Is the DOF that good?

[David Koo]

I think that the image on a 1/2" sensor will be better than 1/3" sensors. But like some have said, I doubt that the image will "blow away" all the competition (except maybe in very poor lighting).

NOW... The shallow DOF with the 1/2" sensor is VERY interesting.... Trying to blur out the background with 1/3" sensor camcorders is HARD—often I'm 30 feet out from the subject, zoomed to the max, and the background is another 40-50 feet away (most rooms are not made to those dimensions)...

Does anyone know how good the XDCAM EX is going to be at achieving that "blurred out" background for interview headshots?

dave

[Alex Amira]

Originally Posted by David Koo

I think that the image on a 1/2" sensor will be better than 1/3" sensors. But like some have said, I doubt that the image will "blow away" all the competition (except maybe in very poor lighting).

NOW... The shallow DOF with the 1/2" sensor is VERY interesting.... Trying to blur out the background with 1/3" sensor camcorders is HARD—often I'm 30 feet out from the subject, zoomed to the max, and the background is another 40-50 feet away (most rooms are not made to those dimensions)...

Does anyone know how good the XDCAM EX is going to be at achieving that "blurred out" background for interview headshots?

dave

While distance between subject and background is use to achieve shallow DOF there is also another way.

Well I can say that since one of the things that affects DOF is a wider aperture on the lens you can use the ND filters to help you achieve this. The ND filter will act as a light gate to help you maintain proper exposure when otherwise the picture would be really overexposed.

Normally if you open the aperture to a low enough f stop (required for that shallow DOF) it will most likely be letting too much light in and the picture will be ruined. By using the ND filters you can dialing a lower that usual f stop and still have the proper exposure.

You can do this on any camcorder. Of course having bigger CCDs helps. I can tell you that this technique works wonders on my VX2000 with it's 1/3 CCDs.

The PMW-EX1 has 2 ND filters just like most of the Sony units do.
You can also attach your matte box and drop some additional ND filters if the built in ones are not enough or you can't control the lighting.

Using some lights indoors will work out nicely for the interviews and help with the fact the low light factor. I assume the EX1 is going to be really good in low light. The article I read did mentioned that since it uses native HD CMOS chips they would not be as good as the SD chips upresed to HD (like the HVX200) because it has more pixels and thus they each get a bit less light since they are smaller. I think this debatable and overall only in some cave you will be able to notice that small edge in low light. Based on Sony's reputation for lowlight the EX1 will be a great performer.

In any case normally most interviews are lit.

I found that the 2ND filters in my VX2000 worked like a charm and I never had to use external ones.

[Bob Grant]

Originally Posted by David Koo

I think that the image on a 1/2" sensor will be better than 1/3" sensors. But like some have said, I doubt that the image will "blow away" all the competition (except maybe in very poor lighting).

NOW... The shallow DOF with the 1/2" sensor is VERY interesting.... Trying to blur out the background with 1/3" sensor camcorders is HARD—often I'm 30 feet out from the subject, zoomed to the max, and the background is another 40-50 feet away (most rooms are not made to those dimensions)...

Does anyone know how good the XDCAM EX is going to be at achieving that "blurred out" background for interview headshots?

dave

Slightly better in theory. Just the ratio of imager size between 1/3 and 1/2 so not a huge amount better, in theory!
However other factors can have an influence. Obviously the shallowest DOF is with the iris wide open and the lenses on the cheaper cams don't perform as well wide open. So a better lens might also help compared to what the theory predicts. Also as you zoom in more the f stop is affected more on cheaper lenses, so again not so obvious factors could make a difference.
As for using the built in ND filters, this worried me a bit. They're very savage in the amount of light they block. Indoors you risk having the camera wind the gain up which isn't what you want. I think you really want to hit the sweet spot of iris wide open with no gain. For that reason more gentle ND filters in front of the lens would seem the way to go.

[David Koo]

Even with the aperture wide open, ND filters in place, maximizing the background distance, and zooming in as much as possible, the DOF is still too wide with 1/3" cameras (in most indoor interview situations). I was hoping that 1/2" chip cameras would offer a narrower DOF without having to resort to using clumsy 35mm adapters...

thanks
dave

[Charles Papert]

If you are comparing to 35mm adaptors, the difference between 1/3 and 1/2 chips is going to be nominal at best.

For your interviews, chances are that for most that can be reasonably achieved without having to go to the extremes that you mentioned, you will see a slight softening of the background with the 1/2 chip. Slight.

Those who shoot with 2/3" chips have to go through similar hoops to get this effect but are able to get something that begins to approximate a 35mm shallow DoF look (generally with at least twice the focal length), so the 1/2 chip will be somewhere between this and the 1/3.

In other words, it won't be "magic". But consider that a late model 1/2 chip will offer noticeable improvements over a 2 year old 1/3 chip even outside of the larger image area.

