Olympus apparent price premium in Europe. (There is an obvious justification: making a profit. The Americans are too poor and competition in Japan is too hard, I guess.)

Josefin · Jul 7, 2010

boggis the cat said:
Do you still claim that a given lens has a different f-number when mounted on different sensor-size bodies, though?

I never made such a claim. Not here and not elsewhere.

boggis the cat said:
That was the actual debated point.

No.

boggis the cat said:
You seemed to be saying that the crop-factor for APS changes the f-number of a lens mounted on by the same multiplier used to calculate the change in apparent DOF. This makes me think that you are confusing DOF effects with f-number, and this is my main complaint against the very loose usage of "equivalence theory": it can cause confusion about what the primary purpose of the f-number convenience is.

The confusion is squarely on those who say an x mm at f/ y lens on 4/3 is equivalent to a 2x mm f/ y on 135 even though it is obvious that they will take different pictures.

I'd say, let's stop the discussion here. We have both stated where we come from and haven't been able to reach an agreement. It has been amiable and fun.

Cheers :-D
Rikke

boggis the cat · Jul 7, 2010

Rikke Rask said:
boggis the cat said:

Do you still claim that a given lens has a different f-number when mounted on different sensor-size bodies, though?

Click to expand...

I never made such a claim. Not here and not elsewhere.

It appeared to me that you did here:

http://forums.dpreview.com/forums/read.asp?forum=1022&message=35725337

Rikke Rask said:
boggis the cat said:

If you put the Nikon 300 on an APS body you get 450 EFL f/2.8,

Click to expand...

No, you get an 450 mm efl f/4.2 equivalent. Remember, cropping is equivalent to using a TC. In this case a 1.5 TC.

It was that statement that made me scratch my head. The context was using lenses to get the telephoto reach, and the comparative cost of 4/3 options against those for (Nikon) lenses mounted on 135 and APS. "Equivalence theory" wasn't part of the discussion.

Presumably you considered tagging "equivalent" on the end to be sufficient to invoke the "equivalence theory".

Josefin said:
Josefin said:

You seemed to be saying that the crop-factor for APS changes the f-number of a lens mounted on by the same multiplier used to calculate the change in apparent DOF. This makes me think that you are confusing DOF effects with f-number, and this is my main complaint against the very loose usage of "equivalence theory": it can cause confusion about what the primary purpose of the f-number convenience is.

Click to expand...

The confusion is squarely on those who say an x mm at f/ y lens on 4/3 is equivalent to a 2x mm f/ y on 135 even though it is obvious that they will take different pictures.

I disagree, as the f-number is a set value and is primarily for exposure calculation. The secondary function is to let you know hoe poor you will be after you buy a lens -- the lower the f-number, the lighter your wallet.

If you want to use equivalent DOF to modify the f-number you should state that is what you are doing to prevent confusion. (Just tagging "equivalent" on the end wasn't enough, because the EFL bit means "equivalent focal length" and the discussion was about actual and equivalent focal length.)

Josefin said:
I'd say, let's stop the discussion here. We have both stated where we come from and haven't been able to reach an agreement. It has been amiable and fun.

So it appears that you were talking about the "equivalence theory", and it did cause confusion. Fancy that.

Josefin · Jul 8, 2010

boggis the cat said:
I disagree, as the f-number is a set value and is primarily for exposure calculation.

Nah, outdoors my camera handles the exposure calculation and I express my disagreement with the exposure compensation function (which I have mapped to the rear dial of my E-3 so I don't have to press the +/- button). The f-number dial is used to select the DOF ;-)

[using aperture priority with auto ISO; with the use of the exposure compensation, it is rather close to shooting manual, but somehow it feels more convenient, probably because the meter gets it right most of the time so I don't have to use both dials every time]

In the studio, however, I always shoot manual with the shutter speed set to 1/250 s, so here the f-number is used for getting the exposure as I want it (I use radio controlled monoblocs).

boggis the cat said:
The secondary function is to let you know hoe poor you will be after you buy a lens -- the lower the f-number, the lighter your wallet.

Oh yes, especially in our 4/3 world. My Summilux-D lightened my wallet quite a bit :-(
A ZD 300/2.8 would cost more here in Sweden than my car is probably worth.

