Why does NTSC video have separate reference blank and reference black levels, and how does this affect format conversion?

Question

Re-reading the Wikipedia page on the IRE unit I noticed something new to me: reference black in PAL and SECAM is 0 IRE, rather than 7.5 IRE as it is in NTSC. This brings up a whole host of questions, the primary one being, why does NTSC set a reference black higher than reference blank?

Some futher questions:

Does RGB with separate sync using NTSC sync frequencies also set reference black to 7.5 IRE (i.e., at slightly higher than 0 V) on the three colour channels? Or is reference black 0 IRE there because it's RGB? What about if it's RGB using sync on green? Whichever way it is, is this something specified by NTSC or did the manufacturers, lacking a specification, just settle on one or the other? (I suppose with typical pro video monitors you'd just adjust them with a test signal anyway, so they might not really have a spec for that, but given that such monitors exist, there must have been video source equipment that had to make this choice.)

The Wikipedia PLUGE page says that the SMPTE colour bars include a PLUGE section of super-black (0 IRE, I am assuming), normal black (7.5 IRE?) and near-black are included for adjusting the black level and contrast, but then goes on to say that these are for NTSC, PAL and SECAM. I take it this means that you need to know the standard your monitor is using and adjust these brightness of differently based on this? That is, if you're adjusting for an NTSC signal, black should not be brighter than super-black, but if you're adjusting for PAL or SECAM black should be slightly brighter than super-black?

How do video conversions between NTSC and PAL deal with this difference in black level?

Do video generation systems (such as old microcomputers and video game consoles) generally follow this standard as well?

I'd also appreicate hearing about any other implications of having black level higher than blank level in NTSC.

Answer 1

Setup, or pedestal, was a 'feature' that helped engineers align the raster on the monitor: it made an active video extent visible by turning up the brightness. When's the last time you did that? Never. So the later PAL and SECAM standards didn't use it. PLUGE is used instead.

Setup requires that the video be 'squished' to allow for the 7.5 IRE added to it and still fit into 100 IRE. So NTSC-M black to white range is reduced compared to no-setup (e.g, NTSC-J) video.

Do videogame consoles and other early systems use setup? Maybe, maybe not. If they're intended for real video production for NTSC-M they do, otherwise their output would look too dark and lose detail in the lowlights. (Video Toaster got it right.)

Do conversions from / to PAL and SECAM to / from NTSC have to deal with setup? Yes. That's the easy problem to solve; frame rate conversion is the bigger challenge.

Does RGB video use setup? It can, but generally doesn’t (not for VGA anyway.) It is used for S-video and sometimes on YUV component (480i / 480p).

What about video production now? With the advent of digital video based on BT.656, things got better. BT.656 sets out same digital range for PAL and NTSC video, regardless of setup: 16-240 for Cb/Cr, 16-235 for Y for 8-bit video. In other words, setup just isn’t a thing in digital video.

Instead, in modern production systems that deal with the NTSC-M legacy format, setup usually gets dealt with by the video I/O hardware that converts analog video from or to BT.656 digital. The flow is this:

• NTSC to digital: setup is subtracted out, then levels normalized BT.656
• digital to NTSC: levels re-normalized to NTSC, then the setup added back in

As long as this is done correctly, you will not see much of a difference between setup and no-setup in the presentation. If it's done incorrectly, the picture will either be too dark (no-setup video shown on a monitor that expects 7.5 IRE black) or washed out (setup video shown on a monitor that expects 0 IRE black.)