Full range & Legal Range for video in adobe Premire question

[38mike]

On the Atomos Ninja V recorder you can select Legal range or turn it off and get full range signal.

the question is can PR read full Range signal when inporting footage? Full Range on wave form would be 0 to 1095 = in 10bit

Legal range would be 64 to 1000 = in 10bit ( broadcast safe.

I herd from youtube channels that its best to keep video footage in Legal range becuse PR only reads video footage as broadcast safe or legal range.

is this correct?

Please Help! 

[R Neil Haugen]

SDR video ... standard dynamic range video media, or "Rec.709" ... has standards set for it. And Premiere Pro follows those standards.

First, understand that the entire legal/limited 'thing' is about how the data is encoded to the file, not the number of values in the file. A common misconception.

YUV (technically Y/CbCr) media, which is nearly every camera out there, is supposed to be encoded as 'limited' or 'legal' range, 16-235. It will be displayed by any properly set monitor as 0-255.

RGB media ... typically 12-bit or higher, and most often produced in only very expensive cameras ... is set by standards to be encoded as 0-255, and displayed 0-255.

Setting that recorder to 'full' for SDR media doesn't get you any additional values for SDR/Rec.709 clips. It just makes a mess out of using them in an NLE. And maybe on many monitors/screens also.

And for most log-encoded and all HDR media, there isn't any limited/legal/full issue to begin with.

Neil

[38mike]

neal, i should keep my prores HQ 422 10 bit on Legal range? I can set it in the fujifilm camera or the NINJA V-  atomos. i am shooting fuji log

[38mike]

In camera I can set it to full or legal range

[R Neil Haugen]

Legal. Period. Doesn't change image quality, just fits all editing apps better.

Neil

[rgr2]

Virtually everything reads both limited and full range. Most current equipment records in full range. Premiere reads full range, but does it quite lamely. Vegas reads without a problem.

[R Neil Haugen]

"Most current equipment records in full range."

Certainly not my experience. Other than the Sony prosumer A7 rigs, and I think some Panny's offer the option also.

What gear do you use? I run both a BMPCC4K, and an Ursa Mini Pro 4.6k G2. Neither do ProRes in full. Not even an option. And of the files I get in to work and test from others, I've never had this.

Except for some of those running A7s, again. And a couple other cameras. 

But even most of the A7 media I've worked with was recorded in legal, as expected.

Because again, you do not get one tiny bit more data in a full level encoding ... and most displays still assume YUV Rec.709 to be legal ... so why create the hassle for absolutely no gain?

[38mike]

Neal my post is old

can you delete it. I am no longer asking questions on this post

thanks

[rgr2]

Sports cameras, phones, cameras -- most record in FR. Even my old DVD burner recorded in the range of 16-255. If you say otherwise, you haven't looked deeper into the file. The declaration that they are limited color range is often wrong, if it exists at all.

[R Neil Haugen]

I've been doing this around pro colorists where dealing with the full details of every file is expected as a routine thing.

So I have long experience in exact color management, at a detailed level most say editors get totally bored with. Including the things like differences in 12 bit RGB versus 10 bit YUV. The variances between color space and dynamic range changes handled separately or together. Wide gamut color correction versus output gamut color correction. When either is appropriate. And a bazillion other things.

Like sitting through long presentations by color scientists. And naturally I have several of the major resource and text books.

Because the people I create tutorials for live by producing outputs that pass QC machines first time every time. They have to. Exact minutia on files on on import is critical.

I've even read the full Rec.709 standards. Not many "normal" people have, of course.

And no, most cameras do not record Rec.709 in full range. Especially used in professional work.

[rgr3]

You write a lot, but there is no valuable content. But when I asked about xvycc you know nothing. Perhaps most professional equipment uses limited range. But popular devices are full range

[Billionaire Watches]

Full Range: Uses the entire 0–255 (RGB) or 0–1023 (YUV) range — best for computer playback.
Legal Range: Uses 16–235 (RGB) or 64–940 (YUV) — standard for broadcast TV.

[rgr3]

To be precise, YUV and RGB use 0-255 (8 bit) and 0-1023 (10 bit) for full range. Limited range is 16-235/240 and 64-940 for RGB and YUV. However, RGB limited range is rare.

[R Neil Haugen]

You are mistaking the levels shown in the two bit ranges for data encoded.

Both 8 bit and 10 bit encode the same data to the file ... it's the encoding math that changes. Both in encoding to file in the recording device, and decoding in the displaying device. As has been demonstrated over and over by color scientists ad infinitum.

For instance, if an 8 bit file only starts with 219 levels, it will have very obvious banding right off the bat. Very obvious banding.

Again, this has been shown in presentations on levels versus bit depth discussions.

And as soon as you touch any tonal (luminance) or color (chrominance) tool, it gets real nasty. 

So if your assumption was correct, an uncorrected 8-bit file wouldn't really ever be useful professionally, period. And you certainly couldn't do anything to the file.

Please ... don't confuse "levels" with "encoded data" ... they are mathematically different parts of the process.

And yea, this is so freaking non-obvious and counter-intuitive. But it's something that colorists have drilled into them off the bat ... if they are trained by experienced colorists.

[rgr2]

LOL. You have no idea what you're talking about.
235-16=239 -- in this range, data is encoded in limited range (8-bit encoding). More (240-16=224) is for chrominance.
And I didn't write anything about levels.

[R Neil Haugen]

Gee, apparently all those guys with the heavy credentials writing the books for SMPTE don't have a clue what they're talking about?

Darn, I've got several major and spendy reference books by these supposedly noted figures ... huh.

[rgr2]

You still need to understand what you are reading. Book knowledge alone is not enough.

[R Neil Haugen]

Your comments are counter to every color scientist, SMPTE color speaker, colorists training and presentations, in other words, every expert I've ever been around, in the technical details.

It isn't just looking at a book. If I've not been clear enough, I've been through this in presentations, in many lengthy in person discussions, as a both watcher of and panel participant in online presentations, so I'm basing my comments on the wealth of technical knowledge of the top people in video color in the world.

Encoding process is not file content data. Though people often conflate the two.

If they actually only encoded 218 levels, as the file data, the file would never be professionally usable. Period