But that is in the 4/3 world. I recently borrowed a 5D with an 85/1.8, actually a rather good lens for a nice price and traditionally 50/1.4 lenses have been among the optically best and at the same time about the cheapest stuff you could get.

--
Rikke

rovingtim · Jul 8, 2010

There's a lot that you have said that is true. Quickly, I wish to address something that I think could be added to ...

John King said:
BUT (and it is a BIG "but") the illumination per unit area of the sensor remains the same for a given f-stop regardless of the sensor size (assuming that the lens throws a sufficiently large image circle to 'properly' illuminate the sensor without showing vignetting etc ... ). That is to say, an f2 lens is always an f2 lens. This is a physical characteristic of a lens, regardless of whether it is even mounted on a camera body or not ...

The problem here is that a larger sensor uses a larger area to collect light. This means that because the exposure is identical, a 35mm sensor will collect 4 times the amount of light as a 4/3rds sensor.

Noticing that the exposure is identical simply accentuates this fact.

I would also like to note that FoV is an asthetic just as DoF is.

John King · Jul 8, 2010

Gidday Tim

rovingtim said:
There's a lot that you have said that is true. Quickly, I wish to address something that I think could be added to ...

John King said:

BUT (and it is a BIG "but") the illumination per unit area of the sensor remains the same for a given f-stop regardless of the sensor size (assuming that the lens throws a sufficiently large image circle to 'properly' illuminate the sensor without showing vignetting etc ... ). That is to say, an f2 lens is always an f2 lens. This is a physical characteristic of a lens, regardless of whether it is even mounted on a camera body or not ...

Click to expand...

The problem here is that a larger sensor uses a larger area to collect light. This means that because the exposure is identical, a 35mm sensor will collect 4 times the amount of light as a 4/3rds sensor.

Exposure is always measured by the amount of light per unit area. The total amount of light is irrelevant unless the amount of light per unit area is different ...

This is why light meters work for all formats and all lenses of all apertures ...

The "total amount of light collected" argument is one of those "facts" that is completely irrelevant for all practical purposes.

It may be of theoretical interest to some who are interested in the minutiae of the design of different makes, models and sizes of sensors. However, we are "stuck with" the sensor that happens to be in our specific make and model of camera of choice.

Sure I would like the E-P1 sensor and parts of the processing engine to be in my E-30, but I would not be prepared to swap the advantages (to me) of my E-30 for the E-P1.

rovingtim said:
Noticing that the exposure is identical simply accentuates this fact.

Sorry, but I missed the point of what you are saying?

rovingtim said:
I would also like to note that FoV is an asthetic just as DoF is.

I agree.

But this is yet another area where "equivalence" falls down. By stating (effectively ... ) that one cannot have the same DoF in physical terms with the same AoV with 4/3rds and any other format (neither can any of those other formats ... ), "equivalence" is basically self-contradictory. It therefore becomes merely a vehicle for attempting to "prove" the superiority of one format over another.

I can achieve the same DoF (or similar enough for all practical purposes), and according to the author of "equivalence" the corners and edges don't matter anyway, effectively stating that this part of framing is irrelevant ...

So it is really comes down to questions that could be framed as follows:
"Can the required DoF be achieved using a 4/3rds lens?" - Yes, it can.
"Can the required AoV be achieved using a 4/3rds lens?" - Yes, it can.

"Will these parameters be exactly the same for other formats?" - No, they will not.

So I ask "Who cares?". I don't.

BTW, I have now printed the "A1" crop referred to in my previous post.

At about a 2.5 metre viewing distance a photographer friend with good eyesight could not tell the difference between the four images in resolution, other than by observing that they were different crops of the image.
At a viewing distance of 250~400 mm, he could clearly see the difference, specially between the 50 x 37 and A0 prints. However, he found that the A1 and A2 prints were very acceptably sharp even at this viewing distance, when compared with the original A4 print.
He is as much into sharpness and resolution as I am in his own piccies.
Personally, I am extremely happy with the sharpness of the "A2" print, and also with the "A1" print when viewed from a normal viewing distance for such a large print.
This was achieved without any special PP of any description ...

I think that this practical 'experiment' demonstrates that 4/3rds is more than capable of producing very acceptable, very large prints even using an entry level body (E-510) and a SG lens (f2.8/25 pancake).

