BLUE LIGHT REDUCING SOFTWARE APPLICATIONS FOR MOBILE PHONE SCREENS: MEASUREMENT OF SPECTRAL CHARACTERISTICS AND BIOLOGICAL PARAMETERS

Full text

RAP Conference Proceedings, vol. 4, pp. 220–224, 2019
ISSN 2466-4626 (online) | doi: 10.37392/RapProc.2019.45
WWW.RAP-PROCEEDINGS.ORG
BLUE LIGHT REDUCING SOFTWARE APPLICATIONS FOR MOBILE PHONE SCREENS:
MEASUREMENT OF SPECTRAL CHARACTERISTICS AND BIOLOGICAL PARAMETERS
S. Mitropoulos1, V. Tsiantos2, A. Americanos3, I. Sianoudis3, A. Skouroliakou1*
1Department of Biomedical Engineering, University of West Attica, Athens, Greece
2Department of Physics, International Hellenic University, Thessaloniki, Greece
3Department of Biomedical Sciences, University of West Attica, Athens, Greece
Abstract. The displays of the majority of electronic devices nowadays are illuminated by Light-Emitting Diodes
(LEDs) or Organic Light-Emitting Diodes (OLEDs). These types of light sources have certain advantages regarding
colour variety, contrast, resolution and the ability to construct thinner screens. Nevertheless, recent research raises
concern of possible negative biological impact of these display types on visual health and the circadian rhythm. The
biological basis of the concern lies in the emission spectra of the light sources. The white LEDs used as backlights in
LED screens have a characteristic emission spectrum with a peak at 450 nm and the Red-Green-Blue (RGB) OLED
emission spectrum has a blue peak. Both of them are very close to the 460nm where the melanopsin retina pigment
presents the maximum absorption. In order to reduce the blue light emission several techniques have been developed
including hardware adjustments, external filters and software applications that control the emission display
characteristics. This study aims to record the performance of several available software applications on different
mobile phone models. The spectral power distributions of the mobile phone screen were recorded by means of a
commercial radiospectrometer, without and with the use of the blue light reducing software application, for various
blue light filtering levels depending on the application. Several photometric and circadian parameters were
calculated from the available spectra such as circadian light input, photopic illuminance and melatonin suppression
index. The results of the study are the recordings of the respective differences in mobile screen output with and
without the use of the blue light reduction application, presented in terms of spectral power and biologically relevant
parameters. The analysis of the measuring procedure and the obtained results lead to an evaluation of the application
performance variation depending on the mobile phone type and a standardised measurement protocol in order to
have comparable results that could be used for blue light reducing software applications performance evaluation.
Keywords: Spectrophotometry, circadian parameters, blue light filtering
* kskourol@uniwa.gr
1. INTRODUCTION
Smartphones are part of our everyday life for
communication, social and professional networking
and leisure applications. Their use is well established
and will probably increase in the near future. The
visible light spectrum emitted by smartphone screens
is mainly determined by screen technology. In most
cases it is based on white LEDs whilst most
sophisticated smartphone models use OLED based
screens. In both cases the emitted light spectrum is
characterized by three bands in the red, green and blue
spectrum regions.
Apart from the visionary system light is also a
significant input for the circadian regulation system.
The human circadian clock organizes the timing of all
daily biological functions and is regulated by the
suprachiasmatic nuclei (SCN) in the hypothalamus.
The temporal pattern of light and dark on the retina
synchronizes the SCN to a 24-h period matching to the
earth's rotation rate on its axis [1]. The human visual
and circadian systems are characterized by different
mechanisms of light processing thus presenting
different sensitivities and responses to retinal light
exposure [2]. The visual system has a large dynamic
response range covering 9-10 orders of magnitude. The
response of the circadian system at optical radiation is
characterized by a high threshold and quick saturation
with a dynamic range of two orders of magnitude. The
maximal sensitivity of the circadian system is at
460nm whereas the visual system presents its
maximum sensitivity at 555nm.
The biological time keeping system regulates day -
night rhythms of behavior, endocrine regulation,
immune response and energy metabolism among
others. The biological clock is influenced by external
light [3]. A number of studies have dealt with the
characteristics of light influence on the circadian
rhythm. Duffy and Czeisler [4] found a dose dependent
relationship with response of the circadian system.
Glickman and Levin [5] suggested that low intensity
levels and short durations seem to affect the circadian
system more. The greater impact of light stimuli on the
circadian rhythm is during the natural dark phase.

