Imaging Chip Size

[BJL]

Since this issue of sensor size and focal length equivalency keeps coming up and keeps being controversial, here is my 2c worth.

   What really allows one to use shorter focal lengths for telephoto shots (ignoring different depth of field behaviour) is the sensor having higher resolution, coming from smaller photosites (and good enough lenses). I mean resolution in the official sense, measured in "line pairs/mm", not total detail ("pixel count"). More resolution is what allows greater magnification from sensor image to print, and hence lenses of shorter focal lengths.
   So a full 35mm format sensor could work with the same focal lengths as the Olympus E-1 so long as it had 6.8 micron photosites, or about 18MP, and it could match lenses with the 10D by having 7.5 micron photosites, or about 15MP: one could just crop for long telephoto shots. This is sometimes called "smart digital zoom". (The more common "stupid digital zoom" of many digicams interpolates up after the crop to pretend that the image has full sensor resolution.)

   I suspect that for a long time yet, and perhaps forever, such huge additional pixel counts will incur substantial costs for sensor and in-camera processing hardware, and will slow frame rates substantially, and so this strategy of buying lots of extra pixels and then using "digital zoom" for long telephoto coverage will not in fact be popular. Thus I expect that smaller sensors with "only" 5 to 10 MP will have a big niche with photographers who do not need the extra total detail enough to pay the extra price.

[RalphC]

One long term advantage of smaller chips is the "multiplication" factor...

There is no "multiplication factor"--it is a cropping factor. The only advantage is as a marketing tool to make unsuspecting buyers think they are getting something they are not.

[flash]

One long term advantage of smaller chips is the "multiplication" factor...

There is no "multiplication factor"--it is a cropping factor. The only advantage is as a marketing tool to make unsuspecting buyers think they are getting something they are not.

I have seen this all too often recently and I must say "so what?" Who really cares if it is or isn't a magnification or cropping or whatever. A 50mm lens on a Canon DSLR gives exactly the same angle of view as an 80mm lens on a Canon film SLR. If everbody accept this why do we have to repeat the obvious over and over and over and over......

Secondly it is not a cropping factor. You are using the full size of the sensor. Cropping would only be correct if you were cropping at the sensor plane.

Thirdly new lenses like those from Nikon  use the full image circle of the lens on their digital SLRs. These are not cropping or magnifying. That's like saying a Canon 80mm on a EOS1v is a crop of a 80mm lens on a 6x6.

End of Rant

Gordon

[BJL]

First, a comment on "multiplication factors" or "crop factors" or whatever one calls them: they do serve one simple, legitimate purpose, but cause much confusion in the process. The legitimate purpose is to help people understand the angular field of view they will get with a certain lens in a new, unfamiliar digital format by referring back to their 35mm format experience: multiplying the actual lens's focal length by this factor tells you the focal length of an imaginary 35mm format lens that would have the same field of view. It does not guarantee that any other optical properties of the actual lens or the images produced are the same as they would be with that imaginary 35mm lens. One could say that 645 medium format has a magnification factor of about 0.6x relativ to 35mm format, but no-one does, because medium format users generally do not need to refer back to 35mm format experience in order to understand the field of view of their lenses.

   About your examples with standard lenses.
a) The angular field of view and perspective would be about the same (i.e. the ratio of sizes of foreground and background objects will be the same) due to using the same camera position.
 The depth of field seen on prints of the same size will be different, since you specified the same aperture ratio: to a good approximation, the DOF will be greater with the smaller format and shorter focal length lens in proportion to the focal length change. Detailed formulas and discussion can be found in another thread on this site!
c) However, if you adjust the aperture ratios in proportion to sensor size and focal length, say f/8 with 80mm in 35mm format and f/5 with the 50mm lens in a format 1.6x smaller, you will get about the same depth of field in each case.
   So as long as you have an adequate range of aperture ratios available, you can get the same combinations of field of view, perspective and depth of field in various different formats. With very small digital sensor formats, it becomes impractical to achieve the very larger aperture ratios needed to get extremely shallow DOF.
d) You did not mention one other ingredient in the appearance of the image: shutter speed. For a smaller format, equal DOF requires a larger aperture ratio, changing the combinations of shutter speed and sensitivity ("ISO") needed: for better or worse, you get more speed with smaller formats (which is why there is not a lot of medium or large format action photography!). If you wish to hold shutter speed constant too, the smaller format needs to be used at lower sensitivity (lower ISO), or with a neutral density filter.
   On the other hand, to avoid sacrificing too much resolution, smaller sensors tend to have smaller photosites, which reduces maximum usable sensitivity, and if the pixel count is kept equal between formats, this should roughly balance out that speed increase. A bit like the greater need for fine grained, high resolution, low ISO films to get good quality with 35mm compared to medium format.