I would expect my HG and SHG lenses to be even more capable in this regard, and ditto if used on my E-30 ...

It's very late here. My eye is hurting badly (bleed ... ). I am very tired (exhausted ... ). I have to go to bed. Please excuse any errors in the above, lol ...

--
Regards, john from Melbourne, Australia.
(see profile for current gear)
Please do not embed images from my web site without prior permission
I consider this to be a breach of my copyright.
-- -- --

The Camera doth not make the Man (or Woman) ...
Perhaps being kind to cats, dogs & children does ...

Gallery: http://canopuscomputing.com.au/gallery2/main.php

Bird Control Officers on active service.

Member of UK (and abroad) Photo Safari Group

rovingtim · Jul 8, 2010

John King said:
The total amount of light is irrelevant unless the amount of light per unit area is different ...

I have to admit I am not technically savvy enough effectively argue this one way or another, but this seems to be the exact point of dispute.

There is some evidence that the digital chips are responding to the total amount of light collected. I don't fully understand why, but some people have found that if we match resolution, noise seems to follow chip size not pixel size.

However, I while I have seen evidence for this, I think more needs to be explored before we can say exactly what is going on.

Tyrone Wellhung · Jul 8, 2010

boggis the cat said:
Tyrone Wellhung said:

Now here's a thing. For a TC to show up on your camera meter as what boggis thinks is the 'correct' f-number, you need a 'reporting TC' which multiplies the reported f-number by the magnification factor.

Click to expand...

You can determine that it's the correct f-number by the exposure.

No, you can't, because whatever the f-number reported the meter will set it's idea of exposure by reference to the light reaching the meter cells, so you will see a 'correct' exposure. If you compare two cameras, same lens, one with a TC, one without the TC you will notice that the exposures are different, but then it's become a circular argument, the 'correct' f-number is the one that gives the 'correct' exposure.

boggis the cat said:
boggis the cat said:

If you use a plain old non-reporting TC the camera thinks the lens has the original f-number, even though boggis 'knows' its f-number has been increased. Of course, if you do your teleconversion by cropping, there's nothing to do the reporting.

Click to expand...

With a "non reporting" TC it will appear to the camera's metering system that the ambient light levels are lower. The point of having the TC announce its presence is to ensure that the correct f-number is indeed reported.

and what is the relevance of the 'correct' f-number if the exposure is 'correct' in both cases?

Tyrone Wellhung · Jul 8, 2010

rovingtim said:
John King said:

The total amount of light is irrelevant unless the amount of light per unit area is different ...

Click to expand...

I have to admit I am not technically savvy enough effectively argue this one way or another, but this seems to be the exact point of dispute.

There is some evidence that the digital chips are responding to the total amount of light collected. I don't fully understand why, but some people have found that if we match resolution, noise seems to follow chip size not pixel size.

However, I while I have seen evidence for this, I think more needs to be explored before we can say exactly what is going on.

The error is to say that total amount of light is irrelevant. It is very relevant to the appearance of the final image. The total amount of light fundamentally determines the amount of visible photon shot noise in an image, and photon shot noise is the dominant noise in the highlights and mid-tones.

Imagine two 12MP sensors, one double the linear dimensions of the other. The larger sensor's pixels are four times the size of the smaller pixel's sensors. If both sensors have the same photon collecting efficiency (aka quantum efficiency) then for the same exposure the larger sensor will collect four times as many photons in each pixel. Since the shot noise signal to noise ratio is given by the square root of the number of photons collected, this means that the SNR of each pixel in the larger sensor will be double that of the smaller sensor. This is an unescapable physical fact.

To extend that discussion, suppose we fit the larger sensor with a lens double the focal length of the smaller one, but with the same f-number, we find that corresponding pixels in the sensors are collecting light from exactly the same portion of the subject (provided the subject is flat and at the plane of focus), but that the larger sensor is collecting four times as many, so is producing double the shot noise signal to noise ratio. Of course if the subject isn't flat, we'll notice that the larger sensor camera produces a shallower DOF. To get the DOF the same, we would need to stop down the lens on the large sensor camera by two stops, thus cutting by a factor of four the light captured by each pixel, and making the images from the two cameras as identical as could be in terms of the shot noise SNR, the angle of view and, indeed, the DOF.

rovingtim · Jul 8, 2010

This is the larger pixel argument which I believe is well understood.

