A few weeks ago I came accross tests you had made on the HD100.

As I remember, at about F4 the ASA was about 300.

Then as you went open to F1.4, the ASA measurement dropped to about 100 or so. My immediate assumption was that although the F-stop was indicating you were allowing more light as you stopped more open, the T-stop was not following the F-stop. Hence, the ASA was falling because less light was really geting through the lens.

But Fujinon claims constant T stop of 1.5. So I'm wondering what your interpretaion of the lower ASA is.

It seems very weird to have a camera that actually gets less sensitive as you open the iris. In fact, it looked like although one opened the iris by 2 stops -- because the ASA dropped in half, you only got 1-stop of extra light.

And as I remember, the ASA changes at smaller F-stops too.

What do you think the ASA is given this variability in the value you measured?

Steve,

We are doing some testings before main photography of a feature to be shot 100% using the HD100. Our results are similar. The test has been really empirical, but in general, the ASA has been around 125 in apertures around 2.8 - 1.4 . Weird... Really no explanation, thought no extensive testing done.

Just my 2 cents, if even worth that...

Respectfully,

Luis Otero

Is it possible that the T stop for the higher apertures is then incorrect, i.e. f4 is more like T2.8?

The manual is very specific about zoom ranges for the F stop specs.

Then for T-stop, it just says 1.5.

It certainly could be an error in the spec, because it seems like 1.5 over the entire zoom range seems too good to be true for a zoom lens that at tele has only a 1.8 f-stop.

Moreover, I can see a T stop varying with zoom. But staying constant?

Luis, are you testing at full wide?

Charles, can T-stop not track F-stop? Is it possible that by 2.8 the T-stop simply doesn't increase in PROPORTION to the f stop?

Luis, what ASA are you getting at f/4 to f/8?

And, at f/11 and f/16?

As I remember, Berry found that at f11 the ASA got even higher. So is it possible, that as the iris is closed the T-stop simply doesn't decrease in PROPORTION to the f stop?

I guess a can imagine that T-stop tracking not tracking F-stop except at the sweet spot of the lens. If this is the case, then we need a table. We also have a headache if one tries to use qa light meter.

Steve, as far as T stops not tracking F stops, I'd well believe it--after all, that's the reason the T-stop concept was devised. With the apparent shortcuts that were taken with this lens to make it available at the given price (as we have seen with the chromatic aberations), I suppose it's not too far fetched to believe that the f-stop markings could deviate from the actual transmission by as much as a stop. Perhaps it's an assumption that the vast majority of users of this camera will set exposure by eye (viewfinder or monitor) rather than mathematically. It would be pretty easy to measure on a waveform; by shooting a gray scale and stopping through the markings on the lens one could measure the output exactly to see if it actually halved or doubled per stop marking. Maybe it would even make sense to have the light source illuminating the gray scale on a dimmer so that you can always set your first stop to the same IRE setting, i.e:

Set lens to f2.0, adjust light to deliver 80 IRE (for example); then change lens to f2.8 and note resultant IRE value.

Set lens to f2.8, dial up light to again deliver 80 IRE and then stop down to f4, noting value...and so on.

In my mind, this would help to calibrate one's lens (you can remark the lens by dialing it as needed to deliver a constant transmission) and perhaps make sense out of this interesting conundrum.

I don't have an HD100 yet so I can't perform this exercise myself, looking forward to hearing someone else's results.

One question--what is the correlation of IRE to stops of transmitted light? For instance, how much should we expect to see the signal drop by as you stop down a single stop? I would have said it should halve or double (i.e. 80 IRE to 40 IRE) but now I am second-guessing myself. Steve?

For instance, how much should we expect to see the signal drop by as you stop down a single stop? I would have said it should halve or double (i.e. 80 IRE to 40 IRE) but now I am second-guessing myself. Steve?

