Why compress 0-255 media to 16-235?

Rich Parry wrote on 8/19/2018, 3:38 PM

I’m a passionate self-taught amateur photographer with 300 videos on Vimeo, all rendered with Vegas over the past 10+ years. I’ve adhered to the broadcast safe standard compressing my media to 16-235 for both my still images and video using either “Color Curves” or "Levels" converting “Computer RGB to Studio RGB”.  I’ve read dozens of posts here on the subject and I still don’t understand why one should stay within the 16-235 color space.

To prove myself right or wrong, I made a 30 second test video switching between 0-255 media and 16-255 media. My conclusion is that it is a mistake to compress media from 0-255 to 16-235 since it results in the loss of deep blacks and bright whites.

https://vimeo.com/285720199

In case it matters, my media will never be broadcast on TV, heck, I’m lucky if I can get my wife to watch it on her cell phone. Perhaps there was a time when 16-235 was the standard, but I don’t think that is still the case. I hope I am wrong, I really don’t want to post process tens of thousands of video clips in 300 videos.

Comments welcome,

Rich

CPU Intel i9-13900K Raptor Lake

Heat Sink Noctua  NH-D15 chromas, Black

MB ASUS ProArt Z790 Creator WiFi

OS Drive Samsung 990 PRO  NVME M.2 SSD 1TB

Data Drive Samsung 870 EVO SATA 4TB

Backup Drive Samsung 870 EVO SATA 4TB

RAM Corsair Vengeance DDR5 64GB

GPU ASUS NVDIA GeForce GTX 1080 Ti

Case Fractal Torrent Black E-ATX

PSU Corsair HX1000i 80 Plus Platinum

OS MicroSoft Windows 11 Pro

Rich in San Diego, CA

Comments

Kinvermark wrote on 8/19/2018, 4:40 PM

This is one of those "chase your tail" endless topics that I am reluctant to engage in because it never ends well. :)

I will point out two things , however:

1) there is NO WAY you should see a

loss of deep blacks and bright whites.

You are only REDEFINING where black and white are numerically, not actually. This means your levels wrangling is wrong.

2) YouTube expects 16-235. TV's expect 16-235. OS & browser Media players often expect 16-235.

 

 

Rainer wrote on 8/19/2018, 5:33 PM

It's more complicated than that. Cameras deliver different levels, put some native GoPro on a timeline with Panasonic see what I mean. YouTube doesn't expect anything. What happens after that depends on your browser and graphics settings. Although you might upload and see 0-255, you mostly can't control what the end user sees. Uploading 16-235 probably has the best chance of delivering the result you want for most viewers. OTOH, many (most?) viewers viewing on their mobile devices are accustomed to black blacks, clipped whites, and may prefer high contrast, or your footage anyway may not have max levels of black and whites. So if it's not for TV broadcast, it may really not matter either way.

Kinvermark wrote on 8/19/2018, 6:09 PM

See, told you it wouldn't end well. There is never any consensus any more.

So, just do whatever you need to do to MAKE IT LOOK GOOD. No crushed blacks or blown out whites (unless that's what you want). FWIW, I use 16-235 video and 0-255 photos with a levels adjustment. Then they all get cc'd using a decent monitor and vegas scopes setup to 16 235. Everything looks great. 😀

Red Prince wrote on 8/19/2018, 8:11 PM

Because the standard was created by video engineers, not by computer engineers. Analog video used a limited range to allow squeezing metadata in the video stream, and that is how video engineers were used to thinking.

Meanwhile, computer engineers would never come up with such a concoction because you can embed as much metadata into a digital stream as you need.

As someone who worked with film and not video for most of my life, I really wish the engineers would just lower their heads in shame and changed the standard as they are coming with newer and more effective data compression methods.

He who knows does not speak; he who speaks does not know.
                    — Lao Tze in Tao Te Ching

Can you imagine the silence if everyone only said what he knows?
                    — Karel Čapek (The guy who gave us the word “robot” in R.U.R.)

Tim L wrote on 8/19/2018, 9:05 PM

@Rich Parry,
Just to give you a data point: On my uncalibrated, consumer LCD monitor, using Google Chrome to view your video, I see this...

On your 0-255 images, all of the lower end blacks mush together.  In the bottom row with boxes for 2, 4, 6, 8, etc, everything looks solid black until I can faintly make out the box at 18, while 20, 22, 24 are easily visible.  Likewise for the white end of the spectrum: I don't see any boxes in 255..243.

In the lower middle pane with color swatches, the left two black boxes have become one, and the top two white boxes likewise have become one.

On the 16-235 images, all of these boxes are clearly visible, except the 2 box on the black end is not noticeable, and the 254 box on the white likewise is not apparent (on my monitor).  The left two black swatches and the rightmost two white swatches in that middle pane are clearly separate colors.

I am just a hobbyist, with an old, inexpensive Samsung monitor.  But from my point of view, if you send 0..255 video to Vimeo, and I view it on my setup, everything below 16 or 17 or so becomes solid, featureless black, and likewise on the white end.

I would be curious what others see.
 

Tim L wrote on 8/19/2018, 9:30 PM

Also, I believe the reason that black starts at 16 (i.e. 7.5 IRE) and 0..15 are not allowed is that dipping down to 0.000v in a composite video signal (tube tv signal) is what marked the end of a display line and triggered the electron beam to return to the left edge of the tube to begin displaying the next raster line.  Sending video output level 0 (0.000v) for a certain time is kind of like pressing the carriage return on a typewriter -- at least in the old days...

But surely there are more knowledgeable pro's here that can speak to this more authoritatively than I can.

Former user wrote on 8/19/2018, 9:39 PM

Tim, although I am not an engineer, the 0 has nothing to do with end of line. In Japan and some PAL countries, 0 ire is the normal black level. It had more to do with transmitters and what a transmitter signal could handle without adding noise. The range from 0 ire to 100 ire (white) is 1 volt. It was basically set as an arbitrary (sort of) black level to minimize some problems in broadcasting. Since then, American Market TVs are calibrated to show 7.5 ire as the black on a TV. That is why if you have 0 black and try to watch on an NTSC calibrated TV, it may look muddled. Unfortunately, this standard will be with us for a long time due to FCC regulations of backward compatability.

Former user wrote on 8/19/2018, 10:37 PM

 

I am just a hobbyist, with an old, inexpensive Samsung monitor.  But from my point of view, if you send 0..255 video to Vimeo, and I view it on my setup, everything below 16 or 17 or so becomes solid, featureless black, and likewise on the white end.

I would be curious what others see.
 

I checked the rgb luminance of the lower row of boxes. The luminance data matches the numbers on the boxes exactly with the limited colour (16-235) version, but with full range version, luminance = 0 for boxes lower than 16, and all the white boxes(243 - 254) are 255.

So that is a really good lesson as to why you use limited colour on the internet. I think what he doesn't understand is that 16 - 235 is expanded out to 0 - 255 by the video player so you do get true black and true white, & all information below 15 & above 235 is lost.

The video it'self still contains the original 0 - 255 levels so I am guessing that is why poster doesn't see the problem, the player must not be decompressing the levels when playing on his monitor?? I dont' know. It seems strange

Musicvid wrote on 8/20/2018, 3:53 AM

Rich, this lighthearted explanation has helped a few people wrap their brains around the conundrum.

I certainly don't plan to write much more about it.

https://www.vegascreativesoftware.info/us/forum/pc-to-tv-levels-a-comedy-of-errors--107325/