The question is how will a 40 mp look compared to a 10 mp sensor that has a quarter of the area (half the size). That is more relevant to the amount of light collected argument.

I'm quite sure people have done tests on this ...

Tyrone Wellhung · Jul 8, 2010

rovingtim said:
This is the larger pixel argument which I believe is well understood.

No it isn't. The size of pixel is irrelevant to the argument. I just chose equal pixel counts to make it conceptually easier, it's easy to see then that each pixel must represent the same portion of the subject. I could as easily argue it from the output image. Imaging an A3/ 300DPI print, with 18MP. Each print pixel in the large sensor print represents four times the sensor area as does each pixel in the small sensor print, so the shot noise SNR in the big sensor print would be twice the SNR in the little sensor print. In practice, in a print made with good scene illumination this will be masked by the fact that the print probably has a low enough output DR that the visible SNR will be the much the same - this is less true in low light where photon counts are smaller and the DR of the senor image is smaller than the print can reproduce.

rovingtim said:
The question is how will a 40 mp look compared to a 10 mp sensor that has a quarter of the area (half the size). That is more relevant to the amount of light collected argument.

No it isn't, the number of pixels is irrelevant to the amount of light collected argument, but having different numbers of pixels makes it more difficult to visualise what is going on. Assume that each pixel collects n photons. If we look at a 10MP output image from both (i.e., the same size) then the small sensor camera has a output pixel SNR of n/sqrt

= sqrt

. An output pixel from the big sensor camera aggregates four input pixels, so its SNR is 4*n/sqrt(sqrt(4*n^2)) = 4n/2*sqrt

= 2*sqrt

- it's output SNR is still twice. Of course, if we choose to look at the two images different sizes, all bets are off.

In practice, upsampling the smaller sensor image to match the bigger sensor one has much the same result. Although one can't really subdivide the 10Mp pixel to make four pixels, interpolation algorithms have a good go at it, and in the process deliver artifacts pretty much equivalent to the per pixel noise of a 40MP small sensor, if they are to get anywhere close to having apparently the same detail.

rovingtim said:
I'm quite sure people have done tests on this ...

So far as the cameras actually on the market allow (you can't find a 40MP FF camera to compare with a 10MP FT camera. You might find a 40MP MF camera to compare with a 10MP APS-C camera, but then you'd find that the sensor technologies and efficiencies were radically different.

On the general issue that pixel size is largely irrelevant to noise at any given output resolution, there is this comparison of the Canon 40D and 50D resampled to different resolutions from IR test images.

As you can see, although the visible character and 'sharpness' of the noise changes a bit, there isn't a lot in the amplitude, either up or down sampled.

steve16823 · Jul 8, 2010

jim stirling said:
What does the Olympus 300mm f2.8 come in at ? It is between £5600 and £6000 here in the UK . Thats the same price as for example a D3x + a 300mm f4

Fair enough but I would guess the better Nikon equivalent would be the 600mm f/4 which is even more ungodly expensive than the Olympus 300mm.

Tyrone Wellhung · Jul 8, 2010

steve16823 said:
jim stirling said:

What does the Olympus 300mm f2.8 come in at ? It is between ï¿½5600 and ï¿½6000 here in the UK . Thats the same price as for example a D3x + a 300mm f4

Click to expand...

Fair enough but I would guess the better Nikon equivalent would be the 600mm f/4

It would be a 600/5.6 if they did it

steve16823 said:
which is even more ungodly expensive than the Olympus 300mm.

The Nikon 300/2.8 with a 2x TC will produce the same range of photographs (same AOV, DOF range and photon shot noise at same shutter speeds) and by B&H prices would cost $5899+350 as against $5585 for the Olympus. You'd get a lower price with a Mk 1 Nikon or a Canon ($4495 +$309).
Still leaves you out of pocket on the D3x, though.

illy · Jul 8, 2010

steve16823 said:
jim stirling said:

What does the Olympus 300mm f2.8 come in at ? It is between £5600 and £6000 here in the UK . Thats the same price as for example a D3x + a 300mm f4

Click to expand...

