A Brief History of Displays and Eye Health

Preamble

This post goes through the most popular types of computer displays from the point of view of their spectral and temporal (flicker) characteristics and corresponding health effects.

More detailed analyses on how these traits affect vision and health can be found elsewhere on our website. As that is a much larger topic, this post only briefly summarizes these relations. Similarly, less popular display types like plasma, field emission, etc., are not mentioned for the sake of brevity.

And finally: our focus is on health. We don't go much into how the discussed topics relate to visual aspects of the display, like color and saturation or motion blur: there's plenty written about these elsewhere.

Cathode Ray Tube (CRT) Displays

CRTs create images by firing electron beams onto phosphorescent screens, creating light pulses where red green or blue phosphors are excited then fade out.

Image: Cottonbro Studio

CRTs: Flicker

The image is repeatedly redrawn line by line (raster scanning) at some refresh rate (e.g., 60 Hz translates to a frame duration of 16.7 ms). Since the phosphor's light emission is significantly shorter (commonly ~5 ms) than the frame duration, portions of the screen become dark between successive refreshes, leading to a noticeable on-off cycle or flicker.

CRT flicker with steady pixel value (compared to the continuous transmission of an LCD; however, see also "Flicker from the backlight" under LCDs below).

Matsumoto, C. S., Shinoda, K., Matsumoto, H., Seki, K., Nagasaka, E., Iwata, T., & Mizota, A. (2014). What monitor can replace the cathode-ray tube for visual stimulation to elicit multifocal electroretinograms? Journal of Vision, 14(9), 2-2.

This inherent flicker production was both a blessing and a curse. CRTs offered precise timing and little motion blur, but the flicker was clearly noticeable and disturbing for many people.

This also prompted more research into the negative effects on health and performance. Connections to epilepsy and the disruption of saccadic eye movements have been uncovered, resulting in attempts to increase the flicker rate of CRTs. Refresh rates increased from 60 Hz to 100-120 Hz, making the flicker less visible but not eliminating the problem.

Health Impact of Flicker

Flicker, whether consciously perceived or not, can lead to a range of adverse effects on visual comfort and overall well-being. These include eye discomfort or fatigue, headaches, nausea, migraines, and visual artifacts.

Flicker has been shown to disrupt saccade planning, resulting in less precise eye movements and reduced visual efficiency, particularly during visually demanding tasks like reading. This exacerbates overall fatigue and can lead to disturbing visual artifacts for sensitive individuals. These effects do not necessarily decrease with increased flicker frequency: depending on the situation, faster flicker rates can also be more problematic.

Initially, flicker fusion rates (the frequency at which flicker becomes imperceptible) for large-field static presentation were the main metric of concern. Based on this, frequencies above 50-100 Hz were disregarded as irrelevant. This was despite the fact that animal research had already alluded to that the retina reacts to flicker even if it is suppressed from conscious perception.

As the applied methodology got more nuanced, frequency thresholds gradually increased and showed dependence on circumstances like saccade velocity, stimulus size and contrast, or the profile of temporal modulation. Today, the highest recorded conscious detection of flicker is above 15 000 Hz — and depending on the paradigm, this can certainly be exceeded as well. Meanwhile, we do not know where there is, if anywhere, an upper frequency threshold for negative health effects that remain outside of conscious perception.

CRTs: Spectrum

CRTs produce a relatively broad RGB spectrum, depending on the phosphors used. In general, these spectra are incomplete, but still less unbalanced than that of newer display types; the three channels together cover most of the visible range to some extent. Reds are deeper than with contemporary screens, which makes CRTs slightly less problematic in terms of eye physiology.

Spectral power distribution of a CRT display. The spectrum is relatively continuous across the shorter half of the visible range. Deep red is also present, although the long-wavelength region is highly spiky.

Sharma, G. (2002). LCDs versus CRTs-color-calibration and gamut considerations. Proceedings of the IEEE, 90(4), 605-622.

Spectral Bandwidth and Eye Health

A broader spectrum means less ability to control colors and their saturation. Therefore, the trend throughout the history of displays (and already during the time of CRTs) has been towards ever narrower RGB spectra, with a narrative that knocks the broader spectral bandwidth of earlier technologies.

We view this differently. From much research in light physiology, it is clear that each frequency in the spectrum (within and outside of the visible range) has physiological effects. These are always in balance in natural light, covering the whole range and dynamically shifting the ratios between different parts of the spectrum according to the time of day and year as well as geographic location. These ratios and their dynamic interactions determine the cyclicity of our physiology with massive health implications.

The narrower the spectral bandwidth in a display, the more these ratios are pushed out of balance, while circadian and annual changes as well as harmonization to the user's location are also missing. Physiological reactions associated with the presented frequencies are continuously triggered, while those that are linked to the missing parts of the spectrum remain dormant. The "blue light" issue is part of this question — but it's not that blue light or any frequency of light would be harmful per se. The problem is that the whackier the spectrum, the more our physiological reactions to it will also be out of whack.