S. Mitropoulos et al., Measurement of blue light spectrum, RAP Conf. Proc., vol. 4, 2019, 220–224
221
Chronic light exposure during the evening causes a
phase delay. The adaptation of the organism to light
exposure is also observed; therefore, the effects are
dependent on repetition. The spectrum of the emitted
light is also critical according to the sensitivity of
photoreceptors to different wavelengths. The spectral
sensitivity of the circadian system is complex and not
yet completely understood. Overall, the greatest
sensitivity is observed at short wavelengths (460nm -
490nm). While disturbance of the circadian rhythm
can occur with any type of light, the effect is more
profound when certain LEDs are used as light sources
and their influence depends on the specific properties
of the particular source.
A number of indices have been proposed to
characterize light interaction with the human circadian
system. These indices are calculated from a measurable
quantity, usually spectral irradiance E(λ).
The human optical system response to light is
characterized by Vλ, the photopic luminous efficiency
function. Depending on the total irradiance, the
photopic or scotopic luminous efficiency functions are
used to convert radiometric to photometric units.
Studies on the impact of light at night regarding the
synthesis of melatonin have shown a peak spectral
sensitivity at 460nm [6], [7]. Melatonin, also called the
sleeping hormone, is produced by the pineal gland and
is released mainly during the night [8]. Two light
variables, intensity and wavelength [9], are responsible
for the suppression of melatonin and an illuminance of
only 1,5 lux may disrupt circadian rhythms.
The melatonin suppression action spectrum
(MSAS) was published in 2001 [6], [7] and represents
to what extent each wavelength is efficient in
suppressing melatonin production. The original dataset
is quite small, ranging from 425nm to 560nm. An
extension from 380nm to 730nm was provided by
Aube [8] who used a combination of two log normal
curves to fit and extrapolate the original data.
Circadian stimulus (CS) and Circadian Light (CL)
are calculated by a mathematical model implemented
by Rea[1], based on the neuroanatomy and
neurophysiology of the retina and on published
psychophysical studies of nocturnal melatonin
suppression. CL is essentially spectrally weighted
irradiance for the human circadian system.
Special spectrophotometers based on the proposed
interaction models have been developed and calibrated
in order to measure light as it is perceived by the
circadian system (Daysimeter, Actiwatch
Spectrum)[10].
Research on the implementation of a
comprehensive quantitative model of the interaction
between light and the human circadian system is
ongoing; nevertheless, two broad suggestions from the
scientific community are the reduction of the
unnecessary high levels of light at night and the
attenuation of the short wavelength spectrum
components.
To that end a number of smartphone software
applications reduce the light emitted in the blue region
of the spectrum by controlling the relative weight of the
signal emitted by each color sub pixel, whilst
maintaining the optimal illuminant quality.
Apple has embedded their Night Shift mode in iOS,
within the quick access setting menu of their phones
from September 2016. The Android operating system
has an embedded Night Light feature in a number of its
versions, while Samsung has a night mode setting
called Blue Light Filter. A number of light filtering
applications are also available for smartphone users.
2. MATERIALS AND METHODS
The Spectral Power Distribution (SPD) of the light
emitted from the screen was assessed for a range of
smartphones. The sample of the measured
smartphones comprised six models operating with
Android operating system and two models operating
with iOS. The screen type was LTPS LCD or IPS LCD
capacitive touchscreen. The blue light filtering was
performed by the built-in filtering application (Night
Shift, Night Light). For two Android smartphones a
third-party filtering application was used and tweaked
to mimic the built-in application settings.
Measurements were made using a modular
(HR4000CG-UV-NIR) CCD array spectroradiometer
(Ocean Optics, Inc., Dunedin, FL, USA), coupled by a
metal jacketed optical fibre with a 400μm core
diameter (Ocean Optics, Inc., Dunedin, FL, USA). The
system was calibrated for spectral irradiance using a
HL-3P-CAL tungsten–halogen lamp. A special black
measuring case was constructed to ensure the
standardized measuring position of the optical fibre,
minimal reflectance of light and shielding from
ambient light. The distance of the optical fibre from the
smartphone screen was 5cm. Measurements were
performed at the center of the screen, with an all-white
background and brightness set at maximum. The first
measurement was performed without the blue light
filter and the second with the blue light reducing
application on. Measurements were performed in the
spectral range of 400nm to 750nm in terms of the
absolute spectral irradiance (μW/cm2).
Raw spectral data were weighted with the photopic
luminous efficiency function Vλ [3] and integrated to
calculate illuminance. Circadian Stimulus (CS) and
Circadian Light (CLa) were calculated from the
application available from the Rensselaer Polytechnic
Institute [11] web page.
Cλ index was calculated by integrating the weighted
SPD with the function c(λ). c(λ) was polynomially
interpolated from the data provided by Gall [12] which
was based on the melatonin suppressing action data by
Brainard and Thapan. Total spectrum and shortwave
(SW) spectrum (420nm-500nm) power were
calculated in radiometric and photometric units.
