EX1 misreading Kino Flo color temp

Bruce Schultz · October 15th, 2009, 09:26 PM

I will be testing this weekend also with both cameras hooked up to a Leader Vectorscope on DSC charts and will post results next week.

Andrew Stone · October 15th, 2009, 11:46 PM

One thing that could be playing into this is "green spike", if you dim the DIVA the green becomes more exaggerated. DIVA's are dimable Kinos so be sure to crank them up if you want to avoid this issue. If it is there it could be affecting your white balance and thereby the temperature reading coming out on the camera display... Just a thought.

I assume you don't have light spill from any other source when you are having this problem.

Bruce Schultz · October 16th, 2009, 11:17 AM

That's another thing I'll be testing - the white balance readings at various dimmed levels. No spill issues so far in my observations of this problem since it's happened in a few very controlled environments.

Bruce Schultz · October 18th, 2009, 01:23 PM

These are my initial test findings with an F900R / EX1 side-by-side testing in a lighting-controlled environment.
( > xxxx Kelvin = readings by camera )

F900R: Joker 5600> 5600 Kelvin
F900R: Diva 5600> 5700 Kelvin
F900R: Diva 3200> 3200 Kelvin
F900R: Diva 2900> 2900 Kelvin
F900R: Mole 3200> 3100 Kelvin

EX1: Joker 5600> 6100 Kelvin
EX1: Diva 5600> 9100 Kelvin
EX1: Diva 3200> 3500 Kelvin
EX1: Diva 2900> 3300 Kelvin
EX1: Mole 3200> 3100 Kelvin

The only thing that is correctly read by the EX1 is the Mole Richardson tungsten light (probably slightly aged and reading 31K instead of 32K) which is a non-gas bulb as opposed to the Joker 200 HMI and the Diva Kino Flo flourescents - both of which have inaccurate readings by the EX1. Readings on the F900R were done using the 1B (tungsten)filter so no CC through the filter wheel is in play. Vectorscope shows the F900R dead on for color representations - not so for the EX1 on presets to match the correct F900R settings. Only when the EX1 is white balanced does it behave as it should on the Vectorscope.

It's easy to remedy the Kino Flo and Joker temperature readings with 1/4 CTO on either light, but that doesn't help a 2-camera shoot - as correcting for one messes up the other especially in the background areas mutual to both shots.

The only reason I'm even noticing this now is that clients who mostly prefer to use an F900R for interview situations are, for declining budget reasons opting to do those with the EX1 and save costs. Having only shot B-roll with the EX1 for almost two years I've just started putting it in my normal lighting environments and discovering these inconsistancies of color temperature.

I'm posting this also to CML to see if anyone there has discovered and remedied this problem.

Leonard Levy · October 18th, 2009, 03:33 PM

Bruce,
I would be very interested to see screen shots of these tests.
I don't care a whit if the EX-1's color temps read out is wrong, but I do care if the pictures themselves indicates a different response to these lights. If the EX-1 reacts to these lights in a way that corresponds with its color temp readings that would of course also be very interesting
To some extent I would expect color to be somewhat different between the F900 and the EX-1 just from the fact they are different cameras and their matrixes may be also set up differently. Its always hard to match two different cameras. However if you're right and the Ex-1 responds way differently to the Diva and the Joker tht also would be cause for serious concern on my part.
Can you post stills from your tests?

Lenny

Bruce Schultz · October 18th, 2009, 04:35 PM

Leonard, I didn't record any stills during this test (my bad) and I've taken it down to prepare for a busy shooting week. All I can say is that I'm convinced that the EX1 has a problem correctly identifying gas-type bulb temperature and I'm going to see if Sony repair, the other geniuses on CML, or my private Sony maintenance person can shed any light on this.

To answer your question without visuals, yes when the EX1 white balances on either a Joker or a Diva it has an accurate DSC chart representation. That's not true when I fudge the WB with a 32K or 56K preset.

Leonard Levy · October 18th, 2009, 05:02 PM

Can you pull still off the video?

Bruce Schultz · October 18th, 2009, 09:45 PM

Quote:

Originally Posted by Leonard Levy

Can you pull still off the video?

Leonard, I didn't feel the need to record these test shots, I just took notes off live camera signal readings. Next time I'll think about recording it.

Bob Grant · October 19th, 2009, 04:04 AM

It seems to me that relying on a single number to describe white light is a bit limiting. Color temperature refers to a black body. Even a tungsten lamp is not exactly a black body and discharge lamps are even further removed from creating light the same as a black body.

I had posted about another oddity with the EX1 some time ago. I bought a set of Kodak grey cards and the CT as reported by my EX1 is significantly different between the white and grey side of the cards. That was under fluro lighting.

Serena from here if I recall correctly checked the same cards under daylight and got a match between the white and grey sides of her cards.

What I suspect is going on is this. We know that Sony extended the range of the red sensor's filter. What we don't know is if they changed the green and blue as well but it's likely this camera is a bit different to any other camera.

I know that some fluro lights designed for video use are made to match the RGB sensors in a video camera. If Sony has changed the colorimetry of the sensors it's not too hard to see how these lights will produce a different outcome with a camera with different colorimetry. I've used so called tungsten fluro lamps from Osram and to the eye they are not quite the same as tungsten lights yet they look the same to a video camera. I'm wondering if this is what the EX cameras are seeing as well.

I wish I had the instruments to measure any of this.

October 15th, 2009, 09:26 PM	#16
Bruce Schultz Major Player Join Date: Apr 2008 Location: Malibu, CA Posts: 480	I will be testing this weekend also with both cameras hooked up to a Leader Vectorscope on DSC charts and will post results next week.

