Has anyone come across information about the actual dimensions (mm) of the EX1's imaging area? I'd like to calculate DOF and hyperfocal distances before the camera gets here. To do that I need to know the physical dimensions of the image at the chips.
Thanks.
Clark

I don't know the sensor size but i think you can work out some things based on the FOV multiplication factor because you know the widest lens and what it's 35mm equivalent is from specs (5.8mm is 31.4mm on 35)

So for example (please shoot me down if im wrong)

the crop factor for the EX is 5.41, the HVX is 7.73 and my EOS 300 is 1.6.

So if the EX was f1.9, that would be f10 on 35mm and f6.4 on the eos and the HVX couldn't open that wide. If the HVX was on f2 that would be around f2.9 on the EX, about a stop which is what others have been saying.

I'm pretty sure aperture is independent to focal length but obviously the magnification of the DOF blur would increase with longer focal lengths (which makes it look like you get more DOF with a long lens)

Can someone back me up here :)

paul

the diagonal size of a CMOS sensor on a 1/2" optical system should be 8mm based upon the formula diag/16=lens format. 8/16=1/2.

that's how i read it. now you have to do the rest of the math to get your HxW.

Vince

"the diagonal size of a CMOS sensor on a 1/2" optical system should be 8mm based upon the formula diag/16=lens format. 8/16=1/2."

I don't know the source of this formula. I tried it and came up with a sensor size of 7mm wide by 4 mm high. The diagonal of 8mm is woefully short of 1/2". I don't think this works.

I tried the crop factor method suggested by Paul. This is a little iffey because I don't know how Sony compared the EX1 to 35mm. The proportions don't match. I took a try comparing the diagonal measures of the two formats. I ended up with a sensor size of 9.1mm by 5.1mm and a diagonal of 10.4mm. This is closer to 1/2", but it's still just a shot in the dark. Lacking confirmed actual measurements I guess I could do my DOF and hyperfocal calculations based on this sensor size. It's less precise than I'd like to be, but nothing about DOF is precise.

Thanks.
Clark

A chance to dust off my American Cinematographer Manual that I never found a use for in school!
According to the manual, a 35mm Academy aperture is 0.868 inches by 0.631 inches. Of course, that's for a 4:3 aspect ratio.
(I knew there was a reason why I was keeping that old book around!)

But these 35mm equivalent focal lengths might be for full-frame still cameras, in which case I measure the aperture as 36mm by 24mm, or about an inch and a half by an inch.

In any case, I'll leave it to someone else to calculate the DOF tables. :)
Ken Hull

i believe that 1/2" refers to the lens not the imager. they call it a 1/2 type exmor cmos. type referring to the optical format which is based on the lens family.

also, if you take the 5.41 crop factor from paul and divide that into 43.2 which is the diagonal measurement of a 35mm still frame it comes out to just about 8.

hey, it seems really small to me too.

vince

here it is:

The actual image size (active area) = 6.4mm x 4.8mm with the image diagonal = 8mm. The 1/2 inch (1/2” = 12.70mm) value that is used to describe the common format name stems from the days of Videcon Tubes and was used to describe the tube diameter and not the actual image size.

From what I've read, a 1/2" sensor has essentially the same DOF characteristics of 8mm film. So the actual dimensions of an 8mm negative should be pretty darn close to the EX-1's sensor size.

Assuming that the sensor diagonal is 1/2 inch (12.7 mm) and the aspect ratio is 16:9, Pythagoras gives the dimensions as 6.23mm x 11.07 mm.

Nick

And that is much closer to Super 16 mm film (7.4 x 12.4 mm) than to Super 8 (4.2 x 5.7 mm).

I don't think it is as simple as Pythagorus could make this, although I wish that it was. I read somewhere on these boards(I think it was Barry Green), that the sensor size is actually quite a bit smaller than the actual chip. They were discussing how the dimension is actually quite a bit decieving to say 1/2" or 1/3" because the original dimension had something to do with a tube size in old cameras. I can't remember the details other than the summation that the chip size may be 1/2 inch, but the actual sensor on the chip is much smaller than that. I definitely could be wrong.

http://de.wikipedia.org/wiki/16_mm

Der 16 mm-Film verfügt über eine Bildfeldgröße von 10,3 x 7,5 mm (Breite x Höhe) und kann sowohl einseitig als auch zweiseitig perforiert genutzt werden, die Bilddiagonale beträgt also 1/2 inch (Zoll) entsprechend 12,7 mm.

but this is seen as letterbox. Because of 16:9 the 4:3 equivelent would be a bit bigger. On the DOF thread somebody has posted a nice image with relatively shallow DOF on a normal focal length (in-room), so the EX-1 should do well for storytelling.

