Ajustes elementales con cartas de ajuste
-
honorable
Re: Ajustes elementales con cartas de ajuste
Este es el texto del pdf escrito por poynton hace un año(es decir muy reciente)
Studio HD technicians and home theater calibrators are familiar with
10-bit HD coding: Reference black is placed at interface code 64 and
reference white is placed at 10-bit interface code 940. (The comparable
8-bit codes are 16 and 235.) These communities are less familiar
with the origins of footroom and headroom, and are generally unfamiliar
with proper treatment of codewords that lie in these regions.
I will address these topics in this column.
Codes 64 and 940 are meaningful across a digital video interface;
however, for signal processing it is much more convenient to declare
reference black and white to be zero and unity respectively. Call those
levels 0 units and 100 units if you like. In standard HD, headroom
extends to about 109 units; the corresponding luminance is about
1.23 times reference white.
Reference
and
peak
are different! Many
ITU-R, EBU, and SMPTE standards get this wrong, mistakenly using
peak white
when
reference white
is meant. Codes 0–3 and 1020–1023
are prohibited across an HD-SDI interface.
The original digital studio video standard was CCIR Rec. 601, established
in 1984. Analog studio video signals drifted somewhat; to introduce
digital video required accommodation of analog signals having
imperfect reference levels. Footroom and headroom were necessary.
That reason has now vanished. However, several good reasons for
footroom and headroom remain
When presented with an input signal containing high frequency
content, any practical filter – whether analog or digital, or lowpass,
bandpass, or highpass – necessarily involves some degree of undershoot
and/or overshoot. Premature clipping of undershoots and overshoots
is detrimental to image quality: Clipping should be deferred to
the last possible point in the signal chain. Footroom and headroom
accommodate undershoots and overshoots. Historically, it has been
Unfortunately, in the consumer domain, manufacturers are motivated
to compromise the headroom region for two reasons. First,
consider a display with maximum luminance of 250 nt. By properly
following the gamma curve all the way up to 109 units (historically,
“IRE”), peak luminance is 250 nt, but reference white luminance is
200 nt. If the display follows the gamma curve up to 100 units then
clips, the manufacturer can claim reference white at 250 nt!
Consumers tend to think bigger numbers are better – but the clipped
picture will suffer. A home theatre calibrator, on assessing the greyscale
response, will be savvy to this trick, and will adjust the display to
follow the gamma curve all the way up to peak. Second, a manufacturer
may be motivated to push the curve up in the midscale, then
roll-off the higher regions in an effort to deliver higher average luminance.
The well-equipped calibrator will similarly reverse the trick
generally agreed – if not properly documented – that signals shouldn't
dwell in the footroom or headroom region for longer than half a dozen
samples. I'll bring this assumption into question below.
A relatively recent reason for footroom is to convey the negative
excursions of camera noise. All sensors generate noise, even around
black. When a camera is sensing true optical black (e.g., capped), it is
sensible to set the average signal value to reference black. However,
noise has excursions above and below that level. If the negative-going
excursions are clipped, then the noise is said to be “rectified”: the
average value of the noise then rises above reference black
It is
important to defer the rectification to the latest possible point in the
signal chain.
A final reason for footroom is that it allows coding of the blackerthan-
black ( -2% PLUGE signal element commonly used to set black
level in studio displays.
A camera engineer typically aligns an HD camera to produce 100
units for a near-perfect white reflector in the scene. However, when
the camera is turned over to the cinematographer, he or she may wish
to convey specular reflections or light sources in the scene, and he or
she may therefore reset exposure to place the white card well below
100 units. Light sources and speculars may lie anywhere above the
cinematographer's white reference, and the speculars and sources
won't let up until they clip! The headroom region ends up carrying
these elements. Typically these elements have momentary excursions,
but they may well be sustained across more than half a dozen samples.
All of the reasons that I have mentioned are arguments against clipping
anything in the footroom and headroom regions! Indeed, to clip
is bound to introduce some degree of visual artifacts. Movie studios
use commercial “QC” outfits to review commercial content prior to
mastering. Many QC outfits are motivated to report “violations.” Most
post houses take the easy way out and clip to “legal” before shipping
content out to the QC houses. The potential visual quality of such
material is compromised, and in the long term, the QC houses should
be educated. Once the QC houses are well informed, the post houses
will stop clipping.
.
Permisos de publicación
- No puedes crear nuevos temas
- No puedes responder temas
- No puedes subir archivos adjuntos
- No puedes editar tus mensajes
-
Reglas del foro