BOKEH : SLR vs DSLR . Please contribute if you know.


Status
Not open for further replies.

marcwang

New Member
Jan 9, 2003
1,112
0
0
East Singapore
www.pbase.com
Good day everyone,

I was wondering if there is a difference in "BOKEH" between a normal film SLR and a DSLR.

I started photography with my A40 powershot, and as I got more into it, I started to crave for the "Bokeh" I see in professional shots. In the A40, or any digicam for that matter, you just cant get a good "BOKEH". Everything is from the subject to the background is in focus. Not good for subject isolation. I learnt that this is because of the very very short focal length in a digicam. So I guess there is no way , other than PS, which I feel is artificial, in achieving a good BOKEH. As such, I bought a Nikon F80. I like the ability to control the depth of field. But now, considering a switch to DSLR, just to upgrade. Probably a S2 Pro or something. No need for DX flash, contrasty. I like it.

But I was wondering, do DSLRs have a shorter focal length as compared to SLRs ? Will I still be able to control the DOP as well as in an SLR ?
If a DSLR and SLR have similar aperture set, focused on exactly the same thing, ceteris paribus in other words........ , will the BOKEH be similar, as in ... as blur or as clear ? This is very important in my decision. Please contribute if you have anything to share.
 

Actually, consider this:

At any given distance, the DoF of say, a 50mm lens at f/2 on a DSLR is smaller than that same lens on a film SLR, assuming you keep subject distance constant.

Thom Hogan and George Lepp has briefly mentioned these - Thom in his D100 book and George in an Outdoor Photographer article. To quote Thom:

Page 117:

Many D100 users question whether DoF on a D100 is the same as on a 35mm body (assuming the same lens, focus point and aperture settings). After dealing with this issue for over a year on D1 models, I've come to conclude that the answer is "no".

<snip>

DoF appears to me to be a bit less than one stop different for a D100 than 35mm film, at least for the sized prints that you're likely to produce (11x14" or smaller).

<snip>
For example, if your 35mm fiolm chart says that the hyperfocal distance is 50 ft for a 50mm lens at f/2 (which it should if it uses the conservative Zeiss CoC value), simply use f/2.8 on the D100).

In other words, to get the same hyperfocal distance on a D100, you need to stop down by a stop. Which means DoF on the D100 is indeed smaller than a full frame body.

DoF calculations are based around a CoC figure. According to Zeiss, this would be 1/1730 * the film diagonal area. Therefore for film/full frame (24mm x 36mm), that would be 1/1730 * (sqrt(24^2 + 36^2)) = 1/1730 * sqrt( 1872 ) = 1/1730 * 43.26662 = 0.02501mm.

For a D100 which has a CCD size of 15.6mm by 15.6mm, the CoC works out to be 0.016401mm.

Now plugging these figures into DoF calculators, you will find that the DoF of that same lens is indeed different (smaller) on the DSLR compared to a full frame body.

Another section of the book says:

Page 169: Watch your focus : If you intend to print at sizes larger than 8x10" (~ISO A4), you should realize that depth of field on a D100 is a bit smaller than for the same focal length, focus distance and aperture combination on a 35mm body.

One thing to remember: The focal length remained unchanged. It is still 50mm. Field of View is changed. And because the image now has to be enlarged to a bigger extent than with a full frame sensor (and the different CoC values because of this), the DoF is affected as well.

Regards
CK
 

Originally posted by ckiang
One thing to remember: The focal length remained unchanged. It is still 50mm. Field of View is changed. And because the image now has to be enlarged to a bigger extent than with a full frame sensor (and the different CoC values because of this), the DoF is affected as well.

Regards
CK
:thumbsup:
You have no idea, how many people, even those on DPReview don't understand this.
 

if its only 1 stop difference, then its negligible. WHen I take pictures with a fixed lense digicam like G2 , A40 .... a F2.8 shot has a DOP of like .... f22 on a Film SLR man.
 

Originally posted by marcwang
if its only 1 stop difference, then its negligible. WHen I take pictures with a fixed lense digicam like G2 , A40 .... a F2.8 shot has a DOP of like .... f22 on a Film SLR man.

Those are a different league altogether.....


Regards
CK
 

The theory aside, what I notice about the bokeh from DSLRs is that the graduation of tones and hues is not as smooth as it should be on film. If you have seen the bokeh of say a prime lens, the bokeh of a digital will look kind of unnatural to you.
 

