angle of view / CCD Size

S

serra

Member
hi all,

for reasons not worth expanding on here, i need to compile
a table that correlates angle of view with focal length. now
theoretically, this is rather simple to do (theres some
basic geometry involved in the computation). but, for some
confounding reason, i am consistently off by roughly a factor
of 2 when comparing my theoretical values to what my camera
actually sees at a given focal length. as the math can't be
the source of the problem, i come to question the CCD
dimensions that i'm using for my calculations. i'm operating
in 16:9 letterbox mode and for that setup, i use a horizontal
CCD dimension of 0.33 in. and a vertical CCD dimension of
0.186 in. can anybody make a statement as to whether
these dimensions are on the money?

cheers and thanks,

serra
 
Re: angle of view / CCD Size

Have you looked at:

http://www.cinematography.net/CINEFILE.HTM

There are a few resources...

pCAM DoF, Field of View etc calculator

Film.zip Max Fords Windows film program, DoF, Lens Angles, Time Lapse, Lighting etc (get it! GB)

Lens equivalents for different formats (XLS)
 
Re: angle of view / CCD Size


I have a cannon focus chart that lists the Angle of view to Focal Length relationship of a 1/3" Video @ 4:3. But that really doesn't help me when I dont' have the exact Z numbers that correspond to the focal lengths. However, I could post some of those numbers if you want.

Also, as for 16:9, correct me if I'm wrong, but if you are using the in-camera bars, then do you even need to calculate the widescreen size? Isn't the image EXACTLY the same as the 4:3 image, other than it having the black bars?

Finally, I'm pretty much clueless and lost on this as well. Can you maybe post up the math you are using to calculate your answers??? :)
 
Re: angle of view / CCD Size

could you post the FOV information for the XLS please?
maybe we can get that data to be of some use to us,
if we make use of the DVX focal length chart i compiled and
which you can find here:

http://www.redshift.ch/dvx/dvx100_focal_length_chart.xls

i came up with this chart by way of a curve fit across the
known (i.e., read from the zoom ring) values of Z vs. f... the
number crunching came up with an absolutely excellent fit and
i have some confidence in the (reasonable) accuracy of the numbers,
but... i'm not making any guarantees in that regard... use at
own risk :)

as for your question regarding FOV in 4:3 vs. 16:9 mode... yes,
the FOV in the horizontal direction would be the same between
the two modes, however, the FOV in the vertical direction would
not, as the 16:9 letterbox mask effectively reduces what is seen
in the vertical dimension, which is another way of saying that
vertical FOV has been decreased...

finally, if the CCD size on a given axis (vertical/horizontal) is "c",
and if "f" is the focal length employed, then the angle of view "a"
can be found by use of this expression:

a = 2 * arctan (c / 2f)

but, as i stated in my original post, marriage of above formula
with my values for focal length vs. Z leads to some
inconsistencies... interestingly these hold up even when testing
against the known values of Z vs. f... which leads me to suspect
that i'm making some wrong assumptions regarding CCD
dimensions.... the subject of my original post... there you have
it... ;)


cheers,

serra
 
Re: angle of view / CCD Size

According to the Panavision calculator (http://www.panavision.co.nz/main/kbase/reference/tblelenseqvform.asp) and assuming your Z to 1/3" relation is correct, the chart you published is wrong. I'm not a valuable technical source myself, but here are the results I got on this link (4:3 ratio) :

Z   1/3" 35mm

0 4.5 18.9
5 4.8 20.2
10 5.2 21.8
15 5.8 24.4
20 6.4 26.9
25 7.0 29.5
30 7.6 32.0
35 8.2 34.5
40 8.8 37.0
45 9.5 40.0
50 10.3   43.4
55 11.3   47.6
60 12.5   52.6
65 14.1   59.3
70 16.1   67.8
75 18.7   78.7
80 22.1   93.4
85 26.4   111.0
90 31.8   134.0
95 38.5   162.0
99 45.0   189.00

Where is the truth ? Can anyone more reliable than me confirm or correct that chart ?

Anyway, thanks a lot Serra for providing us the Z to 1/3" focal length relation, which is the most useful tool we needed.
 
Re: angle of view / CCD Size

I'm sorry to say, Anartiste, that I'm with serra on his Z to Focal Length chart.

I'm not sure where you got those numbers, but if you physically look at the numbers you can emperically deduce that the Z to Focal Length numbers is NOT a linear relationship. We know that 4.5 = Z00 and 45 = Z99, but if you dial the zoom ring, you'll notice that 6 ~= Z15-20, 10 ~= Z43-52, and 20 ~= Z73-77. Those numbers are definately not linear. However, I'm also not sure how you got those numbers serra, but to my eye, 6=Z17, 10=Z48 and 20=Z75. But that's what I saw on my camera.