[Jiri Bakala]

On the upside, maintaining a reasonably good focus won't be quite as difficult as with a full size HD chips. Remember, focus in HD is very critical!
:-)

[Steven Thomas]

I'd be interested in knowing the math for this.
The 1/2" sensor has a 50% increase in size (.5/.333 =1.5).

We know it's not a linear fit for DOF; therefore, DOF would not be 50% tighter at a given focal length and f-stop.

Does anyone know the relationship between sensor size and DOF?

It's hard to tell, but some of the XDCAM sample footage on the net seemed to display a tighter DOF. Of course, that's really hard to tell not knowing how far the camera was away from the subjects, focal length, f-stop...

[Greg Boston]

Originally Posted by Steven Thomas

It's hard to tell, but some of the XDCAM sample footage on the net seemed to display a tighter DOF. Of course, that's really hard to tell not knowing how far the camera was away from the subjects, focal length, f-stop...

The way Sony employed the 1/2 sensors on the full size cameras, they use the full face of the sensor, in effect, giving them a bit more performance than 1/2 sensors of the past. Not quite 2/3, but more than a 1/2 sensor.

Want some proof? Go to the following link and plug in some numbers on the DOF tab. I used a subject distance of 4 feet, a focal length of 5.5 (the widest my lens will go), and a 1.8 f/stop. Look at the calculation for XDCAM HD vs. 1/2 and I think you'll be amazed. Do the same using 2/3 HD and you'll see that XDCAM HD isn't quite as good, but much better than standard 1/2. It's a Java applet.

http://www.fujinon.com/Broadcast/Opt...alculator.aspx

Sony mentioned this little blurb about the XDCAM HD sensors in a press release prior to NAB 07.

-gb-

[Piotr Wozniacki]

Greg, thanks for the great link!
However, either I'm missing something, or it shows shallower DOF for 1/3" than it does for 1/2" imager (at 3m subjct distance, f1.9, focal length of 30mm). Why is that?

[Barry Green]

Originally Posted by Greg Boston

Want some proof? Go to the following link and plug in some numbers on the DOF tab. I used a subject distance of 4 feet, a focal length of 5.5 (the widest my lens will go), and a 1.8 f/stop. Look at the calculation for XDCAM HD vs. 1/2 and I think you'll be amazed. Do the same using 2/3 HD and you'll see that XDCAM HD isn't quite as good, but much better than standard 1/2. It's a Java applet.

Okay, that's really curious though, because when you look at the field of view comparison, that Fujinon calculator disagrees with Sony's claim about "larger surface area." So why would it be showing a DOF difference?

The difference between the XDCAM HD option and the 1/2" option (in that calculator) is the difference between a 16:9 sensor and a 4:3 sensor. Check the angle of view and you'll see that the XDHD image gets wider, but the height gets smaller.

Then, compare the ratio of 1/2" 4:3 to 2/3" 4:3 in the field of view; you'll see that 2/3" is 37.5% wider, and 37.5% taller. Then do the same comparison for XDCAM HD vs. 2/3" 16:9 and you'll see the same ratios hold true: a 2/3" 16:9 sensor has a 37.5% wider field of view, and a 37.5% taller field of view, than the XDHD sensor (according to Fujinon's calculator).

I don't know quite how they're arriving at different DOF numbers based on the same focal length, but I suspect that it has to do with cropping the 4:3 sensor to 16:9 to get an equivalent frame. If so, that means they're having to back up the camera a little ways to get the same width frame (or using a slightly wider-angle lens on the 1/2" option), and that would have the effect of increasing the depth of field.

In any case, if you find that Fujinon calculator's field of view projections to match your real-world experiences, then it's pretty much demonstrating that the claim of their sensor using more surface area to be simply not accurate. The Fujinon calculator is showing exactly what one would expect from a 16:9 1/2" chip, in relation to a 16:9 2/3" chip, and that the differences are identical to the differences between today's 4:3 1/2" chip and 4:3 2/3" chips.

[Barry Green]

Originally Posted by Piotr Wozniacki

Greg, thanks for the great link!
However, either I'm missing something, or it shows shallower DOF for 1/3" than it does for 1/2" imager (at 3m subjct distance, f1.9, focal length of 30mm). Why is that?

I'm not Greg, but I'm pretty sure I can answer. It's because the field of view is so radically different. You wouldn't shoot that shot with the same settings; for a 1/3" imager you'd need to use a wider focal length to get the same field of view. With the same focal length, you'd be getting an essentially telephoto shot, which would optically magnify the background much more and make it appear to be more out of focus.

At a distance of 3 meters, and a focal length of 30mm, with a 1/2" chip you're getting a field of view of 2.079 x 1.559 feet. With a 1/3" chip you'd have to use a focal length of 23mm to get the same field of view. If we then use that new focal length in the DOF calculator, you'll find that the 1/2" system has a DOF of 1.995 feet, and the 1/3" is 2.348 feet (more like what you'd expect).