Moreover, percentile reduction for all parameter values
after the application of the blue light reduction filter
was calculated. All processing was implemented with
Octave software.

S. Mitropoulos et al., Measurement of blue light spectrum, RAP Conf. Proc., vol. 4, 2019, 220–224
222
3. RESULTS
The original measurements provide the Spectral
Power Distribution in terms of irradiance (Fig. 1). The
obtained original spectra are characterized by three
broad peaks in the blue (~450nm), green (~530nm)
and red (~630nm) wavelengths. The case presented in
Figure 1 presents an interesting pattern of the
spectrum in the red region where three distinct peaks
are observed. The application of the blue light
reduction filter selectively attenuates the short (blue)
and medium (green) wavelengths whilst maintaining
approximately the same irradiance value at higher
(red) wavelengths. SPDs were weighted with the
photopic luminous efficiency function Vλ, converted
thus in photometric units (Fig. 2). Vλ has the maximum
weighting value at 555nm where the human visual
system presents its highest sensitivity.
Figure 1. Spectral Power Distributions of a smartphone screen
with the blue light filtering application off (blue line) and on
(red line)
The reduction of the CS, CLa and Cλ parameters as
well as total and SW spectrum power and their
reduction after blue light filter application are
presented in Figure 3.
Figure 2. SPD in radiometric units, Vλ (photopic luminous
efficiency function) and their product (spectral distribution in
photometric units)
At the original SPDs with the brightness set to
maximum the percentage of short wavelength light
(420nm-500nm) was from 36% to 43%. The
corresponding SPDs with the blue light reduction
application on, had a blue light percentage of 29.5% to
39.5%. The application software led to a reduction of
the blue light wavelengths from 12% to 31% depending
on the smartphone model. Total irradiance (400nm-
750nm) was reduced by 20% at the most.
Photobiological indices characterising circadian
interaction with light were reduced from 15% to 30%
when the filtering application was used.
4. DISCUSSION
Owing to the vast use of self-illuminating devices
and the spectral characteristics of the light they emit,
concern on the effects of short wavelengths to the
human circadian system has risen during the past two
decades. Research on the implementation of a model
describing the interaction of visible electromagnetic
radiation with the human circadian system is still
ongoing. Certain photobiological parameters have been
proposed by researchers and they are used to optimize
conditions in everyday illuminating applications. As far
as self-illuminating screens (computers, tablets,
smartphones) are concerned, the reduction of short
wavelength radiation is accomplished in three ways:
external filters (glasses or screen covers), hardware
adjustments (screen technologies) and software
applications. Among the three options, software
applications are the most convenient to use and can be
applied to all smartphones.
In this study, the aim was to accurately measure
and calculate radiometric, photometric and
photobiological parameters in order to quantitatively
evaluate the alterations in the emitted light spectrum
of smartphone screens when a blue light filter software
application was used.
All applications have proven effective in reducing
the short wavelength spectral components. Their
performance, however, does depend on the
smartphone make and model. As expected, the
reduction in the total spectrum power is higher in
illuminance than in irradiance terms, whilst circadian
indices variations correspond better with the SPD in
radiometric terms.
Variations in the values of SPDs as well as photo
biologically relevant parameters are high depending on
the operating system and smartphone model. It should
also be noted that the measurements were acquired
with brightness set at maximum and full white
background of the screen. The adjustment of
brightness is available in smartphones and the
circadian parameters would most likely be reduced
with a lower brightness setting. The built in blue light
filter application of the Android OS provides
adjustment capabilities. Measurements were
performed with the highest level of attenuation
selected.
In general, up to 30% of reduction in
photobiological indices can be accomplished by
applying the filtering application.

S. Mitropoulos et al., Measurement of blue light spectrum, RAP Conf. Proc., vol. 4, 2019, 220–224
223
Figure 3. Percentile reduction in radiometric, photometric
and circadian parameters (different colors correspond to
different smartphone models)
5. CONCLUSION
As the blue light component seems to be affecting
the circadian rhythm, it is advisable to adjust screen
illumination characteristics according to the time of
day. Built in applications in most cases prove very
effective in selectively reducing shortwave visible
electromagnetic radiation no matter which metric is
used, up to 30%. The spectral power distribution in
radiometric terms seems more relevant in
photobiologic calculations and should be measured in
relevant experiments. Research on the exact
quantitative interaction characteristics of the human
circadian system and light is ongoing. However,
smartphone users can effectively reduce their exposure
to blue light maintaining good illuminating conditions
by utilizing the available software applications.