I have always assumed it was 10 IRE to allow 10-stops. (8 IRE to 108 IRE with NTSC and 11-stops 0 to 110, for Japanese NTSC, PAL, and HD.)

But then I read in a textbook that it was 20 IRE per stop. Now, I'm totally unsure.

I'm also confused as to what IRE an 18% gray card should read on a WFM.

Speaking of confusions I continue, to have, and maybe someone could answer these questions:

1) I had always believed the Japanese rate lux by using a gray tone and measuring the light when 100 IRE was reached. IEC?

2) I had always believed the USA ratse lux by using a gray tone and measuring the light when 50 IRE was reached. (That's why are rating arealways 2X better.)

3) What percent gray card do the Japanese use?

4) What percent gray card does the USA use?

5) When Barry measured ASA he didn't seem to use either method. He used one of the gray tones on a chart (which one?) and noted an IRE level (which one?).

I wish I could find his post on the dvx100, but it wasn't all that clear. Which means there may be different folks measuring different ways. But, the number Luis came up with seemed not far from Barrry's number as I remember thinking to myself "Plus X film."

This may help Steve

Click Here (http://www.cybercollege.com/asavideo.htm)

Not sure what a gray card should nominally read at--usually it is one stop down from caucausian skin tone, which the engineers would have us pinning at 70 IRE (too hot, in my opinion). Of course in my test scenario it's only being used as a reference between two camera settings so it could be pinned pretty much anywhere.

Personally, the only time I use my meters on video jobs is for pre-lighting sets or establishing an initial base level, but that's in theory. Honestly, they don't come out of the case in video jobs.

The last feature I shot was on HD, and I'm content with nearly all of the exposures, and I can't recall when I used a meter on it. My second unit DP insisted on using his meter, and some of his day exteriors are too hot. So go figure.

Anyway--let's get to the bottom of this HD100 issue, because I have an important shoot coming up with it and I'd love to know what's going on! (and I thought my only hassle was going to be getting the damn footage into FCP properly--been reading Nate's posts with interest).

But Fujinon claims constant T stop of 1.5. So I'm wondering what your interpretaion of the lower ASA is.
I no longer have an HD100 so I can't re-verify tests, but I can tell you what I remember. I don't believe for a second that the lens is a constant T stop of 1.5. It gets very noticeably darker the further you zoom in. I remember it being about 1 stop slower at full telephoto as versus full wide.

It seems very weird to have a camera that actually gets less sensitive as you open the iris. In fact, it looked like although one opened the iris by 2 stops -- because the ASA dropped in half, you only got 1-stop of extra light.

Well, it's the exact same behavior I observed with the DVX. The conclusion is that the CCD sensitivity is not, in fact, linear. There is no fixed ISO/ASA value. The JVC performed in an exactly similar fashion to the way the DVX did. At f/2.8 the DVX responds at 400 ISO, from f/4 to f/9.6 it responds at 640, and from f/11 to f/16 it goes higher still, up to about 800 or 1000.

The JVC displayed a similar pattern of behavior, with about 100 ISO at the very widest aperture, about 200-250 over most of the exposure curve, and 400 at f/11.

The effective ASA rating of the camera would also be dependant on the gamma curve being used. Also, the auto knee function would give different highlight results under different conditions, so that is why I always set the knee to a manual setting.

As for the Varicam, Panasonic has recommended setting a 18% gray card at 30.5 IRE to determine exposure when shooting for a filmout using FILMOUT gamma. I would imagine the same applies for the JVC HD100 with the FILMOUT curve, but everything to do with the filmout curve on the HD100 seems to be undocumented by JVC. It isn't even mentioned in the manual as an option!

