DOF and Micro Four-Thirds Format

[Dan Kehlenbach]

Hello:

I recently purchased a Panasonic G1 and have been very impressed so far with its performance.

I do have a question about depth of field with the micro four-thirds format:

Since the distance from the lens to the sensor is shorter, and since the 4/3 sensor is smaller than my APS-C Canon DSLR, will there be an inherent increase in depth of field for a given f-stop?  Would there be such things as "f-stop equivalents" between micro four-thirds and APS-C sensors?

Thanks for your help.

Dan Kehlenbach

[DarkPenguin]

It's about a stop.

[BJL]

will there be an inherent increase in depth of field for a given f-stop?  Would there be such things as "f-stop equivalents" between micro four-thirds and APS-C sensors?

Here is my rule of thumb (but beware; many forum wars have been fought on this topic!)

Short answer:
About 2/3 stop difference between Four Thirds and EF-S or DX, 2 stops from Four Thirds to 35mm (24x36mm).
So for example, the f/2.8-3.5 lenses for Four Thirds are comparable to f/3.5-4.5 lenses for EF-S/DX for DOF wide open.


Long answer:
When photographing the same FOV with different formats, and so using focal lengths that differ in proportion to format size, you will get equal DOF if you adjust the f-stop in proportion to the change in focal length.
In other words, the same conversion factor for "equivalent focal lengths" (same FOV) also converts for "equivalent f-stops": equivalent for DOF at least(*).

Since the format conversion factors are
4/3 -> 35mm: 2x
4/3 -> EF-S: 1.25x
4/3 -> DX: 1.33x
EF-S -> 35mm: 1.6x
DX -> 35mm 1.5x
the corresponding corrections in f-stops are roughly
4/3 -> 35mm: 2 stops
4/3 -> EF-S: 0.64 stops
4/3 -> DX: 0.82 stops
EF-S -> 35mm: 1.36 stops
DX -> 35mm 1.19 stops
So about 2/3 stop between Four Thirds and either EF-S or DX.



(*) Equivalent for diffraction effects too, and maybe for "speed", once ISO speed differences are allowed for.

[Dan Kehlenbach]

Thanks for the help - I am looking forward to getting out with the camera and experimenting a bit with some landscapes and close-ups.

All my best,
Dan Kehlenbach

[Er1kksen]

It's also easy to remember this way: While a 14mm lens has the same apparent field of view on m4/3 as a 28mm lens would on 35mm, it's still a 14mm lens, and will have the same dof as a 14 mm lens on 35mm. A 50mm lens on m4/3 will have the field of view of a 100mm lens on 35mm, but still the dof of a 50mm lens. Field of view changes with sensor size, while dof is constant based on lens focal length and aperture, regardless of sensor size.

[Dan Kehlenbach]

Err1ksen,

Your explanation certainly makes sense.  I am wondering how much difference the lens-to-sensor distance makes in DOF.  In Panasonic's other 4/3 camera (L10, I believe), the distance from the lens to the sensor is 40mm while the G1 is 20mm.  

[attachment=12874:MicroFou...Diag_001.jpg]


I am looking forward to getting out and experimenting.

Thanks.

Dan Kehlenbach

[BJL]

It's also easy to remember this way: While a 14mm lens has the same apparent field of view on m4/3 as a 28mm lens would on 35mm, it's still a 14mm lens, and will have the same dof as a 14 mm lens on 35mm.

That is very misleading, because the image used from the 14mm lens in 4/3 is only half the width and height of that from the 14mm lens in 35mm format, so when printed at the same size, it must be enlarged twice as much. That extra enlargement doubles out of focus effects too, and so gives half as much DOF as 14mm in the larger format, and also a very different FOV, if comparing with equal f-stop.

But on the other hand, doubling focal length from 14mm to 28mm at equal f-stop reduces DOF by a factor of four (it goes with the square of focal length.) And doubling f-stop doubles DOF (with same focal length same degree of enlargement.)