[BJL]

A couple of comments about "cropping".

   On one hand, using 35mm format lenses on a smaller format DSLR does involve some cropping of the image produced by the lens; I will leave aside arguments about both the advantages (like avoiding the often lower image quality in the corners of the full frame) and disadvantages (requiring greater magnification to get a print of given size, hence putting greater demands on the resolution of the lens.)

   However, it should be noticed that significant cropping is almost inevitable with any telephoto work: it is easy and natural for lens designs of any focal length to cover a roughly "normal" angular field of view, about 50 degrees, and any lens of narrower angle of view (telephoto) is cropping a bigger image down to what is wanted in a particular situation: this cropping is done by some combination of lens hoods, internal anti-flare baffles, and the sensor/film frame itself.
   Roughly, the degree of cropping that happens depends on the angular field of view used compared to the roughly 50 degree angle of the image circle. For example, the amount of cropping from the full available image circle is the same for 80mm in 35mm format as it is for 50mm on a D30/D60/10D/300D.

   Thus, smaller formats do not involve any more or less cropping of the available image circle to get a given angular field of view, except when one takes lenses designed for normal to wide angle coverage in one format and uses them with a smaller format.

[AJSJones]

Seems to me that the only real "issue" with crop factor ( or whatever else we want to call it ) is the fact that besides size and weight of the lens, you're paying for a lens with coverage that you don't use.  The image circle of any lens is always where the cost / benefit / excellence / size, contortions of any mfg. are wrought.

So, if "all factors" were equal why pay for a lens that will cover a full 35mm frame ( at any given level of excellence, and that will be larger because of this ) when you'll never cover a full 35mm frame if one is "fully digital" and can't afford the likes of a 1Ds "ever?"

Why don't we call it a "format factor"?
 After all 35mm is not the origin of all pictures; however, it has dominated so people feel comfortable saying it's "standard", so we could call it 1.  Then the 10D would be 0.6 ish, as would APS, P&S digicams would be 0.2 or less (most of them), 645 would be 1.7 ish and 4x5 film would be 3.7 ish (quick and dirty diagonal comparisons).  Divide the lens's true (and unchanging) focal length by this factor to get roughly the FL of the 35 mm equivalent needed to get the same image.
 
 Paraphrase of Ansel Adams's reaction to 35mm format : "Nice miniature format you have there."

Scott, As you note there are a variety of parameters to be balanced when designing lenses.  As a 10D user, I'm happy that the lenses I use were optimized for 35mm FF because the lens performance in the image circle area captured by the smaller sensor is inevitably better than that at the edges, so I get a "better" lens.  If Canon started to make a whole series of "digital" lenses, i.e. smaller and lighter as a result of trimming the image circle to match the smaller sensors, they would face the same issue of maintaining performance at the edge of the circle (The EF-S 18-55 comes to mind!)- so many folks would still buy the lens optimized for the 35mm image circle.  I suspect I would.  
 
 Andy

[R Scott Adams]

Why don't we call it a "format factor"?

Andy.... agree completely!

As to the other.... well, if that were all there was to lens design ( make it good to the edges of the image circle and everything else will be even better toward the middle ), I'd agree with you.  Sadly, that's not always the case.