REFERENCES
1. M. S. Rea, M. G. Figueiro, A. Bierman, J. D. Bullough,
“Circadian light,” J. Circadian Rhythms., vol. 8, no. 2,
pp. 1 – 10, Feb. 2010.
DOI: 10.1186/1740-3391-8-2
PMid: 20377841
PMCid: PMC2851666
2. M. G. Figueiro, R. Hamner, A. Bierman, M. S. Rea,
“Comparisons of three practical field devices used to
measure personal light exposures and activity levels,”
Ligh. Res. Technol., vol. 45, no. 4, pp. 421 – 434,
Aug. 2013.
DOI: 10.1177/1477153512450453
PMid: 24443644
PMCid: PMC3892948
3. Opinion on Potential risks to human health of Light
Emitting Diodes (LEDs), SCHEER, Brussels,
Belgium, 2018.
Retrieved from:
https://ec.europa.eu/health/sites/health/files/scientifi
c_committees/scheer/docs/scheer_o_011.pdf
Retrieved on: Jul. 14, 2019
4. J. F. Duffy, C. A. Czeisler, “Effect of Light on Human
Circadian Physiology,” Sleep Med. Clin., vol. 4, no. 2,
pp. 165 – 177, Jun. 2009.
DOI: 10.1016/j.jsmc.2009.01.004
PMid: 20161220
PMCid: PMC2717723
5. G. Glickman, R. Levin, G. C. Brainard, “Ocular input for
human melatonin regulation: relevance to breast
cancer,” Neuro Endocrinol. Lett., vol. 23, suppl 2:
pp. 17 – 22, Jul. 2002.
PMid: 12163843
6. G. C. Brainard et al., “Action spectrum for melatonin
regulation in humans: evidence for a novel circadian
photoreceptor,” J. Neurosci., vol. 21, no. 16, pp. 6405 –
6412, Aug. 2001.
PMid: 11487664
PMCid: PMC6763155
7. K. Thapan, J. Arendt, D. J. Skene, “An action spectrum
for melatonin suppression: evidence for a novel non-
rod, non-cone photoreceptor system in humans,”
J. Physiol., vol. 535, no. 1, pp. 261 – 267, Aug. 2001.
DOI: 10.1111/j.1469-7793.2001.t01-1-00261.x
PMid: 11507175
PMCid: PMC2278766
8. M. Aubé, J. Roby, M. Kocifaj, “Evaluating potential
spectral impacts of various artificial lights on melatonin
suppression, photosynthesis, and star visibility,” PloS
One, vol. 8, no. 7, pp. 1 – 15, Jul. 2013.
DOI: 10.1371/journal.pone.0067798
PMid: 23861808
PMCid: PMC3702543
9. F. Falchi, P. Cinzano, C. D. Elvidge, D. M. Keith,
A. Haim, “Limiting the impact of light pollution on
human health, environment and stellar visibility,”
J. Environ. Manage., vol. 92, no. 10, pp. 2714 – 2722,
Oct. 2011.
DOI: 10.1016/j.jenvman.2011.06.029
PMid: 21745709
10. M. S. Rea, M. G. Figueiro, “Light as a circadian stimulus
for architectural lighting,” Light. Res. Technol., vol. 50,
no. 4, Dec. 2016.
DOI: 10.1177/1477153516682368
11. Circadian stimulus calculator, Rensselaer Polytechnic
Institute, Troy (NY), USA, 2018.

S. Mitropoulos et al., Measurement of blue light spectrum, RAP Conf. Proc., vol. 4, 2019, 220–224
224
Retrieved from: https://www.lrc.rpi.edu/cscalculator/
Retrieved on: Feb. 12, 2019
12. D. Gall, K. Bieske, “Definition and measurement of
circadian radiometric quantities,” in Proc. CIE Symp.
‘04: Light and Health, Vienna, Austria, 2004, pp. 129 –
132.
Retrieved from: http://www.cie.co.at/publications/cie-
symposium-2004-light-and-health-non-visual-effects-
30-september-2-october-2004
Retrieved on: Apr. 11, 2019
13. J. Escofet, S. Bara, “Reducing the circadian input from
self-luminous devices using hardware filters and
software applications,” Light. Res. Technol., vol. 49,
no. 4, Dec. 2015.
DOI: 10.1177/1477153515621946
14. L. T. Sharpe, A. Stockman, W. Jagla, H. Jägle,
“A luminous efficiency function, V*(lambda), for
daylight adaptation,” J. Vis., vol. 5, no. 11,
pp. 948 – 968, Dec. 2005.
DOI: 10.1167/5.11.3
PMid: 16441195