So, doing the comparison sanely, viewing equal sized images of the same subject from the same distance:
- 14mm in 4/3 has half the DOF of 14mm in 35mm at the same f-stop
- 28mm in 35mm format has one quarter the DOF of 14mm in 35mm format at the same f-stop
- 28mm in 35mm format has twice the DOF of 28mm in 35mm format at half the f-stop
and thus
- 14mm in 4/3 has twice the DOF of 28mm in 35mm at the same f-stop
- 14mm in 4/3 has the same DOF of 28mm in 35mm at twice the f-stop. E.g 14mm f/4 in 4/3 vs 28mm f/8 in 35mm.
The last is my rule of thumb, doubling focal length and f-stop for equal FOV and equal DOF (and equal diffraction effects.)


Hopefully we will not now return to debating whether it is more reasonable to compare images with completely different field of view from the different formats, or to compare crops to equally sized portions of the differently sized sensors!

[BJL]

I am wondering how much difference the lens-to-sensor distance makes in DOF.  In Panasonic's other 4/3 camera (L10, I believe), the distance from the lens to the sensor is 40mm while the G1 is 20mm.

That is only distance from the lens mount to the sensor; it has no effect of optical properties like DOF, which will be the same for equal focal length, equal aperture, equal subject distance in any 4/3" format camera.

[aaykay]

I am wondering how much difference the lens-to-sensor distance makes in DOF.

From a DOF perspective, it has absolutely zero effect.  But what that will do, would be to alter the design of the lens itself (especially true for the really challenging ultra-wides and wides), since the new design will not have to compensate for that extra distance to accomodate the mirror.  Also, the design itself becomes a lot more simpler,  with less propensity for coma and other distortions.

[Ray]

Here is my rule of thumb (but beware; many forum wars have been fought on this topic!)

So about 2/3 stop between Four Thirds and either EF-S or DX.

BJL,
The G1 sensor size is 18mmx13.5mm. The Canon EF-S sensor is 22.3mmx14.9mm. The DoF difference between the 4/3rds format and specifically the EF-S cropped format, which is slightly smaller than the DX format, should therefore range between 22.3/18=1.2x and 14.9/13.5=1.1x, depending on the previsualised aspect ratio of the composition.

Assuming one maximises the sensor area in relation to the composition (using a zoom) and assuming one crops only to change the aspect ratio, then the DoF advantage of the G1 in relation to the Canon EF-S cropped format will be, at most, slightly more than 1/3rd of a stop and, at least, slightly less than 1/3rd of a stop.

The best approximation would therefore be 1/3rd of a stop difference, in relation to the Canon APS-C format. Nothing to shout about really. Agreed?  

[BJL]

The G1 sensor size is 18mmx13.5mm. The Canon EF-S sensor is 22.3mmx14.9mm.

No, the 12.1MP image from the G1 sensor comes from a 17.3x13mm "active area": the nominal 18x13.5mm of 4/3" format maybe covers the total pixel count of 13.1MP. So the sensor height and width ratios EF-S/G1 are about and 1.15 and 1.3, with diagonal ratio of 1.24 (26.8mm/21.6mm). That 1.24 is very close to the 1.26 linear factor for 2/3 stop, but if you prefer the extremes, 1.15 is 2/5 stop and 1.3 is 3/4 stop. (A factor if 1.4 is one stop.)

(I used 22.5x15mm for EF-S, but recent EF-S have the slightly smaller dimensions that you gave.)

[Ray]

No, the 12.1MP image from the G1 sensor comes from a 17.3x13mm "active area": the nominal 18x13.5mm of 4/3" format maybe covers the total pixel count of 13.1MP. So the sensor height and width ratios EF-S/G1 are about and 1.15 and 1.3, with diagonal ratio of 1.24 (26.8mm/21.6mm). That 1.24 is very close to the 1.26 linear factor for 2/3 stop, but if you prefer the extremes, 1.15 is 2/5 stop and 1.3 is 3/4 stop. (A factor if 1.4 is one stop.)

(I used 22.5x15mm for EF-S, but recent EF-S have the slightly smaller dimensions that you gave.)

That's very misleading of Dpreview to quote the total sensor area for the Panasonic G1 and GH1 yet on the same style and format of introductory specification sheet quote the smaller 'active' sensor area of the Canon APS-C formats, without mentioning the distinction.