Ah good lord you guys are making my head hurt. Just put the damn thing in auto focus, or just eyeball it. That should be good enough.
Sarcasm aside, reading these message boards constantly keeps me amazed at how much I don't know about my craft... and never knew that I didn't know.

reading these message boards constantly keeps me amazed at how much I don't know about my craft... and never knew that I didn't know.

Hey, we got snowed in today. I just googled till I found the size. And realized when I did find it, I don't even care. But it killed some time.

vince

And that is much closer to Super 16 mm film (7.4 x 12.4 mm) than to Super 8 (4.2 x 5.7 mm).

Video camera sensor sizes given in specs are left overs from the days of tube cameras. The actual image area is smaller than the full 1/2" (or 2/3").

Here's an article on the subject:

http://www.filmalley.com/articles/sensor%20size/

the crop factor for the EX is 5.41, the HVX is 7.73 and my EOS 300 is 1.6.
..
Can someone back me up here :)

paul

There is a difference: 35mm motion film exposes the frames transverse to the film strip direction (Academy-Standard) and a 35mm still camera vertical. So the still 35mm pictures have ~ 2.5 x the size of the frames of a 35mm motion camera and therefore less DOF.
So the real difference is approx 1 - 2 stops smaller I guess.

And here's the CMOS version from silicon imaging. it's near the bottom

http://www.siliconimaging.com/ARTICLES/CMOS%20PRIMER.htm#cmosimgerchar

vince
or just go here and look at the sizes

http://www.dvinfo.net/conf/attachment.php?attachmentid=5323&d=1196679037

perhaps this is of some help:

http://www.dpreview.com/news/0210/02100402sensorsizes.asp
http://www.panavision.co.nz/main/kbase/reference/tbleframelist.asp

unfortunately, these are not always consistent between cameras... notice that at 2/3" on the panavision site there are several listings for different dimensions. there is no listing for 1/2" 16:9 so its hard to say if the ex1's 16:9 adopts the 4:3 1/2" specification's height or width or diagonal.

according to that chart, 16m film actually has a diagonal of about 11mm, 1/2" cmos about 8mm, and 8mm film about 6mm

so 1/2" is somewhere in between standard 8 and 16mm film

who knows, maybe the vignetting issue some people have is because sony went with a 16:9 1/2" sensor that is 8.5x4.8mm, but the lens is intended for 4:3 1/2" sensors and they were too generous with how they came about 16:9 from the normal 1/2" sensor format. totally pulling that out of thin air but that would explain some things.

yes. this might be the explanation - its set too close. If you resize the sensor vertical from 16:9 to 4:3 you get a higher diagonal value.
Tis also explains the good DOF. calculated in 4:3 the sensor may be just a bit smaller then 2/3.

could somebody please cut off the glass to have a look ? ;-)

No HD camera is 4 x 3, they're all widescreen format.

right, although its not unheard of for a camera to have a 4:3 sensor and crop to 16:9. the canon hv20 does this with its 1/2.7" sensor. i was just suggesting the ex1 might do the opposite by expanding horizontally instead of cropping vertically. there does not seem to be that much information on sensors that are 1/2" and 16:9. perhaps we might learn more if we look at sony's other 1/2" xdcams. there are plenty of machine vision 1/2" cameras that are HD-like resolution but most of them are 4:3.

this not an issue of HD or not but a simple question of sensor dimensions.
You can also use a 4:3 sensor and crop it to get 16:9

popular example is the XL2. it has 1/3 4:3 sensors which are cropped for 16:9. the result is a smaller real size and really bad DOF.

I've never seen the actual spec, but my best guess is that the active area of the imager is 7.2mm wide x 4.05mm high.

The first reason that I think this is the measurement is because 2/3" HD cameras are widely known to be 9.6mm x 5.4mm.

2/3" times .75 = 1/2" and so multiplying the 2/3" chip dimensions by .75 gives me the above figures (of course I know that chip sizes are based on old vidicon tube dimensions and hence are not so neatly proportional as this).

The second reason is from measuring the field of view and distance with a tape measure.

At full wide, 5.8mm, I measured the distance to a wall, and the distance between two tape marks on the wall placed at the left and right edges of frame. This gave me enough information to calculate the chip size, assuming two things - that the lens focal length is really 5.8mm and that the EX1's monitor shows the whole frame with no cropping (in fact it does not, but it comes close). From these measurements I estimated 6.54mm width.

