Video recording surgery

[Duncan Say]

On a completely different tack. One of the most successful cameras I've ever used in operating theatre is a Sony FS100. It seems wrong that a large sensor camera would work but in fact it was fantastic because the huge amount of light and the large sensor size meant that you could use a small aperture without the diffraction problems.

About a year ago I asked Alan Roberts the ex-BBC guru what aperture you could go up to before diffraction becomes noticeable and he said that for this size of sensor you'd be looking at F30 according to Lord Rayleigh's criterion.

[Mazhar Celikoyar]

Duncan
I checked out the cameras you suggested and the ones you would start using at the operating theatre. They seem to be way out of my budget. As I might be happy with a relatively small one, yet with manual control options, I am in need to find out the camera with a sensor and lens that would make it possible to record all the details within a surgical field of 2-3 cm width and a depth of 1-2 cm. What else would you suggest?

[Duncan Say]

I hate to say this put you get what you pay for especially in the medical side. Perhaps you should have a further look at the Panasonic AC AG90 that was mentioned up top. With a some ND filters it may do the job for you. I've no idea about its field of view though, maybe you should see a dealer and have a play.

My warning is that to get that field of view most cameras can only achieve this when fully wide, so you will find that although they can get a picture that close-in the downside is that the camera is also only about 5cm from the subject - which would not be appropriate for surgery.

It's horses for courses in this area and you are asking to do something very specialised with equipment not designed for the purpose.

[David Heath]

Originally Posted by Duncan Say

On a completely different tack. One of the most successful cameras I've ever used in operating theatre is a Sony FS100. It seems wrong that a large sensor camera would work but in fact it was fantastic because the huge amount of light and the large sensor size meant that you could use a small aperture without the diffraction problems.

But there's a catch..... the large sensor tends to give higher inherent sensitivity - which will mean it will need to be operated at a smaller f-stop than the camera with the smaller sensor! So you don't gain (or lose) anything. Point is that if it's very bright, you're going to need ND to avoid diffraction limiting - end of story.

About a year ago I asked Alan Roberts the ex-BBC guru what aperture you could go up to before diffraction becomes noticeable and he said that for this size of sensor you'd be looking at F30 according to Lord Rayleigh's criterion.

If you accept f5.6 as the limit for a camera with 1/3" chips (it's one estimate, but depends how critical you are, many may consider it wise to stay wider than f4), then it follows that every time you double the chip area, you can stop down another stop. So in approximate terms, s35 has about 32x the area of 1/3" chips, so you should be able to go about 5 stops further closed with s35 than 1/3" - which would be f64. Alan Roberts figure is based on f4 with 1/3" chips - I wouldn't argue with that, but this is really an illustration of how you can derive a ball park figure for any chip size.

If you want a very rough ball park guide to how sensor sizes relate, then for each of the figures below there is roughly a doubling of area, hence a stop difference for such as diffraction limiting/dof calcs etc

1/3"
1/2"
2/3"
1"
4/3
s35

And I stress these are only approximations!

[Duncan Say]

Originally Posted by David Heath

But there's a catch..... the large sensor tends to give higher inherent sensitivity - which will mean it will need to be operated at a smaller f-stop than the camera with the smaller sensor! So you don't gain (or lose) anything. Point is that if it's very bright, you're going to need ND to avoid diffraction limiting - end of story.

If you accept f5.6 as the limit for a camera with 1/3" chips (it's one estimate, but depends how critical you are, many may consider it wise to stay wider than f4), then it follows that every time you double the chip area, you can stop down another stop. So in approximate terms, s35 has about 32x the area of 1/3" chips, so you should be able to go about 5 stops further closed with s35 than 1/3" - which would be f64. Alan Roberts figure is based on f4 with 1/3" chips - I wouldn't argue with that, but this is really an illustration of how you can derive a ball park figure for any chip size.

If you want a very rough ball park guide to how sensor sizes relate, then for each of the figures below there is roughly a doubling of area, hence a stop difference for such as diffraction limiting/dof calcs etc

1/3"
1/2"
2/3"
1"
4/3
s35

And I stress these are only approximations!

Yes there is a catch but in practice it worked a treat, so my experience tells me that this is a good route to go because you get a much greater depth of field and none of the diffraction problems. And here's Alan Roberts quote in full:

"For HDTV 1920x1080 and 2/3" sensors, the limit comes at F/10.2 for blue, F/5.9 for red. You can scale everything from that, since the 2/3" image is 9.6mm wide. So, going up to full-frame 35mm, 36x24, the limits come out at F/38 and F/22, call it F/30 for green. And so on."