[BJL]

That's very misleading of Dpreview to quote the total sensor area for the Panasonic G1 and GH1.

Yes; or just plain confused: DPReview often uses the traditional (total?) 18x13.5 dimensions, even in its "pixel density" calculations, ignoring the dimensions given in spec. sheets for many Four Thirds cameras, including in some of its own reviews, like for the E-30, which has basically the same sensor as the G1: http://www.dpreview.com/reviews/olympuse30/page3.asp

[Ray]

That 1.24 is very close to the 1.26 linear factor for 2/3 stop, but if you prefer the extremes, 1.15 is 2/5 stop and 1.3 is 3/4 stop. (A factor if 1.4 is one stop.)

(I used 22.5x15mm for EF-S, but recent EF-S have the slightly smaller dimensions that you gave.)

Let's not quibble. Let's call it 2/3rds of a stop advantage. Moving on to the ramifications of the greater DoF of the 4/3rds format, I think we are all searching for the sort of performance in the lighter camera that can equal the performance of the heavier model.

How would a G1 with 14-54/F2.8-3.5, or 14-42/F3.5-5.6,  compare with a 50D with EF-S 17-55/F2.8? Where are the tough comparisons?

It seems to me that a 50D at 17mm and F3.5 might produce a cleaner and sharper result than a G1 or E-620 at 14mm and F2.8. DoF should be about the same, but in low light I suspect the 50D will surpass the 4/3rds format.

When DoF can be sacrificed, or when a shallower DoF is desired, the 50D at 17mm and F2.8 will outperform any 4/3rds camera and lens combination, I suspect.

A 4/3rds format lens equivalent to the EF-S 17-55/2.8 would be a 14-44/F2.2 across the whole range. It doesn't exist.

If we compare the 4/3rds format with full frame 35mm of the same pixel count, for example, the Nikon D700 and G1 or E-620, I suspect the gap widens. With a 2 stop difference in DoF, I'm prepared to accept that a G1 or E-620 at ISO 1600 might be at least as good as a D700 at ISO 6400 regarding noise, but where's the equivalent Zuiko lens to Nikkor 14-24/2.8? It would need to be a 7-12/F1.4. The closest is 7-14/F4.

Another issue is of course resolution at the DoF equivalent F stop. Is the Zuiko 7-14 at 7mm and F4 sufficiently sharp to compete with the Nikkor 14-24 at 14mm and F8? If it is, is it also sufficiently sharp to compete with the Nikkor 14-24 at F4 and F5.6 in circumstances when a shallow DoF is okay?

If a Zuiko 7-12mm/F1.4 were a feasible proposition, woulkd it be about 50% sharper at F1.4, from corner to corner, than the Nikkor 14-24/2.8 at F2.8.? It would need to be to delvier equally sharp image resolution at the 12mp level.

I don't wish to appear dogmatic about such issues, but the great dearth of direct and competently executed comparisons on the interent give me reason to suspect there's really no contest.

[BJL]

Let's not quibble. Let's call it 2/3rds of a stop advantage.

No quibble from me: that 's the facts!

Moving on to the ramifications of the greater DoF of the 4/3rds format, I think we are all searching for the sort of performance in the lighter camera ...

How would a G1 with 14-54/F2.8-3.5, or 14-42/F3.5-5.6,  compare with a 50D with EF-S 17-55/F2.8?

If the interest is a lighter camera (or even in similarly priced kits), lens weight (and cost) matter too, so the EF-S 17-55/F2.8 is the wrong lens to be using: that choice clearly gives better low light performance from a heavier, more expensive kit, just as one expects by changing to a lens with substantially larger aperture diameter and front element size. (The 50D is probably the wrong body too: maybe the smaller, cheaper  500D?)