Fair enough but I would guess the better Nikon equivalent would be the 600mm f/4 which is even more ungodly expensive than the Olympus 300mm.

but this is the point, in the UK you're paying close to the same as a 600mm lens for a 300mm lens
--
http://illy.smugmug.com
every day's a curry day

rovingtim · Jul 9, 2010

Tyrone Wellhung said:
No it isn't. The size of pixel is irrelevant to the argument.

Only if the physics is such that the noise rises directly inversely proportional to photons with no relevance to the engineering of the pixel. Surely a smaller pixel could have different physical characteristics to a physically larger pixel.

Tyrone Wellhung said:
Tyrone Wellhung said:

I'm quite sure people have done tests on this ...

Click to expand...

So far as the cameras actually on the market allow (you can't find a 40MP FF camera to compare with a 10MP FT camera.

But doesn't the 7D have the same pixel size as some recent Oly cameras? Also, I think some point and shoots with different sensor sizes have similar sized pixels.

I'm sure someone has done some research into this.

boggis the cat · Jul 9, 2010

Tyrone Wellhung said:
boggis the cat said:

You can determine that it's the correct f-number by the exposure.

Click to expand...

No, you can't, because whatever the f-number reported the meter will set it's idea of exposure by reference to the light reaching the meter cells, so you will see a 'correct' exposure. If you compare two cameras, same lens, one with a TC, one without the TC you will notice that the exposures are different, but then it's become a circular argument, the 'correct' f-number is the one that gives the 'correct' exposure.

If you have a set scene available, try it out. A TC that reduces the f-number 1 stop produces an EV shift of -1 stop; a TC that reduces the f-number 2 stops produces an EV shift of -2 stops.

I don't have a TC for my Olympus system, but here is what happens with my Canon film SLR (EOS 10 and 35-135 f/4-5.6, at 35 mm f/4) and 2x TC (Tamron 2X C-AF BBAR MC7):

No TC 1/15 s
2x TC 1/4 s.

It works exactly as it should.

Edit:

Just an after-thought -- it is a film camera, so no clever reporting by the TC or other hi-jinks can take place. The camera doesn't know the TC is there, it just meters for the scene.

Tyrone Wellhung · Jul 9, 2010

rovingtim said:
Tyrone Wellhung said:

No it isn't. The size of pixel is irrelevant to the argument.

Click to expand...

Only if the physics is such that the noise rises directly inversely proportional to photons with no relevance to the engineering of the pixel.

The physics says that the shot noise signal to ratio between like size portions of the image projected onto the sensor is given by the square root of the number of photons in that portion of the image. The question of the proportion of those photons that end up being counted, and therefore forms the recorded SNR is a matter of the quantum efficiency (QE) of the sensor. The physics says the QE can't be more than 1, the engineering dictates how close to 1 it gets.

rovingtim said:
Surely a smaller pixel could have different physical characteristics to a physically larger pixel.

There doesn't seem to be a very clear linkage between pixel size and QE. The highest QE sensors out there, with just about the same QE (0.55 or so) are the Nikon D3S sensor and the Sony sensor in the Canon G11 (and several other cameras). The D3S has about the largest pixels to be found in a DSLR, the G11 has pixels far smaller than found in any DSLR. As a general trend, smaller pixels seem to have higher QE than bigger pixels, but that could well just be because smaller pixels tended to be developed later and have newer technology than bigger pixels.

In the end, the differences in QE seem to come down to microlens design these days, the intrinsic efficiency of the silicon is high, so the QE is determined by how good the microlenses are at directing the photons in the image to the photoreceptors. In practice, what they have to do is form an image of the lens exit pupil on the photoreceptor, such that none of that image spills over the edge of the photoreceptor. With a very small pixel/small photoreceptor you need a small FL microlens, a bigger photoreceptor allows a longer FL, which can be a good thing, because the plastic microlens layer has to sit in top of the chip metallisation, so can't get very close to the silicon of the photoreceptor. On the other hand, the microlens has to have a very small f-number. Imagine using an f/16 microlens to image the exit pupil of an f/2 lens, the number of photons collected will be limited by the f/16, not the f/2. There are different solutions to these conflicting constraints. In the D3S, Nikon uses a dual element aspherical microlens presumably to provide a very low f-number microlens. Small pixel sensors get over the long microlens back focus problem by building in a 'light pipe' which carries the light focussed by the microlens to the photoreceptor, or by using back side illumination, which allows the microlenses to be much closer to the photoreceptor, since the chip metalisation is on the other side. In any case, whatever the engineering solutions employed, there doesn't in practice seem to be a whole load of QE difference between different sized pixels.

rovingtim said:
rovingtim said:

rovingtim said:

I'm quite sure people have done tests on this ...

