Zebra 107=Spot 100%=Last 3 HG Bars Gone at DVinfo.net

Michael H. Stevens · February 16th, 2008, 03:48 PM

Did some more tests today with scopes. Very bright weather using CINE1 and CINE4 based profiles.

What I found was that (at least on my camera) the histogram is calibrated wrong; white clipping occurs with three bars still to go! IE maximum exposure occurs when you have eliminated the furthest right three bars of the histogram. If you use the histogram to set maximum exposure you will be severely clipped.

I confirm, as I remember seeing here before, Zebra point 107 (maximum exposure) equates to a spot meter reading of 100%.

Now here is a question for "The experts": Is the Zebra 107 or the Spot meter 100 the IRE value? Or maybe they both are and Zebra 107 really means Zebra 97-107 the Zebra 2 having the 10% spread of Zebra 1 that we thought it did not?

Mike

Piotr Wozniacki · February 16th, 2008, 04:14 PM

Mike,

I cannot confirm your findings.

1. Indeed, the EX1 histogram tends to be narrow, but not more so that Vegas histogram when you load the clip into an 8bit project. However, after you convert it to Computer RGB, the Vegas histogram is indeed much wider, spanning more evenly from 0 to 255 (and the luminance can even shows the range of -10 to 110 IRE.

2. No, Zebra 2 is only displayed in areas above 100% (as apposed to Zebra 1, which shows in the +/- 10% range of the level set).

Michael H. Stevens · February 16th, 2008, 08:17 PM

Quote:

Originally Posted by Piotr Wozniacki

Mike,

I cannot confirm your findings.

1. Indeed, the EX1 histogram tends to be narrow, but not more so that Vegas histogram when you load the clip into an 8bit project. However, after you convert it to Computer RGB, the Vegas histogram is indeed much wider, spanning more evenly from 0 to 255 (and the luminance can even shows the range of -10 to 110 IRE.

2. No, Zebra 2 is only displayed in areas above 100% (as apposed to Zebra 1, which shows in the +/- 10% range of the level set).

How do you convert to computer RGB?

Warren Kawamoto · February 16th, 2008, 08:43 PM

http://www.dvinfo.net/conf/showthread.php?t=114877

I was confused too, the first time I tried to use the histogram.

Piotr Wozniacki · February 17th, 2008, 02:43 AM

Quote:

Originally Posted by Michael H. Stevens

How do you convert to computer RGB?

Michael,

I have answered your question here:

http://www.dvinfo.net/conf/showpost....88&postcount=7

Piotr Wozniacki · February 17th, 2008, 03:51 AM

Michael,

As I wrote in the thread Warren linked you to, in the V1's histogram there is a threshold vertical line displayed at the value yoy have the zebra set to. If I remember correctly, it'still some distance to the left of the RH histogram end when the zebra is set to 100% - meaning, the RH histogram end represents super-whites of up to 108 IRE, and NOT 100 IRE.

Whereas Zebra 2 on the EX1 will show at 100% already (BTW I don't understand your "Zebra at 107" term; you cannot tell what Zebra 2 is showing with this precision; it looks the same for 100, 101, ... 108).

Perhaps this is the reason of your confusion?

Bob Grant · February 17th, 2008, 04:59 AM

I just had another play around with the camera to test something else and I think I can see why it's so confusing. Yes, it can be very hard to tell when you're into clipping. I could get the camera to go into Too Bright alarm and it still wasn't obvious from the histogram that I was clipping. Change what the camera was looking at so that most of the frame would go into clipping around the same time and the graph in the histogram changed dramatically, going into clipping was very obvious.

Changing the gamma also has quite an effect as well.

Randy Strome · February 17th, 2008, 08:01 AM

With different PP's the brighness cutoff is changing and the histogram is subsequently ending at a differnt max brightnesses. This is not true "off the chart clipping", but will produce very white blocking which still looks like clipping.

Michael H. Stevens · February 17th, 2008, 10:44 AM

Quote:

Originally Posted by Randy Strome

With different PP's the brighness cutoff is changing and the histogram is subsequently ending at a differnt max brightnesses. This is not true "off the chart clipping", but will produce very white blocking which still looks like clipping.

Randy: You may have hit on something here that I am confusing. What is the difference between white blocking where all detail in the whites is lost and clipping? I'm treating them as the same thing.

Randy Strome · February 17th, 2008, 11:05 AM

Quote:

Originally Posted by Michael H. Stevens

Randy: You may have hit on something here that I am confusing. What is the difference between white blocking where all detail in the whites is lost and clipping? I'm treating them as the same thing.