Originally posted by Prismatic
The theory aside, what I notice about the bokeh from DSLRs is that the graduation of tones and hues is not as smooth as it should be on film. If you have seen the bokeh of say a prime lens, the bokeh of a digital will look kind of unnatural to you.

Hmmm.... I shoot both film and digital, never really noticed that. Which lens/DSLR are you using?

Regards
CK
 

Originally posted by ckiang
Hmmm.... I shoot both film and digital, never really noticed that. Which lens/DSLR are you using?

Regards
CK

Well, no specific lens/DSLR cos never really test out the different. However, shot several pictures of the same scene using both a DSLR and SLR before using the same lens, the bokeh looks sort of different.

On film, if you have a good bokeh, the graduation of tones is quite smooth, what I have noticed for digital pictures is that the bokeh seems too define. For example if you have a boundary between hues, instead of a smooth transition of hues, what I got is a well-defined region of blur in between the shades.
 

Originally posted by Prismatic
Well, no specific lens/DSLR cos never really test out the different. However, shot several pictures of the same scene using both a DSLR and SLR before using the same lens, the bokeh looks sort of different.

On film, if you have a good bokeh, the graduation of tones is quite smooth, what I have noticed for digital pictures is that the bokeh seems too define. For example if you have a boundary between hues, instead of a smooth transition of hues, what I got is a well-defined region of blur in between the shades.

Jpeg? Try shooting in RAW and see if it is improved. :)
 

just itching to give my 2 cents worth.

1) First we talk about bokeh. Bokeh should be the same if the same lens is used for both SLR and DSLR. (FYI, bokeh is caused by a lens imperfection known as spherical aberration, which in this case is good) Can it be that when the same lens is used on a DSLR, the smaller CCD (which caused the multiplying effect on the focal length) gives a relatively different bokeh effect. Note it is different bokeh "effect", that is the perceived effect, and not the actual bokeh for the lens which shouldn't change if this lens is used on both a SLR and DSLR.

2) Next, we talk about tone and shade gradation. This is has to do with the pixel density of a given image. The higher the density, or finer the grain, the better tonal gradation, just like a medium or large format enlargment has better tones and shade gradation compared to small format for the same size enlargement. A 24mmx36mm frame has the equivalent of 14 megapixel. A d100 image is 6 megapixel. Hmm....what do we have here.....
 

In RAW, it's significantly better, but then it's still visible. However, it's really a minor problem unless you are printing in large sizes.
Though I still prefer the resolution of details on film.
 

Originally posted by Ansel
just itching to give my 2 cents worth.

1) First we talk about bokeh. Bokeh should be the same if the same lens is used for both SLR and DSLR. (FYI, bokeh is caused by a lens imperfection known as spherical aberration, which in this case is good) Can it be that when the same lens is used on a DSLR, the smaller CCD (which caused the multiplying effect on the focal length) gives a relatively different bokeh effect. Note it is different bokeh "effect", that is the perceived effect, and not the actual bokeh for the lens which shouldn't change if this lens is used on both a SLR and DSLR.

2) Next, we talk about tone and shade gradation. This is has to do with the pixel density of a given image. The higher the density, or finer the grain, the better tonal gradation, just like a medium or large format enlargment has better tones and shade gradation compared to small format for the same size enlargement. A 24mmx36mm frame has the equivalent of 14 megapixel. A d100 image is 6 megapixel. Hmm....what do we have here.....

Great for the explanation. But that's basically what the thread is about; how the bokeh appears differently for DSLR and SLR. After all, it's the end-result that's important.
 

Originally posted by Prismatic
Well, no specific lens/DSLR cos never really test out the different. However, shot several pictures of the same scene using both a DSLR and SLR before using the same lens, the bokeh looks sort of different.

On film, if you have a good bokeh, the graduation of tones is quite smooth, what I have noticed for digital pictures is that the bokeh seems too define. For example if you have a boundary between hues, instead of a smooth transition of hues, what I got is a well-defined region of blur in between the shades.

I know what he is talking about. I have noticed the same thing with my digital prints. They look fine on the screen, but when printed out, there is a definitely noticeable gradation (stepping) of colours. Rather ugly.

Not sure whether it has to do with jpeg compression or colour gamut. Will try some shots with RAW and some shots with Adobe RGB, print them out and report back later.

PS I refer to commercial digital printing, not to home printing.
 

Originally posted by Prismatic
Great for the explanation. But that's basically what the thread is about; how the bokeh appears differently for DSLR and SLR. After all, it's the end-result that's important.