I did the same thing at full tele, 81.2mm. The number I got was 7.75mm, with the same two assumptions.

The dimensions 7.2 x 4.05 fall neatly between these two estimates.

It's clear to me that the actual lens focal lengths on the EX1 are a bit different from the lens markings. I don't believe the lens goes quite as wide or as long as its markings indicate, but that's for another thread, and I can certainly measure that definitively once I find out the size the 1/2" HD chip.

If someone finds the actual dimensions of the active area of the EX1 chips, please let me know.

I read somewhere on these boards (I think it was Barry Green), that the sensor size is actually quite a bit smaller than the actual chip. You've read it on these boards from both Barry Green (at least once) as well as myself (several times).

Barry touches on it here:

http://www.dvinfo.net/conf/showpost.php?p=748366&postcount=21

Two weeks earlier, I had mentioned it here:

http://www.dvinfo.net/conf/showpost.php?p=740131&postcount=20

Plus, both of those occasions were in this particular forum.

...the result is a smaller real size and really bad DOF.There is no such thing as "bad" DOF (depth of field). There is only shallow DOF or deep DOF or points in between, but you cannot possibly characterize deep DOF as being a "bad" or undesirable quality. There will be circumstances in which a person may want shallow focus and there will be other circumstances in which a person may want deep focus... one is not better or worse than the other; it's simply an aesthetic choice.

Keep in mind that one of the classics of American cinema, Orson Welles' 1941 masterpiece "Citizen Kane," was mostly composed of shots using very deep focus, that is, copious amounts of DOF, to the extent where they had to resort to composites for some scenes because the lenses they were using couldn't present the extremely broad depth of field they wanted.

Hand me a cheap consumer camcorder with a tiny 1/6th-inch chip, and I'll show you how to produce a very shallow DOF with it. See http://www.dvinfo.net/articles/optics/dofskinny.php for more info. There is no such thing as "bad" DOF.

To further that point, Chris, I would suggest that if there were in fact a "bad" depth of field, it would be shallow DoF and not the deep DoF of these small format cameras.

In 35mm film production, much effort and money is expended to get "enough stop" to hold focus for a particular scene - to hold focus between an actor's eyes and nose for example. This can mean doubling the amount of lighting equipment, pushing the film or whatever. In so many cases in the past (another example is miniature photography) I would have loved to have been able to use one of these miraculous cameras with which everything is in focus!

Over the past decade, with the rise of small format digital acquisition and post, shallow DoF has become a sought after effect and not simply a byproduct of the medium. This is is exemplified when I work with CG artists who might ask me "do you want me to add more depth of field to that?" And I say "no - there's already too much - make the the background softer"

Our definitions of DoF are opposite - to me, more depth of field means deeper depth of field. To them it means shallower DoF, because in their world, everything is in focus, and "depth of field" is a filter that's added in post.

35mm adapters are lovely, but at times I think that they are superfluous added "effect" - a shortcut to an appearance of quality at the expense of other qualities. There are plenty of ways to tell a great story while embracing the small format medium and its wonderful deep DoF.

...I work with CG artists who might ask me "do you want me to add more depth of field to that?" And I say "no - there's already too much - make the the background softer"

Our definitions of DoF are opposite - to me, more depth of field means deeper depth of field. To them it means shallower DoF, because in their world, everything is in focus, and "depth of field" is a filter that's added in post.Let me assure you then...

On *this* site, CG artists are no, "more depth of field" definitely means deeper depth of field. If I have to drive to somebody's house with my twenty-year-old copy of Citizen Kane on VHS and whack them over the head with it, "more DOF" equals a deep focal plane.

"Bad DOF" is what my Rottweiler is when I come home from the dentist.

when the DOP / director WANTS or NEEDS the DOF to bee deep it is a free artists decision. But if he wants to have a shallow DOF and can´t archive it due to technical limits of the camera, it´s simply bad.
And of course I always can get a (more) shallow DOF when using tele, but this also means a very long distance from camera to object and no chance to get shallow DOF on in-room shots.
Even the EX-1 has a bad DOF for me, but it touches the lower range of what is useful.
For me the DOF of the EX-1 looks close to 16mm.

You can also use a 4:3 sensor and crop it to get 16:9.

Any HD camera I've used that offered 4x3 cropped either end of the 16 x 9 image, rather than cropping a 4x3 top and bottom to create 16 x 9. This was for shooting DV.

That's true Brian, but the post by Carlos that you're referring to:

http://www.dvinfo.net/conf/showpost.php?p=787906&postcount=22

...makes it clear that he was speaking about the standard definition Canon XL2.