In my tests with a chip chart and gray card (assuming the lens is actually T1.5 at full open/10mm - maintaining full open and dimming source lights instead of stopping down.) I found the effective ASA on the various curves:

Using 18% gray at 55IRE standard:
CINELIKE GAMMA NORMAL/KNEE 100%/MB NORMAL/0dB: 100 ASA (100% white only peaks at 95 IRE)
STANDARD GAMMA NORMAL/KNEE 100%/MB NORMAL/0dB: 200 ASA (100% white only peaks at 98 IRE)
CINELIKE GAMMA MAX/KNEE 100%/MB NORMAL/0dB: 400 ASA (100% white peaks at 100 IRE)
STANDARD GAMMA MAX/KNEE 100%/MB NORMAL/0dB: 600+ ASA (100% white peaks at 102 IRE)

Using 18% gray at 30IRE:
FILMOUT GAMMA/KNEE 100%/MB NORMAL/0dB: 200 ASA (100% white only peaks at 92 IRE)

Is it possible that the T stop for the higher apertures is then incorrect, i.e. f4 is more like T2.8?
Considering the low-cost of the lens, I would assume that the marked ƒ stops are in no way representative of actual transmittance - except at full open. The manual for the lens claims Maximum Photometric Aperature to be T1.5.
It also states that the relative aperature is "ƒ1.4 (5.5mm~68mm) ~ ƒ1.8 (88mm)" So it definitely doesn't hold through the zoom range.

The manual for the lens claims Maximum Photometric Aperature to be T1.5.

Do you read this as the "MAXIMUM" (i.e., BEST) T-STOP IS 1.5 -- which implies that there are conditions where the T-stop might be much less, for example T/2.8?

Or, do you read "MAXIMUM APERATURE" as a term that differentiates the value from "RELATIVE APERATURE."

Or, do you read this as, AT "MAXIMUM APERATURE" -- the T-stop is 1.5?

In your post you say "assuming the lens is actually T1.5 at full open" which seems to be the third interpretation.

The reason I ask, is from a shooter's point of view the third interprestation is the best -- since at low light one would like t/1.5 to be the case when the lens is at f/1.4. Not much loss at all.

However, with either of the other two interpretations, it leaves the possibility that the T-stop at full open is t/2.8, for example. This would indicate your reported ASA values would be larger. (Right?)
---------------------------------------
Are you saying you think the variation in sensitivity folks see in camcorders from wide open to near closed is a result of their not considering gamma?
----------------------------------------
Since the actual light falling on the CCDs is the same at F16 as at F1.4 (because we control the light in our testing) -- I can't see how the camera can tell 100 IRE at f/1.4 from 100 IRE at f/16. If it can't tell, then I don't see how/why sensitivity can vary with the aperature opening.

It seems like the sensitivity variation must be an artifact of the lens transmitting more or less light than the F-stop indicates. Which I'm really willing to believe, but then I don't understand the 1.5 T-stop spec. Unless, it is the first interpretation, i.e., the BEST (MAXIMUM) T-stop the lens can do is 1.5. Which is likely Wide open with mid- to small-aperature sizes.

This is a very misleading spec because I don't think expensive cine lens are speced like that.

It gets very noticeably darker the further you zoom in. I remember it being about 1 stop slower at full telephoto as versus full wide.


Hey Barry,

If we are using the JVC with the Micro35, we need to zoom into the achromat to the ground glass. In that case, we will lost 1 stop?

does the DVX lose light when zoomed in?

Yes to both questions. All these cameras lose at least 1.5 stops over the course of their zooms, the DVX goes from 1.6 on the wide to 2.8, and so does the Sony FX1/Z1; the Canon has a longer zoom so it loses more -- it goes to 3.5 IIRC.

Now, if someone were to design a custom relay lens instead of having to use the JVC lens, you'd get *much* better performance, I'm sure. You'd eliminate the issues with the chromatic aberration and the loss of edge sharpness, and you'd also avoid the light loss from zooming in.

Do you read this as the "MAXIMUM" (i.e., BEST) T-STOP IS 1.5 -- which implies that there are conditions where the T-stop might be much less, for example T/2.8?

Or, do you read "MAXIMUM APERATURE" as a term that differentiates the value from "RELATIVE APERATURE."