Scott

[AJSJones]

BJL,
I was not trying to imply that a lens designed for a smaller image circle would automatically be worse - just that in balancing all the design criteria Scott listed, one could run into the same compromise resulting in a "small image circle" lens having performance at the edges of its image circle that is not as good as the central part of a lens designed for the FF sensor.  If the designers focused on performance across the whole (smaller) image circle, there's no reason to think they couldn't do it - rather it could end up either more expensive or with some other compromise  such that the FF lens might still be preferred by a prospective purchaser.

Whether one uses 35 mm as a "standard" or some other baseline, we agree that the key parameter is the image diagonal.  However, I'd venture to say that 35 mm is already pretty well-entrenched and is as good as any to use as a base - viz. most digicams are described in those terms already - I was only suggesting a way out of the interminable "crop vs. magnify" morass with something a little more acceptable and quite easily explained in familiar terms.  Which is better - the foot or the meter   The resistance to change is based in familiarity with the present!

Andy

[Jack Flesher]

I like "format factor" as the appropriate term...

[Ray]

In fact, resolution needs are a good reason to think that lenses designed for the job (smaller image circle) are likely to perform better than ones originally designed for a diferent purpose. Canon and many lens makers measure lens sharpness through MTF curves at 30 lp/mm, and probably their designs and quality control are oriented to good performance at upto 30lp/mm. For smaller sensors and smaller pixels, higher resolutions are important, and thus Olympus works with MTF 60lp/mm for its E system lenses. The E lenses' MTF curves at 60lp/mm are close to Canon L lenses at 30lp/mm. (As an aside, according to tests at the PhotoDo site, even very good medium format lenses tend to have somewhat worse MTF at 40lp/mm than 35mm format lenses of the same focal length, but the difference is not enough to offset the extra detail given by the larger image size.)

  

BJL,
You have a point. A quick look at the Photodo site will reveal that even the best MF lenses rarely get a rating above 4. If a smaller than 35mm format such as the 4/3rds catches on, one could expect eventually a greater proportion of lenses for such cameras will tend to have a higher rating than 35mm lenses.

I think the highest rating Photodo has given to a lens is 4.8 (Canon 200mm F1.2, is it?)

However, it's reasonable to suppose that such lenses are still going to be very, very expensive. Most people will have to settle for less and the cheaper 4/3rds format lenses are still going to suffer from a worse performance at the edges and at full aperture. This seems to be a lens fact of life.

And not just worse performance in relation to the the same lens' performance at the centre, but worse performance compared with a comparably priced 35mm lens as measured 12mm from its centre.

Of course, this effect is going to be much greater with wide angled lenses. The Sigma 14mm prime was considered to be quite good with the D30, but unacceptable on a 1Ds. When I was faced with the choice of getting a Sigma 15-30 zoom or the much more expensive Canon 16-35, I compared them in the shop, photographing newspapers using my D60. I could see no performance advantage to the Canon. Both series of shots seemed equally sharp corner to corner. I bought the Sigma and saved a heap of money.

[Ray]

Comparing the same lens, or lenses of the same focal length, with different sensor sizes is not the relevant comparision; or at least, not the one I am making.

Not directly relevant, true. However, I recall from the days I was using film, both my Sigma 20mm lens and highly regarded Sigma 24mm prime (as good or better than the Canon) had problems in the corners and at the edges. I don't get a sense of this using the Sigma 15-30mm zoom (at 15mm) with the D60.

It's a pity Photodo have withdrawn their MTF charts because these were very revealing of edge performance. At 40 lp/mm, the MTF response of all but a small handful of lenses takes a steep dive in that part of the frame between about 12mm and 22mm from the centre - just about the same area that is cropped by the 10D. Some of the cheaper wide angle lenses seem to have almost zero MTF response at 40 lp/mm in the corners, at the larger apertures, and not at all impressive at f8.

But chop off those curves at the 12-14mm point along the horizontal axis, and a relatively cheap lens begins to look like a really expensive lens.

One should also bear in mind that the difference in size between the 4/3rds format and 35mm is similar to the difference between 35mm and the next step up to 6x4.5. I don't think the best Canon 50mm lens is much better over all than the best Mamiya 80mm lens. But I'd bet the Mamiya 80mm would outperform the Canon at a point 20mm from the centre of the frame, at any aperture you'd care to nominate.  