I'm sure I could map those damn numbers and get a decent Focal Length numbers, but I thought people already came up with answers for this kind of problem. Doesn't eveyone want to know what the Hyperfocal distance is? Or is everyone trying to eliminate DOF.

Anyhow, the Angle of view vs Focal Length chart I got from Cannon has nothing to do with XLS. I was at a convention, and Canon gave away focus charts with the aforementioned chart on its back. I'd believe those numbers than anything else I derive. I'll post the DVX range listed on that chart:

Angle of View Focal Length 1/3" Video (4:3)
(Horizontal) (CCD witdth of 4.8mm)
56.1 4.5
51.2 5.0
43.6 6.0
37.9 7.0
35.5 7.5
29.8 9.0
27.0 10.0
22.6 12.0
21.7 12.5
18.2 15.0
15.6 17.5
13.7 20.0
12.2 22.5
11.0 25.0
10.0 27.5
9.1 30.0
8.4 32.5
7.8 35.0
7.3 37.5
6.9 40.0
6.5 42.5
6.1 45.0

But honestly, I don't see people using Angle of Views all that much. It's the Focal Length that people reference for all the DOF stuff. Those numbers are useless to me unless I get a good Z to Focal length chart.

Someone HELP!
 
Re: angle of view / CCD Size

ok guys... confusion seems to reign...

let me explain some things about how the Z to f chart was derived:
as we've all noticed the Z to f relationship is definitely not linear, hence a functional relationship between Z and f (i.e. f(Z) ) is not readily determined without some serious number crunching in the sense of subjecting the known values of f(Z), i.e.:

f[0] = 4.5
f[17] = 6
f[48] = 10
f[20] = 76
f[99] = 45

to a curve fitting algorithm, in the hope that that will yield a function that describes the observed data points with sufficient accuracy. such a (polynomial) function was found and the result of that curve fit is what lies at the foundation of the XLS chart that i published. so these numbers were not pulled out of
thin air... rather they were derived based on standard mathematical procedure...

now that the "how" ought to be better understood, the question of accuracy arises. curve fits yield a "correlation factor", for the data in question. the better the fit, the closer the factor is to some ideal value. for the f(Z) curve fit, this value was close to something suggesting a "perfect fit". given that, a possible uncertainty, or fuzziness, arises due to the somewhat imprecise nature with which the known datapoints (listed above) are determined... basically one dials in the f-markings listed on the zoom rings and then reads off the corresponding Z value. this can be done only with so much precision, hence the difference between what SJ_Lee reported for f(75) to what my table shows for f(75)... this "error" is negligible however, my table yields 20mm at f(77), which to me is close enough for it not to warrant any worry...

the other values SJ_Lee reports correspond 100% to my table... simply plug in the Z values you need an f for and you'll
see that we're on the same page here... i simply chose 5 Z increments for brevity's sake... if you want you can easily
adapt the table to show you f(Z) in 1 Z increments...

whew! all of that means that i'm pretty darn certain that my
table holds up to scrutiny... as far as f(Z) for DVX focal lengths is concerned...

the f(Z) values for 35mm are another matter... basically i proceeded roughly the same way for the 35mm values as i did for the DVX f-values... panasonic publishes 35mm equivalents (endpoints) in the tech addendum... those were the departure points for what i did (i'll spare you the details)... the question in light of what the panavision site's calculator yielded is: just what 35mm system is panasonic referring to? 35mm (4:3) film cameras, or 35mm SLRs...??? i'm really not sure what the answer is to that one... but i suspect it's 35mm SLR they're referring to... which would explain why my values don't match what Anartiste obtained....

all in all... i stand behind f(Z) for DVX focal lengths... f(Z) for 35mm focal lengths need to be taken with a grain of salt... ;)

sorry for the biblical length post...

serra
 
Re: angle of view / CCD Size

me again... sorry... ;D

a minor correction:

Anartiste, what i published is a functional relationship between DVX focal lengths and Z-values... CCD size does not factor into that in any way...

this in reference to your statement:
[ Anyway, thanks a lot Serra for providing us the Z to 1/3" focal length relation, which is the most useful tool we needed. ]

just needed to clarify that,

serra
 
Re: angle of view / CCD Size

As I've said before, I think serra is right on this. The only difference is that the errors that were attributed to the emperical analysis of the Z numbers and Focal numbers around the dial.

What remains to be seen is the EXACT Z number to Focal Length relationship for the DVX100. Obviously, such a list exists, lest it be buried in the vaults of Panasonic engineers. What I'd like to know is that why don't they give us those numbers???. I'd like to know, partially because I'm curious, and mostly because I think it'd be useful for finding the exact DOF for this camera.