[David Heath]

Originally Posted by Duncan Say

Yes there is a catch but in practice it worked a treat, so my experience tells me that this is a good route to go because you get a much greater depth of field and none of the diffraction problems.

No, it may well have worked - but assuming the same lighting level, angle of view, and no NDs it won't have gained you anything over much smaller chips. If you work it through, then the depth of field characteristics and the limit for diffraction softening are more or less identical between f4 on 1/2" chips and f16 on s35. That's no accident - in each case the physical diameter of the aperture will be the same in mm. (To keep the same angle of view, the focal length of the lens must change - if the physical diameter of the iris stays the same, the f no must then change as f no is defined as focal length/iris diameter.)

And since bigger chips means bigger ISO, then it's exactly what will happen - the higher base ISO of the s35 sensor will demand a smaller aperture (f stop).

That's not the whole truth as it's assuming all else equal as the sensor size changes. In practice, a small chip camera (like an EX1) will be three chip, versus the single chip of the FS100. Single chip loses about a stop of sensitivity compared to 3 chip, so in the example above, if the EX1 had to be at f4, I'd expect the FS100 to be at f11 rather than f16. Hence about a stop better for diffraction limiting issues, a stop shallower depth of field.

And here's Alan Roberts quote in full:

"For HDTV 1920x1080 and 2/3" sensors, the limit comes at F/10.2 for blue, F/5.9 for red. You can scale everything from that, since the 2/3" image is 9.6mm wide. So, going up to full-frame 35mm, 36x24, the limits come out at F/38 and F/22, call it F/30 for green. And so on."

But Alan is there referring to full-frame sensors quite explicitly - "going up to full-frame 35mm, 36x24" - whereas the FS100 is s35, about half the area.

From the Sony website, the dimensions are 23.6mmx13.3mm. Doing the same maths as Alan did for full-frame, the figures for blue and red work out as f25 and f14.5. So expect diffraction limiting for the FS100 to become an issue somewhere around f16-22.

That corresponds to my rough and ready table which predicts limiting for s35 to start about 3 stops smaller than 2/3". So (using Alan's figures) if it starts at f8 for 2/3", the table would predict f22 for s35, f4 for 1/3".

[Mazhar Celikoyar]

Hi David, Hi Duncan,

I thank you both, for the comments and the discussion. As you both are highly qualified in this field and I am just an amateur, I need some guidance with the camera model and the equipment that is needed along. So far, most of my colleagues and I, we have been able to use videos obtained by the hand held camcorders. At this point, I would like to carry the equipment up one or two levels, and that is why my inquiry.

My recordings will be of macro, which means lens, with an aperture wide enough, yet without losing depth of field. With manual focus, to avoid blurring caused with the instruments getting in and out of the field.

The recording might continue for several hours, which precludes using my digital SLR.

White balance is another prerequisite, which I think only some handheld camcorders come with that feature.

The benefit of having such a small camera is, having it fixed at about 40 cm.s from the surgical field, we can get almost an identical view the surgeon gets.

So, I would appreciate it very much, if you guys or anyone could indicate me proper camcorders.

Once again, thank you.

Mzhar

[Duncan Say]

Originally Posted by David Heath

No, it may well have worked - but assuming the same lighting level, angle of view, and no NDs it won't have gained you anything over much smaller chips. .

Which is the big difference, you have to use ND or some other method to control the light the with a small chip camcorder and you didn't with the FS100. QED it worked.

[Duncan Say]

Originally Posted by Mazhar Celikoyar

Hi David, Hi Duncan,

I thank you both, for the comments and the discussion. As you both are highly qualified in this field and I am just an amateur, I need some guidance with the camera model and the equipment that is needed along. So far, most of my colleagues and I, we have been able to use videos obtained by the hand held camcorders. At this point, I would like to carry the equipment up one or two levels, and that is why my inquiry.

My recordings will be of macro, which means lens, with an aperture wide enough, yet without losing depth of field. With manual focus, to avoid blurring caused with the instruments getting in and out of the field.

The recording might continue for several hours, which precludes using my digital SLR.

White balance is another prerequisite, which I think only some handheld camcorders come with that feature.