(Sorry! That last sentence is a bit confusing. I meant the 20mm point on a 6x4.5 lens would have about the same angle of view as the outer corner of a 50mm lens designed for the 4/3rds format, although I haven't done the maths. I might be out a bit, but you get the drift.)

To take a specific example (I managed to download a few of Photodo's MTF charts before they withdrew them). Let's compare 2 lenses; the Hasselblad Planar CF 80/2.8 with a Photodo rating of 3.7, and the Minolta 50/1.4 with a rating of 4.4.

At f8 and 40 lp/mm and 22mm from centre, the MTF of the 80mm lens is about 47% for both curves.

The 50mm lens at 21mm from centre shows an MTF response of less than 10%[/i]. Okay, let's move away from the absolute corner. At 18mm from the centre the response is better; 50% for the solid line and 28% for the broken line. Moving a bit closer to the centre, say 15mm - to just outside the cropped area of a 10D, the situation improves dramatically for the meridional curve (about 74%) but the sagital curve is still in the doldrums and lower than the Hasselblad sagital curve 15mm from centre.

I don't have any reason to suppose that lens technology applied to the smaller format lenses will be able to solve this problem of lower performance at the periphery of the image circle. If any advances in technology do make it possible, then such advances can surely be applied to all lenses across the board.

As a matter of interest, the MTF response of the Canon EF 200/1.8 at 40lp/mm is superb, and at full aperture. So it can be done. (For both lines, it fluctuates between 70% and 60% right out to the edges. What a lens!)

[dbarthel]

One long term advantage of smaller chips is the "multiplication" factor. I shoot a lot of birds, so there is never too much length. My standard setup is the 500mm F4 + 1.4x teleconverter for base. This is a very managable setup weight wise, and can be hand held in a pinch. On my 10D this becomes an 1100mm lens @5.6. If I had to have a real 1100mm f5.6 lens, I'd need a crane to setup and move the lens (remember the Canon 1200mm?). So for sports and wildlife I hope there will be a long life ahead for less than full frame sensors. Interestingly, if you crop a 1Ds image to the size of the 10D, you lose some pixels, so full frame and cropping is not a valid suggestion. I do strongly agree with Michael that given both Canon and Nikon have small sensor solutions, Olympus has a tough up hill battle to sell both a format and a lens arsinal to a new user. Canon folks have the best of all worlds: 1.6x, 1.3x and full frame.

[Dean]

