Posterization

Started 6 months ago | Discussions thread
Steve Bingham
Forum ProPosts: 20,557Gear list
Like?
Re: Raw Histogram
In reply to Jack Hogan, 6 months ago

Now that's an answer! . . . and no, I don't "see" any posterization. Between you and E J Martin my head hurts. I usually convert my raw files to 16 bit tiffs and without seriously pushing the image I rarely run into any posterization. (I know jpgs are only 8)

Now excuse me while I reread your explanation for the next 5 minutes.

Jack Hogan wrote:

Sergey_Green wrote: What bothers me the most is the limitation in producing smooth gradients regardless of the viewer. I have several of those frames from winter, where the moment I darken the sky (just barely) I start seeing the gradients breaking down. It's like there is no way around it.

A couple points of note:

1) In the absence of noise humans apparently perceive steps in gradients as smooth if the change in Luminance from step to step is less than about 1%. They may see them as discrete steps if the change in Luminance is greater than about 1% (Weber-Fechner).
2) In the presence of Noise the 1% threshold is raised significantly (see here for a very good demonstration by ejmartin)

Since Raw data is supposedly proportional to Luminance (unlike gamma corrected jpegs or TIFFs) it is easy to verify whether the information from the camera is posterized: just look at the NEF data and see whether the green channel values within the gradient jump by more than 1% at any point (that would be 1 level around DN 100, 2 around DN 200 and so on).

In your file this is the histogram of a two-bit wide vertical slice more or less in the center of the yellow rectangle to the left of the flower (if you squint you can see the thin selection):

Raw Histogam of thin 2 pixel wide by 2000 pixel high column. Green Channel is superposition of G1 and G2

Ignore Red and Blue which are pre-conditioned for WB. Green is actually the superposition of G1 and G2, so it represents a true 'single pixel' vertical gradient 2000 pixel high within the yellow box . The green histogram appears fully populated. Nowhere is more than one level skipped from ADU 116 to 530 (the value range in the box), excluding the very first pixel at 111. Therefore if you see posterization anywhere within the yellow box area it is clearly not the fault of your camera but it is being generated by downstream processing - with 8 bit conversions, stretching by profiling, video drivers and lower bit depth video panels typical guilty parties.

If you have a real 8+ bit panel with 12+ bit internal LUT, calibrate it properly by adjusting its Brightness and Color Temperature through the hardware knobs instead of letting the profiling software do it for you automatically through its typically 8-bit look up tables there is little else you can do other than ETTR letting undesirable highlights blow and possibly adding a touch of random noise in post to hide the steps.

Jack

-- hide signature --

Steve Bingham
www.dustylens.com
www.ghost-town-photography.com

 Steve Bingham's gear list:Steve Bingham's gear list
Nikon D300 Nikon D7100 Nikon D610 Nikon D5300 Nikon AF-S DX Nikkor 18-105mm f/3.5-5.6G ED VR +21 more
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
ThxNew
thxNew
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow