IEEE Consumer Electronics Magazine - May 2018 - 39

ICtCp color space in a framework of an HDR and WCG visual
data compression was previously performed [20]. The theoretical performance of the reshaper was assessed in [21]. However,
the aforementioned analyses are restricted to the use of objective metrics only.
In this article, subjective experiments evaluating the performance of ICtCp are presented, and their results are analyzed
and discussed in detail. Furthermore, a reshaper, adjusting
pixel intensity values to optimize compression algorithms, is
also evaluated, and its performance is analyzed. Our main aim
is to assess the two following assertions: 1) ICtCp signal quality will be preferred or equally perceived over YCbCr signals,
and 2) reshaped signals, enabling 10% of bit rate savings, will
be perceived as having a similar quality as compressed signals
using current algorithms.

BAckgrOuNd

have been proposed and described within the Society of Motion
Picture and Television Engineers [25], [26]. Among others, the
British Broadcasting Corporation hybrid log-gamma (HLG),
the Dolby Vision end-to-end service, and the Samsung solution
for dynamic metadata have been proposed. In the scope of this
article, we consider the improvements of data representation
using either a new color space or a mapping function (reshaper)
enabling bit rate saving.

ColoR SpaCeS
The use of each particular color space leads to specific compression artifacts. For example, significant color artifacts
result from the use of nonconstant luminance (NCL) YCbCr
representation as color impairments spread to the luminance
channel, where distortions are more noticeable. Using the
constant luminance (CL) Y'CbCr representation, prevents the
prior mentioned color artifacts; however, it requires more
computation power due to higher complexity.

HDR anD WCG
YCbCr Color SpaCe
The ITU standardized the YCbCr color space within the first
version of the ITU-R BT.601 in 1982 [16]. YCbCr is composed
of an NCL channel Y and the blue- and red-difference chroma
components Cb and Cr, respectively. Several variants derive
from YCbCr to transcend its limitations. A nonlinear filter,
such as a gamma correction, can be applied to the luminance
channel to reduce quantization errors perception. The resulting
representation is commonly referred to as Y'CbCr. When nonlinear filtering is also applied to the chrominance components,
the obtained color space is referred to as Y'Cb'Cr'. The CL
Y'CbCr (or Yc'CbcCrc) addresses the issue of the correlation
between luma and chroma components.

0.9
0.8

ITU-R BT.2020

0.7
ITU-R BT.709

0.6
0.5
(y )

HDR and WCG imaging allow a device to capture and represent a wider color gamut containing a larger amount of reproducible colors when compared to SDR. This fact allows users
to be provided with a more realistic sensation and more accurate visual information.
In general, the HDR technology aims at a more realistic
representation of visual information. It copes better with the
HVS ability to generally perceive approximately ten orders of
magnitude of luminance [7] and accommodates roughly five
orders of magnitude simultaneously [24]. The improvement of
the reality sensation is mainly due to a more accurate rendering
of the nuances in colors and luminance present in a scene. For
instance, the extension of the dynamic range permits having
more accurate highlights and shadows. HDR's new characteristics include higher peak luminance, lower minimum luminance, greater contrast range, and improved precision, minimizing
quantization errors that are beyond the capabilities of existing standards.
However, a color gamut is a subset of colors representing the
proportion of the visible spectrum that a rendering and/or visualization system can reproduce. It is usually represented by a
triangle in the International Commission on Illumination (CIE)
1931 XYZ color space xy chromaticity diagram. WCGs, such
as BT.2020 [11], are required to accurately represent visual
information since they allow more colors to be rendered than
standard color gamuts, such as BT.709 [17]. Figure 1 shows
the CIE 1931 XYZ color space xy chromaticity diagram and
the representation of the BT.709 and BT.2020 color gamuts.
HDR and/or WCG representation of visual information
requires more storage when compared to SDR. Therefore, an
efficient compression algorithm of HDR and WCG content is
needed. Current standards for HDR image and video compression are Joint Photographic Experts Group (JPEG) 2000, JPEG
XR, JPEG XT [1], and high efficiency video coding format
range extension [18]. Nevertheless, the compression process is
not the only solution to reduce the bitstream bandwidth and to
improve HDR and WCG content distribution. Several solutions

0.4
0.3
0.2
0.1
0.0
0.0

0.1

0.2

0.3

0.4
(x )

0.5

0.6

0.7

0.8

FIgure 1. a representation of the BT.709 and BT.2020 color gamuts.

may 2018

IEEE Consumer Electronics Magazine

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - May 2018