Do most DSLRs have a "multiplication factor" or a "crop factor"?  In other words, is the use of a smaller sensor equivalent to using a longer lens?  I suspect that the answer is yes and no, depending on the citcumstances.  Imagine three cameras all set up at the exact same location, each with the same hardware and software, except one camera has a smaller sensor. All the cameras take a photo at the same moment.  (I know that this senario is not possible, but a mind experiment allows us to do the impossible.)  In addition, the full frame cameras have lenses that perform exactly the same, as far as resolution, contrast, color, flare, etc., but one lens is a 50mm and the other lens is an 80mm.  The camera with the smaller sensor has an exact clone of the 50mm lens, which gives the same field of view on the small sensor camera as the 80mm does on the full frame camera. (Is it really possible for the field of views to be the same?)  In summary, all things are held constant other than the size of the sensors and focal length of the lenses.  In this manner, we can imagine the effect of diffedrent sensor sizes without being influenced by other changes.  PHOTO NUMBER ONE: there is a person in the near foreground with mountains in the distance, the lenses are all focused on the person and shot at f2.8.  All photos are processed identically and printed identically (and there is no difference in appearance due to the use of more or less pixels), except the photo made with the full frame camera and 50mm lens is cropped to the exact size as the small sensor photo.  In other words, we now have three 8 x 12 prints that are very similar, but are they the same?  Will they all show the same field of view?  Will the distant mountains in the print shot with the 80mm lens be as sharply focused as in the other two prints?  Will the mountains shot with the 80mm lens appear to have been closer than they appear in the other two prints?  I expect that the two prints shot with the 50mm lens will be exactly the same, but that the print shot with the 80mm lens will be noticably different.  Am I correct?  If I am correct, then in this situation the small sensor camera has a "crop factor" and use of the smaller sensor is not equivalent to using a longer lens.  PHOTO NUMBER TWO: This photo is the same as in photo number one, except that all photos are shot at f8, which results in the distant mountains being in equally sharp focus in all three prints as far as can be discerned by looking at the 8 x 12 prints.  Will all three prints be identical, or will there be a difference in the field of view or apparent distance of the mountains?  I expect that the photo shot with the 80mm lens will still appear to be different.  Am I correct?  PHOTO NUMBER THREE: In this photo there is nothing in the foreground.  The photo is just of mountains.  The nearest object in the photos is at least 1/2 mile away.  Will the three prints be identical?  They should all be equally sharp, but will they all show the same field of view?  Will there be any difference in the apparent distance of the mountains?  I expect that all three prints will be identical.  Am I correct?  If I am correct, then in this third senario the small sensor camera does have a "magnification factor" and it is equivalent to using a longer lens.  ---  If my expectations are correct, then it would appear that, in real world landscape photography where there are near and far objects in my photo, it would not be possible to to get the same photo using a full frame camera and a smaller sensor camera.  Not that one would necessarily be better or worse, just different.   For example. when there are foreground and distant objects in a photo, even a small change in the position of the camera will change the apparent relationship of the near and far objects.  Therefore, in that circumstance,  if I change the position of the camera, change the focal length of the lens or change the sensor size, or any combination thereof, I will also change the photo, with one exception.   I may be able to get similar photos, but they would be noticably different.  The lone exception is to shoot with a full frame camera and smaller sensor camera using the same lens and then crop my full frame photo to match the size of my smaller frame photo.  On the other hand, if all objects in my photo are very distant, then I may be able to obtain virtually identical photos with a full frame camera and one with a smaller sensor by just changing the focal length of my lens. Am I on the right track here or have I made a mistake somewhere?

[jeffreybehr]

Do most DSLRs have a "multiplication factor" or a "crop factor"?

Yes, 'most' do.  Only the Canon 1Ds and the now-discontinued Contax (?) have full-frame (= same size as 35mm film) sensors.

The cropping factors vary from the Canon D30/D60/10D/Rebel's 1.6 to Nikon's 1.5 to Canon 1D's 1.3.  The smaller the sensor, the higher the cropping factor.

[RalphC]

One long term advantage of smaller chips is the "multiplication" factor...

There is no "multiplication factor"--it is a cropping factor. The only advantage is as a marketing tool to make unsuspecting buyers think they are getting something they are not.

I have seen this all too often recently and I must say "so what?" Who really cares if it is or isn't a magnification or cropping or whatever. A 50mm lens on a Canon DSLR gives exactly the same angle of view as an 80mm lens on a Canon film SLR. If everbody accept this why do we have to repeat the obvious over and over and over and over......

Secondly it is not a cropping factor. You are using the full size of the sensor. Cropping would only be correct if you were cropping at the sensor plane.

Thirdly new lenses like those from Nikon  use the full image circle of the lens on their digital SLRs. These are not cropping or magnifying. That's like saying a Canon 80mm on a EOS1v is a crop of a 80mm lens on a 6x6.

End of Rant

Gordon

I admire your passion, unfortunately you are totally wrong.

A rant without research is like catsup without tomatoes.

[jeffreybehr]

Secondly it is not a cropping factor. You are using the full size of the sensor. Cropping would only be correct if you were cropping at the sensor plane.

If one understands that the situation includes 35mm-framed camera bodies, lenses made for those bodies, and the image created by the sensor, COMPARED with the image that would be on 35mm film, it is indeed a crop, and it is indeed at the sensor and film plane.  Even if you're including anyone's lenses designed for the smaller digital sensors, the sensor's image STILL is a crop compared with the image that would be on 35mm film.

[flash]

I admire your passion, unfortunately you are totally wrong.

A rant without research is like catsup without tomatoes.

Wrong? How? I don't see you presenting any alternative. As far as I know what BJL said is absolutely spot on.