I do have a question for serra though? Why are concerned about the 35mm Focal Lengths? Even more strangely, angle of view for that matter? Unless you want to match the DOF for the 35mm, I don't see the need for this.

The focal length chart I've mentioned above gives a list of:

16mm Standard
16mm (Super-16)
35mm
35mm Full Aperture
35mm Still

The chart is a Film Lens chart so I presume it's accurate, unless Canon has incompetant people writing up those tables
The 35mm has Focal Length almost 5 times that of the 35mm.
 
Re: angle of view / CCD Size

[quote author=serra link=board=Events;num=1081605832;start=0#7 date=04/11/04 at 12:29:14]Anartiste, what i published is a functional relationship between DVX focal lengths and Z-values... CCD size does not factor into that in any way...[/quote]

The DVX having a 1/3" 3CCD sensor, I was assuming that what was true for the DVX was true for any camcorder with the same sensor size — i.e. Canon XL1, Sony PD150, etc. Am I wrong here ?

So, in order not to scramble everything more for today, can we assume the Z to DVX100 focal length relation is correct ?

If so, all we have to do now is to find a 1/3"... sorry : a DVX to 35mm (cinema) equivalence chart and we'll be quite happy.

[quote author=SJ_Lee link=board=Events;num=1081605832;start=0#8 date=04/11/04 at 15:22:01]Why are concerned about the 35mm Focal Lengths?  Even more strangely, angle of view for that matter?  Unless you want to match the DOF for the 35mm, I don't see the need for this.[/quote]

I think it's useful if we work with a DP and/or a cameraman familiar with 35mm cameras, so that they can take their marks. Also useful for us to get used to feel the equivalence between our baby's optics and 35mm optics.

It's not replacing taste, nor will it prevent us from bad framing or composition, but I find it a funny sport.
;)
 
Re: angle of view / CCD Size


Assuming is alway deadly when dealing with numbers, but unless someone can prove otherwise, serra's numbers are the right ones, or very close to them.

As for the Focal Lengths, I think they are important. I was a bit puzzled with the need to know the Angle of View.

Since the subject is in full swing, can I ask what the Circle of Confusion for the DVX100 is? I've seen writings on the web that says that the 16mm CoC is anywhere between 0.015 to 0.025. Which is it? More importantly, what is the CoC for the DVX100?
 
Re: angle of view / CCD Size

Circle of Confusion size is rather subjective, and depends on how you're going to be viewing the footage.  For purposes of developing DOF charts for the ASC, the circle of confusion size was set based on the idea that the footage would be blown up to a theatrical 35mm print and projected on a large theater screen.

The ACM recommends .025mm for 35mm film, and .015mm for 16mm film that's to be blown up to theatrical size.

If you're not projecting on a large screen, your Circle of Confusion can be much larger and still yield acceptable focus.  It is the magnification to theatrical size that demands tight circle sizes.  The Circle of Confusion for 16mm (intended for blow-up) is typically much tighter than it is for 35mm, because 35mm is not magnified as much... 16mm gets optically magnified to about 5x as much as 35mm does, to fill a theater screen.  The end goal is to end up with circles the same size after blowup.*

Much (most?  All?) DVX footage will never undergo a theatrical blowup, so the circle can be much, much larger when taking into account television screen sizes.  The smaller the end screen, the larger the acceptable circle size.

When going for a theatrical blowup, the circle size you use in those circumstances would need to be incredibly small, maybe .002mm, probably smaller still.

So, that's a lot of rambling to basically say: it's up to you as to when you think something crosses the line from in-focus to out-of-focus, and the larger your final screen size, the tighter the circle will need to be.

*
(interestingly, or diversionarily, S16 undergoes only 3x as much magnification, as opposed to 16mm which undergoes a 5x magnification.  So the circle on S16 can be quite a bit looser.  Using R16 circle size to determine DOF will yield much deeper DOF on S16 footage when blown up to 35mm.)
 
Re: angle of view / CCD Size

IWhat's this Circle of Confusion thing ? Right now, I'm standing at the center of it... :-/ ??? ??? :-/
 
Re: angle of view / CCD Size


I can't remember the technical definition for it but Circle of confusion is a circle of representing a point on the film/CCD an image point formed by the lens. This determines the the DOF.

As for it being a subjective size, yes I do remember reading that this is a subjective size, but it's obviously a tangiable number that will eventually translate into actual focal/hyperfocal distances when used in formulas. So, those numbers actually do matter as even a small differentiation, e.g. 0.0005mm, has a big effect.

I guess the difference, as you were saying, is that film/video numbers vs. photographcs as may be what differentiates the CoC.
 
Re: angle of view / CCD Size

The Circle of Confusion is a constant used in formulas to calculate effective depth of field. What it means is the diameter at which a pinpoint is no longer considered "in focus" and crosses the line to being "out of focus".