[Piotr Wozniacki]

Originally Posted by Barry Green

I'm not Greg, but I'm pretty sure I can answer. It's because the field of view is so radically different. You wouldn't shoot that shot with the same settings; for a 1/3" imager you'd need to use a wider focal length to get the same field of view. With the same focal length, you'd be getting an essentially telephoto shot, which would optically magnify the background much more and make it appear to be more out of focus.

At a distance of 3 meters, and a focal length of 30mm, with a 1/2" chip you're getting a field of view of 2.079 x 1.559 feet. With a 1/3" chip you'd have to use a focal length of 23mm to get the same field of view. If we then use that new focal length in the DOF calculator, you'll find that the 1/2" system has a DOF of 1.995 feet, and the 1/3" is 2.348 feet (more like what you'd expect).

Barry, thanks for this excellent explanation! I knew I was doing some mistake... Now I know exactly how it works!

[Evan Donn]

Originally Posted by Barry Green

You wouldn't shoot that shot with the same settings; for a 1/3" imager you'd need to use a wider focal length to get the same field of view. With the same focal length, you'd be getting an essentially telephoto shot, which would optically magnify the background much more and make it appear to be more out of focus.

This is a great point that gets missed a lot - the size of the chips doesn't directly affect the depth of field. A lens of a given focal length and aperture should always have the same depth of field, regardless of what you put behind it. Where the chip size comes in to play is that wider lenses have to be used in order to achieve a standard field of view on a smaller imager, and, in general, the wider the lens the greater the apparent depth of field.

Am I correct in thinking also that longer focal lengths should mean less barrel distortion at wide angles? Another issue I haven't seen mentioned much - shouldn't the larger imager make it easier to produce a lens with sufficient resolution to max out the imager capabilities simply because glass imperfections become less critical?

[Steven Thomas]

Using the above mentioned lens calculator:
http://www.fujinon.com/Broadcast/Opt...alculator.aspx

Both cameras opened up to 1.9 f-stop (XDCAM EX wide open. Note HD100 opens to 1.4).
I'm not sure the effects of wide open on the EX, but the HD100 at 1.4 goes a tad soft.
For the purpose of this test, both were calculated at 1.9 f-stop.

XDCAM 1/2:
Subject 3M away and a Focal length 20mm giving us a Field Of View of: 3.407ft x 1.906ft

To match this with the JVC HD100:
Subject 3M away with a focal length 15mm giving us a Field Of View of: 3.408ft x 1.920ft

The Field Of Views are now closely matched.

Giving the above FOV match to the XDCAM
HD100:
DOF Near: 2.664M
DOF FAR: 3.433M
Therefore: Depth = 0.768M
Hyper: 23.684M

Giving the above FOV match to the JVC HD100:
XDCAM:
DOF Near: 2.596M
DOF FAR: 3.553M
Therefore: Depth = 0.957M
Hyper: 19.139M

Just to clear any confusion:
So in this case, by adjusting the 1/3" sensor camera focal length (HD100) wider to closer match the FOV of the 1/2" XDCAM, the DOF technically are the same.
I believe this thought is correct?, but if not, please set me straight.
If not, then the only plus would be as we move to the wider end of the 1/2" sensor lens.

So, with that in mind...

XDCAM 1/2":
Subject 3M away and a Focal length 5.8 mm (max wide on XDCAM EX)giving us a Field Of View of:
11.806ft x 6.640ft

JVC HD100 1/3":
Subject 3M away with a focal length 5.5 mm (max wide on HD100)giving us a Field Of View of:
9.325ft x 5.252

At their given widest VOF and both cameras at thier max aperture, (XDCAM EX at 1.9. HD100 at 1.4), and focused on a subject 3 meters away, the HD100 (according to the calculator used here?), far DOF falls off at 9.77M, where the XDCAM EX runs from a meter to infinity. What the? I thought it would of been a bit tighter at full wide than the HD100? What's up with this calculator or my math? LOL

[Greg Boston]

Originally Posted by Barry Green

The difference between the XDCAM HD option and the 1/2" option (in that calculator) is the difference between a 16:9 sensor and a 4:3 sensor. Check the angle of view and you'll see that the XDHD image gets wider, but the height gets smaller.

Yup, I missed that in the calculator. The 16:9 vs. 4:3 part of the equation.

As to equivalent field of view, I presumed (incorrectly) that they were compensating for that in the calculator. I was thinking that as I clicked between XDCAM HD and 2/3 16:9 I should be multiplying 5.5x1.37 for the equivalent FOV.

So disregard what I said. I had seen that calculator once before and went back to find it when this topic came up.

Nonetheless, I still think the DOF on the camera is pretty good and Sony did indeed make that claim about the sensor performance in a pre NAB press release.

-gb-