The benefit of having such a small camera is, having it fixed at about 40 cm.s from the surgical field, we can get almost an identical view the surgeon gets.

So, I would appreciate it very much, if you guys or anyone could indicate me proper camcorders.

Once again, thank you.

Mzhar

It may well be that if you were to buy the Panasonic that has been recommended and the purchase some Close-up diopters (they come in a range of strengths that could work for you) then you will probably achieve the field of view that you require from the distance that you can place the camcorder. But I can't promise it.

[David Heath]

Originally Posted by Duncan Say

Which is the big difference, you have to use ND or some other method to control the light the with a small chip camcorder and you didn't with the FS100. QED it worked.

I'm sure it will have worked - but no better than a smaller chip camera, that's the point.

Compare the FS100 (and s35) with an EX1 (and 1/2" chips) and for depth of field purposes then they will give identical results at f4 and f16 respectively - this follows from the s35 chip size being (roughly) 16x the area. Now stop the EX1 down more than about another stop and diffraction limiting will start to take effect - but stop the FS100 down by a similar result from f16 and it's the same problem.

Basically, for a given system resolution there's an effective limit to how great the depth of field can be for any given angle of view. To increase depth of field, you use a smaller f stop. But the limit is set by diffraction limiting, and physically that is down to actual physical dimensions of the iris. Sensor size becomes irrelevant. The point of large format cameras is exactly the opposite - they can enable you to achieve SHALLOW depth of field that is physically impossible with small sensors, but that's not what is relevant here. (If the dof required was given by s35 at f2, you'd need a theoretical f0.5 lens for 1/2". Practically, it becomes virtually impossible to make lenses with that sort of aperture, there's very few that have f stops wider than around f1.4 at best.)

So the message to Mazhar is forget about sensor size! It's irrelevant if you want deep dof, that will be limited solely by diffraction limiting as you stop down and the limit will be independent of sensor size.

[David Heath]

Originally Posted by Mazhar Celikoyar

Hi David, Hi Duncan,

My recordings will be of macro, which means lens, with an aperture wide enough, yet without losing depth of field. With manual focus, to avoid blurring caused with the instruments getting in and out of the field.

The recording might continue for several hours, which precludes using my digital SLR.

White balance is another prerequisite, which I think only some handheld camcorders come with that feature.

The benefit of having such a small camera is, having it fixed at about 40 cm.s from the surgical field, we can get almost an identical view the surgeon gets.

So, I would appreciate it very much, if you guys or anyone could indicate me proper camcorders.

Once again, thank you.

Mzhar

I agree with a previous post of Duncan's when he mentions Sony and Panasonic solutions especially designed for medical use. Apart from specific approval, they both feature a small head connected to a base station by cable - hence what needs to be located 40cm from the surgical field is as small and lightweight as possible, and that consequently eases mounting issues as well.

For these sorts of reasons I would NOT consider the AG90, or indeed any sort of general purpose camcorder.

If the true medical solutions are too expensive, then if you can get it cleared for operating theatre use, then consider this - Panasonic AG-HCK10 (AG HCK10) 3MOS HD minicam with 12x optical zoom and Image Stabiliser - Ideal for use with AG-HMR10 recorder , for which you will also need the remote control box/recorder. (AG-HMR10).

As well as giving the features you require (manual zoom, focus iris etc) they are operated from the control box, which seems to me far preferable than having to make adjustments on the camera itself. The spec doesn't say what it's minimum focussing distance is - but does say it has a 43mm filter size. Consequently, I'd be inclined to get a 43mm 2 dioptre CU filter - to bring the max focusable distance down to 50cm, then rely on internal focussing to trim for closer distances.

For the reasons stated in the post above, you'll want it to be operating at an f stop just open of the diffraction limiting point. If the level of lighting is such that for correct exposure that is too small, then just use 43mm ND of the right amount to hit the sweet spot. (You can stack the CU and ND.)

I don't believe this camera has specific medical approval - but then neither does such as the AG90 or any normal handheld camera. At least this head is small and lightweight and easily mounted.

[Tim Maloy]

We used a Sony HXR-MC1P very successfully for live surgery webcasting.
The light weight of the head allowed us to mount on a boom arm, leaving the operator well outside the sterile sphere.
We combined the footage with two EX3's.
It all worked great.
The bit I was happiest about was taking all the gear needed to Germany and still being underweight on EasyJet!.