And he's right, it's certainly possible to use a 4:3 sensor and crop it to get 16:9, which Canon did with the XL2, as I've explained here: http://www.dvinfo.net/canonxl2/articles/article06.php

.... it's certainly possible to use a 4:3 sensor and crop it to get 16:9, which Canon did with the XL2, as I've explained here: http://www.dvinfo.net/canonxl2/articles/article06.php

Yes, he's correct regarding SD cameras. However, I'm not saying that you can't crop a 4 x 3 sensor to 16 x 9, just that HD cameras don't do it that way.

this may not get us any closer, but the brochure for the ex-1 does mention that at 5.8mm, the ex1 lens is equivalent to a 31.4mm lens on a 35mm camera. assuming we are talking about 35mm still cameras (and we'd probably have to be given the number ranges) if they did this based on horizontal field of view, the magnification factor for the ex1 would be 5.4x over 35mm still film. and dividing this into 36mm (width of still film) would give us a sensor width of 6.65mm. but then again who knows how they are equating the FOV between these image plane sizes when 35mm film has an aspect ratio of 3:2 and the ex is 16:9. horizontal fov would be my first guess though.

other thing is perhaps someone can check out the dof calculator built into the camera and figure out the sensor size based on the dof readout. otherwise, anyone have an ex1 they want to take apart and a micrometer caliper? j/k

http://www.users.qwest.net/~chemman/fov.JPG

According to a little spread sheet I use to calculate fov, 1/2 inch sensor should be 6.4mmX4.8mm. That will probably be inaccurate for 16X9(6.46mmX3.63mm). Using the 35mm equivalent in the brochure you end up with 6.46mm or 6.47mm (given 35mm is actually 35mm and not 25.273mm for film, if so then it is 4.67mm)

Another way would be to read off the focus ring and measure the distance from a known sized object, like a 100 foot long building. A little simple math and wallah sensor H and V



I am guessing to get the true active area of the sensor is to have a data sheet for the sensor as it is used in the EX1.

Or as stated above, a hammer and a set of calipers.

The actual image size (active area) = 6.4mm x 4.8mm with the image diagonal = 8mm. The 1/2 inch (1/2” = 12.70mm) value that is used to describe the common format name stems from the days of Videcon Tubes and was used to describe the tube diameter and not the actual image size.

That makes for a 4:3 sensor. Is the EX1's sensor 4:3?

Sorry to dredge up this old thread but there's some discussion of this topic on my blog:

ProLost: Sensor Size Cheat Sheet (http://prolost.blogspot.com/2008/08/sensor-size-cheat-sheet.html)

I'd love to add the EX1 to the cheat sheet if I can get some authoritative measurements.

-Stu

Stu, I just received my EX1 today, and doing some really quick tests for sensor dimension measurement it seems to come out as 6.56mm wide (which would mean 3.69mm tall). I'll do it again and try to be more accurate tomorrow after I've gotten sleep. This is still just based on the focal length markings on the lens and the image FOV displayed on the LCD screen, didnt get to check for overscan yet so it could be off a bit.

Panavision NZ Film,Video and Photography Frames & Formats, aspect Ratios and dimensions. (http://www.panavision.co.nz/main/kbase/reference/tbleframelist.asp) lists 1/2" 16x9 as 6.97x3.92mm, 35mm equivalent comparison from the manual implies 6.65x3.74mm, and this quick test says 6.56x3.69mm. Who knows if Sony stuck with the same dimension standards panavision was referencing. I'd say if you made a box representing 6.65x3.74mm or 6.56x3.69mm on your chart (id lean toward 6.65x3.74mm since my measurements have more likelihood of error than Sony's) I dont think anyone would really notice the <0.1mm difference. I dont have enough faith in my tapemeasure to guarantee much less than +-0.1mm room for error anyway...

BTW just finished your DV Rebel book this past weekend and really enjoyed it even though I am not accustomed to reading materials that are not displayed on an LCD.

As far as I know 1/2" means 8.0mm diagonally. For 16x9 that means 6.97mm x 3.92mm.

This "inch fraction" business is notably inexact. I suppose you could calculate from measured VOF and distance, but I'd say even a few hundredths of a millimeter won't really make a difference.

For reference you may want to just use "the norm".

George/

P.S. I couldn't resist checking Wikipedia (http://en.wikipedia.org/wiki/Image_sensor_format)

Thanks guys: ProLost: Sensor Size Cheat Sheet Update (http://prolost.blogspot.com/2008/08/sensor-size-cheat-sheet-update.html)

-Stu

Thanks for the update, Stu. A neat resource.

Jus.