I am no expert on this, but can tell you what I am seeing. When, for (an extreme) instance PP2 is selected, you will never record over 100% (out of a possible 107%-which may also be lending to the confusion). That PP simply disallows brighter information to be recorded, but as you will note as you pogressively open your iris, that more and more pixels are registering at the same (the brightest value) in the case of PP2 that is 100 (which will fall short of the right side of the histogram). That will cause white blocking in your image that looks just like clipping.

The same is practically true with certain cine settings. When you are at the high end of proper exposure, so as that you are just starting to see zebras with the fixed (above 100%) zebra, and you proced to open the iris, there will be a barrier where the histogram will stack at a certain value (for instance 104%) but not reach the far right.

February 16th, 2008, 03:48 PM	#1
Michael H. Stevens Major Player Join Date: Sep 2007 Location: Palm Desert, California Posts: 311	Zebra 107=Spot 100%=Last 3 HG Bars Gone Did some more tests today with scopes. Very bright weather using CINE1 and CINE4 based profiles. What I found was that (at least on my camera) the histogram is calibrated wrong; white clipping occurs with three bars still to go! IE maximum exposure occurs when you have eliminated the furthest right three bars of the histogram. If you use the histogram to set maximum exposure you will be severely clipped. I confirm, as I remember seeing here before, Zebra point 107 (maximum exposure) equates to a spot meter reading of 100%. Now here is a question for "The experts": Is the Zebra 107 or the Spot meter 100 the IRE value? Or maybe they both are and Zebra 107 really means Zebra 97-107 the Zebra 2 having the 10% spread of Zebra 1 that we thought it did not? Mike

February 16th, 2008, 04:14 PM	#2
Piotr Wozniacki Inner Circle Join Date: Aug 2006 Location: Poland Posts: 4,086	Mike, I cannot confirm your findings. 1. Indeed, the EX1 histogram tends to be narrow, but not more so that Vegas histogram when you load the clip into an 8bit project. However, after you convert it to Computer RGB, the Vegas histogram is indeed much wider, spanning more evenly from 0 to 255 (and the luminance can even shows the range of -10 to 110 IRE. 2. No, Zebra 2 is only displayed in areas above 100% (as apposed to Zebra 1, which shows in the +/- 10% range of the level set). __________________ Sony PXW-FS7 \| DaVinci Resolve Studio; Magix Vegas Pro; i7-5960X CPU; 64 GB RAM; 2x GTX 1080 8GB GPU; Decklink 4K Extreme 12G; 4x 3TB WD Black in RAID 0; 1TB M.2 NVMe cache drive

February 17th, 2008, 03:51 AM	#6
Piotr Wozniacki Inner Circle Join Date: Aug 2006 Location: Poland Posts: 4,086	Michael, As I wrote in the thread Warren linked you to, in the V1's histogram there is a threshold vertical line displayed at the value yoy have the zebra set to. If I remember correctly, it'still some distance to the left of the RH histogram end when the zebra is set to 100% - meaning, the RH histogram end represents super-whites of up to 108 IRE, and NOT 100 IRE. Whereas Zebra 2 on the EX1 will show at 100% already (BTW I don't understand your "Zebra at 107" term; you cannot tell what Zebra 2 is showing with this precision; it looks the same for 100, 101, ... 108). Perhaps this is the reason of your confusion? __________________ Sony PXW-FS7 \| DaVinci Resolve Studio; Magix Vegas Pro; i7-5960X CPU; 64 GB RAM; 2x GTX 1080 8GB GPU; Decklink 4K Extreme 12G; 4x 3TB WD Black in RAID 0; 1TB M.2 NVMe cache drive

February 16th, 2008, 08:43 PM	#4
Warren Kawamoto Trustee Join Date: Nov 2006 Location: Honolulu, HI Posts: 1,435	http://www.dvinfo.net/conf/showthread.php?t=114877 I was confused too, the first time I tried to use the histogram.

February 17th, 2008, 04:59 AM	#7
Bob Grant Trustee Join Date: Nov 2005 Location: Sydney Australia Posts: 1,570	I just had another play around with the camera to test something else and I think I can see why it's so confusing. Yes, it can be very hard to tell when you're into clipping. I could get the camera to go into Too Bright alarm and it still wasn't obvious from the histogram that I was clipping. Change what the camera was looking at so that most of the frame would go into clipping around the same time and the graph in the histogram changed dramatically, going into clipping was very obvious. Changing the gamma also has quite an effect as well.

February 17th, 2008, 08:01 AM	#8
Randy Strome Major Player Join Date: Jan 2008 Location: Haiku, HI Posts: 203 Images: 12	With different PP's the brighness cutoff is changing and the histogram is subsequently ending at a differnt max brightnesses. This is not true "off the chart clipping", but will produce very white blocking which still looks like clipping.