Click to expand...

So far as the cameras actually on the market allow (you can't find a 40MP FF camera to compare with a 10MP FT camera.

Click to expand...

But doesn't the 7D have the same pixel size as some recent Oly cameras? Also, I think some point and shoots with different sensor sizes have similar sized pixels.

But it still isn't your 40MP vs 10MP test. The 15MP vs 10MP test I showed you seems pretty conclusive to me. There's no reason to expect a 18MP to 10MP test to produce radically different results.

rovingtim said:
I'm sure someone has done some research into this.

I'm pretty sure no one has, since there is no point. All this is well established physics. There is no real mileage in setting up complex experiments to verify what is already well known (apart from on DPR).

Sensor designers know the physics of sensor design, and they will test their engineering designs, and publish the outcomes (in terms of QE) in their data sheets. from that, an engineer can predict the shot noise outcome for different illumination levels.

Tyrone Wellhung · Jul 9, 2010

boggis the cat said:
Tyrone Wellhung said:

boggis the cat said:

You can determine that it's the correct f-number by the exposure.

Click to expand...

No, you can't, because whatever the f-number reported the meter will set it's idea of exposure by reference to the light reaching the meter cells, so you will see a 'correct' exposure. If you compare two cameras, same lens, one with a TC, one without the TC you will notice that the exposures are different, but then it's become a circular argument, the 'correct' f-number is the one that gives the 'correct' exposure.

Click to expand...

If you have a set scene available, try it out. A TC that reduces the f-number 1 stop produces an EV shift of -1 stop; a TC that reduces the f-number 2 stops produces an EV shift of -2 stops.

I don't have a TC for my Olympus system, but here is what happens with my Canon film SLR (EOS 10 and 35-135 f/4-5.6, at 35 mm f/4) and 2x TC (Tamron 2X C-AF BBAR MC7):

No TC 1/15 s

2x TC 1/4 s.

It works exactly as it should.
Tyrone Wellhung wrote:

If you compare two cameras, same lens, one with a TC, one without the TC you will notice that the exposures are different, but then it's become a circular argument, the 'correct' f-number is the one that gives the 'correct' exposure.
Look at it a different way, maybe you'll see it.

Imagine we have a 12MP camera and take a photo of a grey card such that n photons are recorded at every pixel location. Now we fit a 2x teleconverter in front of the lens. This magnifies the image projected onto the sensor to four times the area, but the number of photons in that image is the same. This means that each pixel location records a quarter of the number of photons, n/4.

Now we take the TC off and mount the lens directly. Since our camera is a modular E-5 we can slip out the 12MP sensor and replace it with a 48MP one. With this new sensor the pixels each has a quarter of the area, so each records n/4 photons. We now crop out the centre 12MP and compare the resulting image with the one taken with the TC. We find that in both cases we have 12MP with n/4 photons recorded at each pixel location.

boggis the cat said:
Edit:

Just an after-thought -- it is a film camera, so no clever reporting by the TC or other hi-jinks can take place. The camera doesn't know the TC is there, it just meters for the scene.

A film camera is a completely different kettle of fish. If we use the same speed film, n/4 photons over 1/12Mth of the frame will produce a completely different density from n/4 photons over 1/48Mth of the frame, because film records light non-linearly. A digital sensor, on the other hand, records light linearly. There is no 'density' until we read the photon count recorded in each pixel and use software to map that count to some density.

Tyrone Wellhung · Jul 9, 2010

Just reviewing the discussion in this thread, I just thought that some misconceptions about the 'equivalence' case need to be rectified here, and since no-one else has...

John King said:
"Equivalence" as espoused makes the totally unwarranted assumptions that :

Somehow aesthetic parameters (such as DoF desired) are part of the exposure parameters required to make an exposure "correctly" as desired by the photographer (these considerations are part of choosing which member of an exposure parameter set are used by the photographer, as is the FL of lens and format used ... ); and

Equivalence doesn't make such an assumption. If the desire is to set a particular DOF using a given format, then physics dictates the outcome in terms of illuminance at the sensor for a given scene luminance.