Cropping to me implies that I take an image or image sensor and use only part of it by using a mask on the sensor or printing only a part of the recorded information. Therefore a DSLR does not "crop". It uses a completely different format. It is like saying 35mm is a crop of 6x9 which is a crop of an 8x10. Just because the entire image circle isn't used doesn't mean an image is a crop does it?

We use the term "crop" or "magnify" when in fact neither occur. I would personally prefer to say that lens "X" on a Canon/Nikon DSLR has the same angle of view as a "Y" on a 35mm film camera or a "Z" on a 6x9 camera.

Please feel free to show me where I went wrong. Your katsup analogy would apply equally to someone who argues without actually presenting a case.

Gordon

[R Scott Adams]

Seems to me that the only real "issue" with crop factor ( or whatever else we want to call it ) is the fact that besides size and weight of the lens, you're paying for a lens with coverage that you don't use.  The image circle of any lens is always where the cost / benefit / excellence / size, contortions of any mfg. are wrought.

So, if "all factors" were equal why pay for a lens that will cover a full 35mm frame ( at any given level of excellence, and that will be larger because of this ) when you'll never cover a full 35mm frame if one is "fully digital" and can't afford the likes of a 1Ds "ever?"

[BJL]

I see one long term problem in defining a "format factor" relative to 35mm; in not too long, the great majority of all cameras sold will be below 1/4 on this scale. this is because (a) digital camera sales are about now passing film cameras [editted to remove reference to anes item that has been denied by Nikon] ( digital cameras sales are dominated by compact digicams with sensors in 1/1.8" format and smaller; even 2/3" sensors are now used mostly in the relatively small niche of more expensive EVF cameras with big zoom lenses.

   So rather than entrench reference to a format that the great majority of future camera users will never use (or even if they do keep using it), why not use a simple old-fashioned measure of format size: image diagonal? 35mm is about 43mm diagonal, APS-C about 30mm, 4/3 is 22.5mm.

   Then field of view comparisons can be based on the ratio of focal length to image diagonal: an 80mm lens is an 80/43 or about 2x in 35mm format, about a 3x on a 10D, and so on. This FOV measure is a magnification relative to the normal lens for that format, and does not involve giving any special status to any one particular format.

[BJL]

As a 10D user, I'm happy that the lenses I use were optimized for 35mm FF because the lens performance in the image circle area captured by the smaller sensor is inevitably better than that at the edges, so I get a "better" lens.

I have heard this argument many times: (a) lens performance is worse near the edges than at centre, so cropping avoids the worst part of the image [maybe] ( if a lens the same focal length is then designed specially for a smaller format, and so needing only a smaller image circle, the bad performance will show up closer to the centre, and so be more noticable within the smaller frame.

   Part ( makes absolutely no sense to me. As I have said in my discussion above of cropping, for normal to telephoto lenses, the image circle is easily big enough anyway, so the main design changes would be things like having smaller openings in any lens hood and interior flare control baffles, which would provide better flare control, and so possibly improve image quality. With shorter focal lengths, the change would be from a wide angle design (field of view significantly more than 50 degrees) to a design of normal or at least less extremely wide angle coverage. Avoiding the need for very wide angle coverage allows simpler, smaller, cheaper and generally better lens designs.

   Either way, designing lenses for the image size needed allows design flexibility that can increase quality; there is no reason to think that it would reduce quality compared to using lenses of the same focal length but designed for a larger format. To suggest otherwise, you would have to think that the people designing lenses for smaller formats are too stupid to realize that something they are doing is degrading image quality near the corners; I do not think that the optical designers at Olympus, Nikon or Pentax are that stupid.

    Canon has not yet really designed from scratch a smaller image circle lens; they apparently produced their 18-55 AF-S lens by modifying an existing, cheap, slow, 22-55mm lens design left over from their APS SLR effort, and that was originally designed for full 35mm format coverage. This cheap stop-gap product is not a good measure of what a smaller format lens can achieve; the Nikon 12-24 DX, Olympus 14-54 E and the forthcoming Nikon 17-55 DX are far better test cases.