OK I have a suggestion for a test. But we need someone who has a DSLR and SLR. I can't afford a DSLR so I can't help here, so need a volunteer with booth types of camera of the same brand so we can use the same lens.

1) With the anologue SLR, take a picture which shows a clear subject, an out of focus foreground, and an out of focus background. Note aperture, and focal length and distance.

2) Attach the same lens to the DSLR take a picture of the exact same subject, framing it exactly as in (1) above. You may have to either move further away, or change the focal length (if it is a zoom lens). In fact, try both: take a picture by changing the focal length to create the same frame, then set the focal length back to the same as (1) above, then move further back. For both cases, maintain aperture, note focal length and distance.

Post the two (or three) pictures and we can have a go at coming to our conclusions.
 

Originally posted by Ansel
OK I have a suggestion for a test. But we need someone who has a DSLR and SLR. I can't afford a DSLR so I can't help here, so need a volunteer with booth types of camera of the same brand so we can use the same lens.

1) With the anologue SLR, take a picture which shows a clear subject, an out of focus foreground, and an out of focus background. Note aperture, and focal length and distance.

2) Attach the same lens to the DSLR take a picture of the exact same subject, framing it exactly as in (1) above. You may have to either move further away, or change the focal length (if it is a zoom lens). In fact, try both: take a picture by changing the focal length to create the same frame, then set the focal length back to the same as (1) above, then move further back. For both cases, maintain aperture, note focal length and distance.

Post the two (or three) pictures and we can have a go at coming to our conclusions.

You can't really conduct this test without introducing too many variables:

1. Film scan introduces few more variables : Film scanner scan quality, scan resolution, sharpening.

2. Film used adds few more variables : Film grain, inherent sharpness, contrast, colour rendition

3. Post processing required on the film scan adds some more variables, including but not limited to USM amount. That same amount CANNOT be applied to the DSLR image.

Yes, we are talking about bokeh here, but the variables in the scanning process can effectively make the test moot.

Regards
CK
 

Originally posted by ckiang
You can't really conduct this test without introducing too many variables
Regards
CK

So CK, how do you think we can conduct an objective test or verification of the observations of marcwang and others on this thread?
 

Using the same lens, same aperture, same subject, same distance, bla bla bla... on a DSLR and an SLR, assuming that it's a full size CCD (fsccd), the bokeh swill be the same.

However, many Dcams in the market, and some DSLR included, does not utilise a fsccd, therefore, it is observed that the bokeh is different.

Now, note that this does not translate to a DSLR having different bokeh to an SLR, because a fsccd imitates the exact dimension of a 35mm film, therefore those DSLR equipped with it will have the same bokeh as a similar photo taken with an SLR.

The whole issue here is the film size, or rather, the issue with the Digital is the CCD.

That's basically the core of the whole chunk of text that Thom and George were talking about. It aint that hard to understand after u carefully analyse the equations.
 

Originally posted by elsanto
Using the same lens, same aperture, same subject, same distance, bla bla bla... on a DSLR and an SLR, assuming that it's a full size CCD (fsccd), the bokeh swill be the same.

However, many Dcams in the market, and some DSLR included, does not utilise a fsccd, therefore, it is observed that the bokeh is different.

Now, note that this does not translate to a DSLR having different bokeh to an SLR, because a fsccd imitates the exact dimension of a 35mm film, therefore those DSLR equipped with it will have the same bokeh as a similar photo taken with an SLR.

The whole issue here is the film size, or rather, the issue with the Digital is the CCD.

That's basically the core of the whole chunk of text that Thom and George were talking about. It aint that hard to understand after u carefully analyse the equations.

Excellent summary of my sentiments right from the start. I know that the bokeh on a 24x36mm film frame and a 24x36mm CCD sensor with the same lens, same distance, same focal length same aperture, is the same. Logically and theoretically yes.

But is there a way we can produce an empirical proof of it, given that we are talking about two different media here.
 

Until I can afford the N Digital body (SGD $14k) for my lenses, nope, I'm afraid I can't proof that.

But again, that'll be a 6mb body, not a 14mb... Yes, under an 8r or 12r print the difference (theoretically) is noticeable, though not significant... Which adds onto the "experiment" a whole new factor to be considered.

On 2nd thoughts, I do not need a digital body... good luck with your quest.

;p
 

Originally posted by Ansel
So CK, how do you think we can conduct an objective test or verification of the observations of marcwang and others on this thread?

Can't think of a very objective way. Comparing film and digital always introduces other variables. So no point. :)

Regards
CK
 

Status
Not open for further replies.