How about trying to stay in roughly the same size and weight range with something like
50D or 500D body with EF-S 18-55 f/3.5-5.6 IS vs G1 body with Panasonic 14-45/F3.5-5.6 (45, not 42) or E-620 with Olympus 14-42/F3.5-5.6
or
50D or 500D body with EF-S 17-85 f/4-5.6 IS vs E-620 with Olympus 14-54/F2.8-3.5 (or G1 w/ 14-54/F2.8-3.5, but then IS is lost)

The latter pairing actually shifts the DOF control advantage to the 4/3 kit, and maybe the wide open low light performance too, because that Canon 17-85/4-5.6 has uniquely small maximum aperture diameters for its class and price range.

[250swb]

It would need to be a 7-12/F1.4. The closest is 7-14/F4.

Oh yeah, you are really going to be working the DOF hard with a 7mm lens..... I can't see where its going to make much difference if its f1.4 or f4, who is going to use a 7mm lens when they need shallow DOF? I don't think the sums are reflecting the real world, in which many people who use Olympus lenses have given up worrying if they are 'sufficiently sharp' to compete, because 'sharpness', or lack of it, is why they dumped Canon or Nikon in the first place.

Steve

[Ray]

No quibble from me: that 's the facts!

BJL,
That's not quite correct. Those are the mathematical calculations based upon the simplest of formulas. In practice there are other factors, including lens design and distance to subject, which influence those theoretical DoF differences.

I've noticed myself, when comparing the noise differences between the 40D, 50D and 5D, after appropriate adjustment of ISO and aperture to equalize shutter speed, that the 1.6x factor applied to F stop # often does not result in equal DoF between the cropped format and full frame.

In other words, using the same zoom lens (Canon 24-105), adjusting FL to equalise FoV, stopping down by 1 & 1/3rd stops on the 5D whilst increasing ISO by 1 & 1/3rd EV to maintain the same shutter speed, results in the 5D image having noticeably less DoF.

Focussing on subjects at fairly close distances of 6 to 12ft (in my living room), there appears to be something like a 2 stop difference in DoF between the Canon APS-C format and FF 35mm. That is, the same difference that your formulas attribute to the 4/3rds format versus FF 35mm.

Now, I haven't taken the trouble to compare DoF with the plane of focus at numerous distances, from near to far. I've got other things to do. But it would be an interesting exrecise for someone with the time.

If the interest is a lighter camera (or even in similarly priced kits), lens weight (and cost) matter too, so the EF-S 17-55/F2.8 is the wrong lens to be using: that choice clearly gives better low light performance from a heavier, more expensive kit, just as one expects by changing to a lens with substantially larger aperture diameter and front element size. (The 50D is probably the wrong body too: maybe the smaller, cheaper  500D?)

My interest is in getting the best performance possible within the parameters of reasonable cost, reasonable weight and and reasonable convenience of use. I'm not interested in sacrificing a noticeable and significant loss of image quality for the benefit of a noticeable and significant loss of weight. But I am interested in sacrificing  a small and insignificant loss of image quality for the benefits of a substantial loss of weight.

Only when comparable test images and MTF charts are available can one make sensible choices regarding these matters. Fortunately, Photozone had tested the EF-S 17-55/2.8 and their results influenced my choice of this lens. I'm clearly not interested in saving a few grams of weight, and a few dollars, by getting a slightly lighter lens which is clearly not as sharp and useful in low light. Are you?

[Ray]

Oh yeah, you are really going to be working the DOF hard with a 7mm lens..... I can't see where its going to make much difference if its f1.4 or f4, who is going to use a 7mm lens when they need shallow DOF? I don't think the sums are reflecting the real world, in which many people who use Olympus lenses have given up worrying if they are 'sufficiently sharp' to compete, because 'sharpness', or lack of it, is why they dumped Canon or Nikon in the first place.

Steve

You missed the point. I'm not arguing that the larger format has the benefit of a shallower DoF at any given aperture (although it does) because there's an equally strong argument that the smaller format has a greater, more extensive DoF at any given aperture. Such differences can be reduced to the principle of always using the best tool for the job. If shallow DoF is often a preference, then the larger format camera may be the best choice.

The 4/3rds format was designed to 'take on' the the 35mm format by capitalising on the superior quality of Zuiko lenses. I merely point out, if you want to get wide-angle performance from the 4/3rds format on a par with the D700/14-24/F2.8 combination, you need a Zuiko 7-12/F1.4 lens which is 50% sharper at F1.4, from corner to corner, than the Nikkor lens at F2.8, and 50% sharper at F2.8 than the Nikkor at F5.6.