A properly focused lens will render a pinpoint of light as a tiny super-sharp -- well, pinpoint. The more out of focus it is, the larger and softer that pinpoint becomes. The Circle of Confusion diameter is the point at which a person looks at the pinpoint and declares "that is now out of focus."

Lenses can only ever truly focus on one infinitely-thin plane at a time. The area in front of, and behind, that plane are progressively more and more out of focus, the farther they are from that plane. To start with, they're acceptably sharp, but there comes a point where they cross the line from in-focus to out-of-focus. That area, measured from in front of the lens to behind the lens, where things are still considered "in focus", is called "depth of field".

The question becomes, what defines "in focus"? You might have a stricter standard of tolerance than I do, for example. So circle of confusion diameter is somewhat subjective. Furthermore, depending on how large you're magnifying the image will determine what's considered in focus. Something may look perfectly good on a small television screen, but when projected in a theater it'd look horribly out of focus (witness the entire movie "28 Days Later").
 
Re: angle of view / CCD Size

Thanks a lot Barry for clearing up this concept new to me.

Could we please folks go back to our first concern : Z to 1/3" to 35mm (film) ? I think this is something that could be useful to a lot of us DVXusers. So...

1/ Can anyone confirm or correct the Z to DVX focal length Serra gave us ?

2/ Can anyone tell us for sure if the DVX has a 35mm focal range equivalent of 19mm—>190mm or 32.5mm—>325mm ?

3/ Do you professional or confirmed cinematographers think it's important to know such things ? (I sometimes get obsessed by such useless issues...)
 
Re: angle of view / CCD Size

1 - no, because I'm too lazy, but it sure looks good.

2 - both. 19-190 when compared to a 35mm movie camera, 32.5-325 when compared to a 35mm SLR still-photo camera. Movie cameras use about half the frame size of a still camera, so they're not the same.

3 - Not really. I mean, it's nice to know the basics of ballpark comparisons (as in, max tele = about 325mm on an SLR) but when you get down to working with it, in day-to-day stuff I never bother with what the exact focal length is in relation to my Z-setting, etc. If you were doing a project for blowup to 35mm, and you were hyper-sensitive to the DOF issue, you may want to know, but I suspect that with most indie feature work you'll be far too busy, shooting 50+ setups per day, to worry about something like whether you're at 21mm or 22mm on the lens.
 
Re: angle of view / CCD Size

A lot of us being wannabe filmmakers/cinematographers, my point was more to use this chart in order to get a "physical" knowledge of what a 28mm, a 75mm or a 150mm look like in the cinema world, so that we can have them in mind if we one day get lucky enough to shoot on 35mm (film size here, not focal length).

So, as useless as my approach may be, I'm a happy man now. Thanks Barry, Serra and the rest of you :)
 
Re: angle of view / CCD Size

To bring my share to the pot, I rewrote Serra's chart  — thanks again Serra :) —, replacing his original 35mm SLR focal lengths with 35mm cinema focal lengths and removing the decimal part which is not very useful in 35mm. For whatever use it can be to anyone, I also added angles of views (width, height 4:3 and height 16:9).

This is not a real scientific chart, mostly an approximation. Hope I'm not too far from reality...

Z      DVX f   35mm f   WIDTH   HEIGHT (4:3)   HEIGHT (16:9)
0     4.5         19        55.8°         43.3°             34.5°
5     4.8         20        53.4°         41.3°             32.8°
10   5.2         22        49.1°         37.9°             30.0°
15   5.8         24        45.5°         34.9°             27.6°
20   6.4         27        40.9°         31.2°             24.6°
25   7.0         29        38.3°         29.2°             23°
30   7.6         32        34.9°         26.5°             20.9°
35   8.2         34        33.1°         25.0°             19.7°
40   8.8         37        30.4°         23.0°             18.1°
45   9.5         40        28.2°         21.4°             16.8°
50   10.3       43        26.3°         19.9°             15.6°
55   11.3       48        23.7°         17.9°             14.0°
60   12.5       53        21.5°         16.2°             12.7°
65   14.1       59        19.3°         14.6°             11.4°
70   16.1       68        16.8°         12.7°              9.9°
75   18.7       79        14.5°         10.9°              8.5°
80   22.1       93        12.3°          9.3°               7.3°
85   26.4       111      10.4°          7.8°               6.1°
90   31.8       134        8.6°          6.4°               5.0°
95   38.5       162        7.1°          5.3°               4.2°
99   45.0       189        6.1°          4.6°               3.6°
 
Re: angle of view / CCD Size

Hrmmm... I'm not quite finished with this post :|

Can someone explain why the 35mm still has a different focal length than film?

Dosn't 35mm determine the size of the negative? How can this be different?
 
You must log in or register to reply here.
Back
Top