White Clip

[Eugene Kosarovich]

I'm still trying to tweak some settings, and one I'm not sure of is white clip.

The Sony provided scene files have have it set for 100 IRE for DVNORMAL, 105 IRE for DVCINE, and 109.5 IRE (max) for HDNORMAL and HDCINE.

Since technically HD should still be in the 0 to 100 IRE range, I take it those are maxed since they expect people to be doing more grading in HD post?

And with the Cine curves, one shouldn't be ending up over 100 IRE much anyway.

But what about for the "normal video looks"?

Is there any difference in having the camera do the white clip versus just legalizing in post? Is the in-camera white clip just a limiter or is it something more that would look better than limiting in post? Is using white clip cleaner than if the camera just hit its own max?

Does the white clip level affect anything else? Like how DCC functions (or at what levels)?

Thanks.

[David Heath]

Difficult to generalise, but the likelihood is that on transmission, material will just get a hard clip at 100IRE, end of story.

Hence, set the camera clipper to anything higher, and you may well see detail in the highlights whilst shooting/editing etc, but it will be plain white for the viewer.

If you are CERTAIN that material shot will undergo proper grading, it may make sense to clip at higher than 100%, to get a little extra headroom. If there is any chance it won't get a full grade, play safe and set the clipper to 100 - that's the only way you can guarantee the viewer will see the same as in the edit suite.

[Simon Wyndham]

It certainly makes sense to turn the white clip off if the footage will be used for broadcast and the web. 100ire white on the web looks like a very light grey.

Like how DCC functions (or at what levels)?

No, white clip or level has no effect on this. Although I wouldn't recommend using DCC for anything. It is a common misconception that it helps highlights, but in fact it doesn't always. If it deems the highlights in your picture not to be important it will let them hard clip. You don't want the camera taking that sort of decision for you.

[Daniel Epstein]

In the analog days the Broadcast level of 100 Ire is not usually where Sony specced the white clip on many cameras. Some of the cameras used 110 ire for the white clip. Many Tv Stations set their equipment including cameras through transmitters so signals wouldn't go above 100 for legal broadcast. Most video signals could go much higher than 100 before causing problems except in broadcasting. This means some detail in white objects like clouds could be higher than 100. Dynamic Contrast Control was a specific kind of knee control which allowed allowed brighter looking highlights to not be clipped. The actual sensors like CCD's or CMOS can handle highlights much higher than than the contrast range one volt of video can resolve so compression circuits squeeze the signal into the video standard based on settings. The problem with DCC is it would often not be consistent depending on the scene and if misadjusted change the highlights in ways that looked worse than not using it. Setting the white clips lower than what Sony wanted also interfered with the DCC circuits accuracy.

[Simon Wyndham]

The best way to think of DCC is to think of it as an auto focus function for the knee. Not good.

[David Heath]

Originally Posted by Daniel Epstein

In the analog days the Broadcast level of 100 Ire is not usually where Sony specced the white clip on many cameras. Some of the cameras used 110 ire for the white clip. Many Tv Stations set their equipment including cameras through transmitters so signals wouldn't go above 100 for legal broadcast.

Trying to throw my mind back a long way I believe there was a fundamental difference between the full spec for PAL and NTSC, over and beyond the most basic ones of line structure and the way the colour is carried.

In each system, black and white are defined as 0 and 100 IRE, but I seem to remember that the official NTSC spec allows for clipping to be set higher than 100, and transmitters to be specced to allow for these higher luminance levels. That wasn't the case in PAL - peak white was defined as 100, and this is where clipping should be set to adhere to the PAL standard. Anything higher will be hard clipped off at the transmitter.

As far as the Sony and other manufacturers spec go, then most cameras then were first built as NTSC (with manuals written accordingly), with a PAL version following later with manual amendments. Consequently, the instructions were written first and foremost for NTSC, and that included clip settings. They should have been altered for the PAL version of the manual, but generally didn't, hence the lineups which allowed too high luminance levels all the way to the transmitter, and hence the tales of highlight detail being visible in the edit suite, but not on transmission.