These conditions are not only to enable equal DoF in all circumstances (except at the smallest apertures) but equal sharpness and low noise performance in low light. In other words, a 12mp 4/3rds camera with 7mm lens at F1.4 and ISO 1600 could produce results comparable to (and perhaps even better than) a D700 with 14mm lens at F2.8 and ISO 6400. Or if that's too extreme, F1.4 at ISO 800 compared with the D700 at F2.8 and ISO 3200.

[250swb]

You missed the point.

I'm sorry Ray, but to some extent I think you have missed the point.

Not wanting to get into all your DOF calculations and combinations again because they are fundamentally sound in theory, the practical world bites back. For example, a top Olympus lens like the 7-14mm wide open at f4 is sharp, and sharp across to the corners. I don't know much about Nikon lenses, but the Canon 'L' lenses I have used are not as sharp at f2.8 (or f4 if thast is the faster aperture), and even more unsharp at the corners. So it isn't a practical starting point assuming the FF lens is everyday usable at f2.8 just because it has the facility to go one stop faster. Similarly I'd be happy to use an Olympus 12-60mm wide open at f2.8 for a landscape shot, but I wouldn't use a Canon 24-70mm at f2.8 . I'm just saying that 'like for like' is more complicated than the numbers on a page.

Further complicating the issue is the recent Amateur Photographer test that shows the G1 gives finer resolution just plonking a Sigma non-Four Thirds lens on it than a Nikon D3 gives using the same lens. This result was confirmed also using the G1 kit lens. So at the very least it implies that there is a choice for Nikon D3 user between taking advantage of shallower DOF, or wanting the image sharper and therefore needing to use an optimal aperture. So again, I'm just saying perhaps in the DOF calculations all the Nikon and Canon aperture ranges should have at least the widest aperture eliminated to get a truer 'like for like' overall performance both in DOF and overall resolution. Its about the camera settings people actually choose to use, and not those they would theoretically like to use.


Steve

[Ray]

I'm sorry Ray, but to some extent I think you have missed the point.

Not wanting to get into all your DOF calculations and combinations again because they are fundamentally sound in theory, the practical world bites back. For example, a top Olympus lens like the 7-14mm wide open at f4 is sharp, and sharp across to the corners. I don't know much about Nikon lenses, but the Canon 'L' lenses I have used are not as sharp at f2.8 (or f4 if thast is the faster aperture), and even more unsharp at the corners. So it isn't a practical starting point assuming the FF lens is everyday usable at f2.8 just because it has the facility to go one stop faster. Similarly I'd be happy to use an Olympus 12-60mm wide open at f2.8 for a landscape shot, but I wouldn't use a Canon 24-70mm at f2.8 . I'm just saying that 'like for like' is more complicated than the numbers on a page.

Further complicating the issue is the recent Amateur Photographer test that shows the G1 gives finer resolution just plonking a Sigma non-Four Thirds lens on it than a Nikon D3 gives using the same lens. This result was confirmed also using the G1 kit lens. So at the very least it implies that there is a choice for Nikon D3 user between taking advantage of shallower DOF, or wanting the image sharper and therefore needing to use an optimal aperture. So again, I'm just saying perhaps in the DOF calculations all the Nikon and Canon aperture ranges should have at least the widest aperture eliminated to get a truer 'like for like' overall performance both in DOF and overall resolution. Its about the camera settings people actually choose to use, and not those they would theoretically like to use.


Steve

I'm talking about specific Canon and Nikkor lenses, the EF-S 17-55/2.8 and the Nikkor 14-24/2.8, the best in their category. They should be compared with the best equivalent Zuiko lenses in the same category.

Where are the comparisons? If I could get an E-620 with Zuiko 7-14/F4 lens which had nearly the performance of a D700 with 14-24/2.8 lens, I'd do it. I'll sacrifice a small amount of image quality, which might be noticeably only on really large prints (say 24"x32"), for the sake of the lighter weight.