October 15th, 2009, 11:46 PM	#17
Andrew Stone Major Player Join Date: Sep 2008 Location: Vancouver, Canada Posts: 975	One thing that could be playing into this is "green spike", if you dim the DIVA the green becomes more exaggerated. DIVA's are dimable Kinos so be sure to crank them up if you want to avoid this issue. If it is there it could be affecting your white balance and thereby the temperature reading coming out on the camera display... Just a thought. I assume you don't have light spill from any other source when you are having this problem.

October 16th, 2009, 11:17 AM	#18
Bruce Schultz Major Player Join Date: Apr 2008 Location: Malibu, CA Posts: 480	That's another thing I'll be testing - the white balance readings at various dimmed levels. No spill issues so far in my observations of this problem since it's happened in a few very controlled environments.

October 18th, 2009, 01:23 PM	#19
Bruce Schultz Major Player Join Date: Apr 2008 Location: Malibu, CA Posts: 480	These are my initial test findings with an F900R / EX1 side-by-side testing in a lighting-controlled environment. ( > xxxx Kelvin = readings by camera ) F900R: Joker 5600> 5600 Kelvin F900R: Diva 5600> 5700 Kelvin F900R: Diva 3200> 3200 Kelvin F900R: Diva 2900> 2900 Kelvin F900R: Mole 3200> 3100 Kelvin EX1: Joker 5600> 6100 Kelvin EX1: Diva 5600> 9100 Kelvin EX1: Diva 3200> 3500 Kelvin EX1: Diva 2900> 3300 Kelvin EX1: Mole 3200> 3100 Kelvin The only thing that is correctly read by the EX1 is the Mole Richardson tungsten light (probably slightly aged and reading 31K instead of 32K) which is a non-gas bulb as opposed to the Joker 200 HMI and the Diva Kino Flo flourescents - both of which have inaccurate readings by the EX1. Readings on the F900R were done using the 1B (tungsten)filter so no CC through the filter wheel is in play. Vectorscope shows the F900R dead on for color representations - not so for the EX1 on presets to match the correct F900R settings. Only when the EX1 is white balanced does it behave as it should on the Vectorscope. It's easy to remedy the Kino Flo and Joker temperature readings with 1/4 CTO on either light, but that doesn't help a 2-camera shoot - as correcting for one messes up the other especially in the background areas mutual to both shots. The only reason I'm even noticing this now is that clients who mostly prefer to use an F900R for interview situations are, for declining budget reasons opting to do those with the EX1 and save costs. Having only shot B-roll with the EX1 for almost two years I've just started putting it in my normal lighting environments and discovering these inconsistancies of color temperature. I'm posting this also to CML to see if anyone there has discovered and remedied this problem.

October 18th, 2009, 03:33 PM	#20
Leonard Levy Trustee Join Date: Feb 2006 Location: USA Posts: 1,684	Bruce, I would be very interested to see screen shots of these tests. I don't care a whit if the EX-1's color temps read out is wrong, but I do care if the pictures themselves indicates a different response to these lights. If the EX-1 reacts to these lights in a way that corresponds with its color temp readings that would of course also be very interesting To some extent I would expect color to be somewhat different between the F900 and the EX-1 just from the fact they are different cameras and their matrixes may be also set up differently. Its always hard to match two different cameras. However if you're right and the Ex-1 responds way differently to the Diva and the Joker tht also would be cause for serious concern on my part. Can you post stills from your tests? Lenny

October 18th, 2009, 04:35 PM	#21
Bruce Schultz Major Player Join Date: Apr 2008 Location: Malibu, CA Posts: 480	Leonard, I didn't record any stills during this test (my bad) and I've taken it down to prepare for a busy shooting week. All I can say is that I'm convinced that the EX1 has a problem correctly identifying gas-type bulb temperature and I'm going to see if Sony repair, the other geniuses on CML, or my private Sony maintenance person can shed any light on this. To answer your question without visuals, yes when the EX1 white balances on either a Joker or a Diva it has an accurate DSC chart representation. That's not true when I fudge the WB with a 32K or 56K preset.

October 18th, 2009, 05:02 PM	#22
Leonard Levy Trustee Join Date: Feb 2006 Location: USA Posts: 1,684	Can you pull still off the video?

October 19th, 2009, 04:04 AM	#24
Bob Grant Trustee Join Date: Nov 2005 Location: Sydney Australia Posts: 1,570	It seems to me that relying on a single number to describe white light is a bit limiting. Color temperature refers to a black body. Even a tungsten lamp is not exactly a black body and discharge lamps are even further removed from creating light the same as a black body. I had posted about another oddity with the EX1 some time ago. I bought a set of Kodak grey cards and the CT as reported by my EX1 is significantly different between the white and grey side of the cards. That was under fluro lighting. Serena from here if I recall correctly checked the same cards under daylight and got a match between the white and grey sides of her cards. What I suspect is going on is this. We know that Sony extended the range of the red sensor's filter. What we don't know is if they changed the green and blue as well but it's likely this camera is a bit different to any other camera. I know that some fluro lights designed for video use are made to match the RGB sensors in a video camera. If Sony has changed the colorimetry of the sensors it's not too hard to see how these lights will produce a different outcome with a camera with different colorimetry. I've used so called tungsten fluro lamps from Osram and to the eye they are not quite the same as tungsten lights yet they look the same to a video camera. I'm wondering if this is what the EX cameras are seeing as well. I wish I had the instruments to measure any of this.