[Greg Boston]

In NTSC, the video portion of the transmitted signal is amplitude modulated. If a carrier is modulated beyond 100 %, it will create spurious emissions on other frequencies... a big no-no as far as the FCC is concerned.

Therefore, the maximum white level is 90%, which modulates the carrier down to 10%. The minimum black levels are set up so that video black doesn't interfere with the horizontal and vertical sync pulses.

So 100 IRE is the reference for white which is only 90%. 110 IRE is the absolute max signal before clipping because that would correspond with 100 % modulation and you don't want to cross that line with your transmitter signal.

-gb-

[David Heath]

Originally Posted by Greg Boston

In NTSC, the video portion of the transmitted signal is amplitude modulated.

Not quite - analogue systems use vestigial sideband modulation. If amplitude modulation was used, the transmitted signal would have over twice the bandwidth of the video signal, and be inefficient in power usage. In the UK, the channel spacing is 8MHz, and transmitted video bandwidth is 5.5MHz, not allowing for sound. If AM was used, channel spacing would have to be greater. (See Amplitude modulation - Wikipedia, the free encyclopedia )

So 100 IRE is the reference for white which is only 90%. 110 IRE is the absolute max signal before clipping because that would correspond with 100 % modulation and you don't want to cross that line with your transmitter signal.

Yes, but it's a bit more complicated than that because of the colour subcarrier. In practice, colour subcarrier can go much higher than peak white, and much lower than black level - as is obvious if you look at colour bars on a waveform monitor. Practically, it's the highest voltage value that the subcarrier peaks can ever reach (in the 100% yellow bar) that restricts modulation level. At the other end of the scale (blue), then whilst blue subcarrier troughs will go well below black level, they will never go below sync level at 0 volts.

[Greg Boston]

Originally Posted by David Heath

Not quite - analogue systems use vestigial sideband modulation..

I was simplifying a bit. I did hold an FCC First Class Radiotelephone License a few moons ago. I was also an active ham radio operator for many years. Sideband transmission is still a 'form' of amplitude modulation, it just omits the inefficiency of transmitting the carrier which is re-inserted by the receiving end in sideband transmission. I used to listen to AM broadcast stations on my ham receiver which only had upper/lower sideband reception. You just get the audio tone difference of the transmitted carrier and your own local carrier insertion, but it's minimal when you are right on frequency.

-gb-

[Simon Denny]

Regarding the DCC switch on the f350 should I have this set to on when outdoors? More to the point what is the DCC doing. I have found some info in the subject but it's rather loose on it's use. My understanding is the DCC rolls of the top end in bright areas. Is this correct?

Should I use the DCC or leave it off?

Thanks

[Simon Wyndham]

My understanding is the DCC rolls of the top end in bright areas. Is this correct?

DCC is an auto knee. It adjust the knee point and I believe the slope(?) on a dynamic basis. The marketing blurb tells you that it will help you capture high contrast situations. In practise, although it sometimes works, like all auto functions it is flawed. In reality the DCC adjusts your knee to adapt to what it thinks is important. So if it deems your highlights to be unimportant you can end up with more blow outs than you would do normally!

The DCC function ha sa long history of messing things up. I like to know what the knee is doing from shot to shot so I never ever use the DCC function. Certainly if you are doing a dramatic style shoot I would never recommend using DCC.

But as per all things, try it and see. Flick it on and off to see what the different result is. Sometimes it will protect your highlights, and other times it will blow them out more for you.

[Uli Mors]

"sometimes it does... sometimes it blows out" - is EXACTLY what I encountered after dozens of DCC experiments with the XDCAM HD line.

a) the f3xx series is very sensitive to overexposure - if you overexpose, DCC makes everything even worse...

b) knee/auto knee (+DCC) in the f3xx has no "true skin" function - resulting in heavy overblown skintones (Yellow-reddish blown out faces) when overexposing / going into knee, again, DCC can make everything worse

c) I see advantages when filming shots with sky + clouds. Simply try switching DCC on /off and see how the contrast in the sky/clouds change.

My recipe: Better avoid DCC with faces (at least at the PDW-F3xx series) - you have better manual control over everything.

Uli