John King said:
One has to use the same lens (or a similar one, i.e. with a similar effective focal length {= EFL}); and

Equivalence does not dictate which lens any photographer may or may not use. The need to fit different FL lenses to achieve the same angle of view with different formats is, as you pointed out, not controversial.

John King said:
One is not "allowed" to use a lens with a different EFL to achieve the aesthetic effects that one might achieve by using a different lens on a different format; and

As said above, no version of equivalence I have ever seen attempts to dictate to any photographer what he is allowed or not allowed to do.

John King said:
One actually wants to achieve the same DoF effect using two or more different formats! For me, part of the purpose of using different formats has always been precisely the opposite of "equivalence" - I do so in order to achieve different results ...

That is of course your choice. The notion of equivalence allows one to predict easily how to set the controls to achieve a desired DOF. Many pictorialists wish to control DOF, not take a 'one size fits all' dictated by the system they are using. This approach avoids having to pack multiple camera/lens systems should one wish to take photos with different DOF's during the same assignment.

John King said:
None of these assumptions are necessarily correct for any given photographer. I would even go so far as to say that most photographers could not care less about these things.

I wouldn't presume to talk for most photographers. While you may not be so fastidious, the fact that most photographers are not packing Olympus gear is quite good evidence that their views do not necessarily align with yours.

John King said:
John King said:

Picking only the part of equivalence that is advantageous is misleading. This habit caused much bad blood on this forum.

Click to expand...

What causes the " bad blood " is the continued pushing of "equivalence theory" down our throats, usually with the corresponding agenda of using it to demonstrate the inferiority of 4/3rds as a system ...

The 'pushing down the throats' seems to be quite popular with many. This thread

http://forums.dpreview.com/forums/read.asp?forum=1022&message=34702202 is number 5 in the most bookmarked on this forum. Once again, I'm not sure it's a safe assumption that everyone's views align with yours.

John King said:
John King said:

Oly lenses are more expensive and Oly bodies are cheaper. Whether the Oly system works for a user is up to that user. If it works, then it is irrelevent if another system is better for someone, else isn't it?

Click to expand...

This has always been my point as well.

It is not really that much different from those who constantly "inform" us here that one might be able to make a print up to (shock, horror ... ) even as large as 11 x 17 inches using a 4/3rds camera ...

Yesterday I printed some crops that are equal to a full size print of:

50 x 37 inches (150 dpi) - not critically sharp at 10" viewing distance, but indistinguishable in sharpness from the "A2" crop at about 2.5 m viewing distance,

A0 (46 x 33") (300 dpi) - still not critically sharp at 10" viewing distance, but indistinguishable in sharpness from the "A2" crop at about 2.5 m viewing distance and

A2 (33 x 23") (300 dpi) - critically sharp at 10" viewing distance, even using a 5x loupe at about this distance. At a "normal" viewing distance for a print this size, the "A2" print appears razor sharp ...

I have not yet printed an "A1" size crop. I am going to try to get that done today.

These are all slightly larger than 11 x 17" ... .
An A2 print is 4x the area of an 11 x 17" print ...

BTW, I did not take the source image with either of my "pro" bodies.
I took it with my E-510 and SG f2.8/25 pancake.
The crop is taken from the very edge of the image ...

I'm not sure what is the point being made here. I cannot recall ever seeing anyone suggest that the Olympus E system is not capable of excellent results.

John King said:
John King said:

When people post hero worship fantasy while comparing Oly to other systems, responses are not only likely, but probably justified.

Click to expand...

And when people post " ... hero worship fantasy ... " in the OSTF about "FF" or even better, about APS, that's OK?

Maybe you could give an example of such a post. I don't recollect them being common.

No "Bad" Olympus digital lenses...

Veteran Member

Veteran Member

Senior Member

Veteran Member

Senior Member

Veteran Member

Forum Pro

Veteran Member

Leading Member

Leading Member

Veteran Member

Leading Member

Senior Member

Leading Member

illy

Guest

Veteran Member

Veteran Member

Leading Member

Leading Member

Leading Member

Similar threads

Keyboard shortcuts