Or, do you read this as, AT "MAXIMUM APERATURE" -- the T-stop is 1.5?

In your post you say "assuming the lens is actually T1.5 at full open" which seems to be the third interpretation.
Yes, that is my interpretation.

Are you saying you think the variation in sensitivity folks see in camcorders from wide open to near closed is a result of their not considering gamma?
No. I think the variation in sensitivity is based on the assumption that the F markings on the lens are T markings. They are not. Therefore, the logical conclusion is that the markings do not correspond to actual transmittance at all markings. If the camera is set to completely manual image processing, then I don't see how the sensitivity could possibly change at any time.

Since the actual light falling on the CCDs is the same at F16 as at F1.4 (because we control the light in our testing) -- I can't see how the camera can tell 100 IRE at f/1.4 from 100 IRE at f/16. If it can't tell, then I don't see how/why sensitivity can vary with the aperature opening.
Exactly.

It seems like the sensitivity variation must be an artifact of the lens transmitting more or less light than the F-stop indicates. Which I'm really willing to believe, but then I don't understand the 1.5 T-stop spec. Unless, it is the first interpretation, i.e., the BEST (MAXIMUM) T-stop the lens can do is 1.5. Which is likely Wide open with mid- to small-aperature sizes.
Probably wide open at 5.5mm.

This is a very misleading spec because I don't think expensive cine lens are speced like that.
They have T stop markings so the spec is the fastest T stop.

I agree -- I think 1.5 is the best you get. Just like when you say a 1.4 lens you mean the best.

The difference is that when you say a 1.4 lens we all know you mean the number f1.4 marked on the lens AND with a full Open aperture.

I think Fujinon is saying, yes it has a T-stop of 1.5 -- but not disclosing that this value is only reached at some-point in the aperture range. Maybe in the middle or maybe only around f/11. And, certainly not at full open!

Worse, even this 1.5 value may be only wide open. It could increase dramatically at full Tele.

Which means at full Tele and at f/1.4 the actual T-stop could be 5.6 or worse!

So to REALLY answer Charles's question -- someone needs to create a 2D table with Zoom across the top and f-stop down the side. Then put the ASA values in the cells. Also choose the cell with the highest ASA value and make it a 1. Then rate all the other cells in LOWERED SENSITIVITY is "stops of light."

I think Fujinon is saying, yes it has a T-stop of 1.5 -- but not disclosing that this value is only reached at some-point in the aperture range. Maybe in the middle or maybe only around f/11. And, certainly not at full open!

Huh????

Worse, even this 1.5 value may be only wide open. It could increase dramatically at full Tele.

Which means at full Tele and at f/1.4 the actual T-stop could be 5.6 or worse!
4 stops? I highly doubt it. I just measured complete consistency from 5.5mm to 40mm and then a drop of 10 to 15 IRE with heavy vignetting as I approach 88mm. BTW, the vignetting seems to start subtley at around 20mm.
So to REALLY answer Charles's question -- someone needs to create a 2D table with Zoom across the top and f-stop down the side. Then put the ASA values in the cells. Also choose the cell with the highest ASA value and make it a 1. Then rate all the other cells in LOWERED SENSITIVITY is "stops of light."

A simpler and better solution would to simply measure actual transmittance at each marked F stop using the published spec of T1.5 as the starting point for full open. However, based on Barry's experience with his lens, each lens would probably have slight variances with results.

Yes indeed. I wasn't looking at the variance in light transmittal across the zoom range, only the question of why the sensitivity of the camera appeared to shift across the APERTURE range, so my theory and suggested experiment were designed to measure the actual transmittance (i.e. T-stops) of the lens against the markings (f-stops) which I thought might shed light on the conundrum. I can't see why a camera would exhibit different sensitivity (ASA) at different apertures--the imager doesn't "care" what the aperture setting is, it only knows the amount of light hitting it.