From this perspective, it is a sad coincidence that our vision works in much simpler ways than our physiological reactions to light. As a consequence, narrow-band RGB light can trick us into seeing a large range of vivid colors, incentivising displays to become less and less natural in their output. Through this post, we'll see how ever narrower RGB display spectra were created, boosting visual performance but also deepening the negative health impacts.

Each color channel and pixel is controlled by the electron beam and can be selectively turned off, resulting in deep blacks and good control of the spectrum. For example, if the image did not contain any blue, then the output was also lacking the blue frequencies — unlike with some more contemporary technologies.

In sum, CRTs are mainly remembered for their horrendous flicker, which was their primary problem not just for health but even for usability. Their spectral output was relatively broad and less problematic than that of most newer display types. Their reign lasted until about the early 2000s, when the elimination of severe flicker and a huge reduction in display size and weight made LCDs a much welcomed change.

Liquid Crystal Displays (LCDs)

LCDs use color filters organized in a pattern of red, green, and blue pixels, to modulate (filter) a white backlight source that provides homogeneous illumination from behind the screen. The light output of the display therefore depends on two factors:

The backlight source: its spectral properties and temporal modulation (flicker)
The filtering properties of the LCD:

how the backlight spectrum is further reduced by the LCD (depending on the bandwidth of the RGB phosphors)
whether the LCD itself introduces another source of temporal modulation (flicker)

Image: American Optronics, Inc.

Relative to CRTs, LCDs allowed for slimmer, lighter panels with no geometric distortion; however, contrast and black level performance were initially inferior. Notably for health, the spectrum is less directly dependent on the image presented, as the backlight source would always shine through to some degree. This means that the spectral and temporal characteristics of the backlight source are crucial determinants of the final light output and its physiological impact.

Early LCDs using cold cathode fluorescent lamp (CCFL) backlights

LCD spectrum with CCFL backlight. Deep reds are lacking and the center of the visible range is quite spiky. The spectrum is not as blue-dominant as with later display types.

Sharma, G. (2002). LCDs versus CRTs-color-calibration and gamut considerations. Proceedings of the IEEE, 90(4), 605-622.

Early LCDs had less flicker than CRTs — their CCFL backlights flickered with much higher rates and sometimes lower depths. To many flicker-sensitive people, this was the golden age of displays when overall flicker characteristics were least disturbing (though still present due to the inherent control and dimming of CCFLs).

Spectral properties, on the other hand, were generally worse compared to CRTs. CCFLs are infamous for spikes in the spectrum, creating 'weird' colors and this was not different with displays either. The pleasant and deep reds of CRTs were gone, and so were crisp edges and great black levels.

Later LCDs using light-emitting diode (LED) backlights

Spectrum

Most LED backlights produce blue-rich white light where the emission spectrum is characterized by a blue spike (created by a blue "pump" LED source) and an additional orange spectrum (created by the phosphor that is excited by some portion of the blue LED). The two together give white light with less broad spectrum compared to earlier CCFLs, approximately doubling the relative contribution of blue in the 435-465 nm range.

LCD phosphor transmission and LED backlight spectra with contemporary PC-white LED backlight source. While the red channel would be capable of letting through deep red wavelengths, these are completely missing from the backlight spectrum. Note also that the spectral output is dominated by the LED's blue pump.

This was a step forward in visible display characteristics (color, brightness and adjustability) and with improvements also in LCD technology, screens have become more vivid, offering better contrast and visibility also in bright conditions.

On the other hand, they introduced new issues regarding comfort and health. While the blue-dominant backlight output helped to create more attractive colors, it also had a greater effect on circadian rhythms and caused additional eye discomfort, particularly during evening use.

This happened also at the time when the topic of harmful effects of blue light took off, with more people feeling the stressor effects of these new display emission spectra, inducing discomfort and sleep disruption.

A growing body of research started to make clear that short-wavelength light can induce oxidative stress and accelerate the aging of tissues, especially of the most sensitive central parts of the retina.
These cells also happen to be the most exposed during display viewing.
With the myopia epidemic worsening and screen-times increasing globally, a connection between the two was also becoming more broadly recognized. While basic research on animals already in the 1990s and 2000s demonstrated a clear link between light environment and eye development, the topic only gradually got more well-known.

Flicker from the backlight

Similarly, flicker characteristics changed in novel and largely unfavorable ways with LED backlights. LEDs are more efficient and easiest to dim when they are driven in a temporally modulated fashion (called pulse-width modulation), where current to the LED is cut periodically. Here, the frequency of modulation can stay constant while changing the proportion of the 'on' and 'off' phases of the flicker cycle, the apparent brightness of the source can be set through a broad range. While this method itself also applies to earlier light sources (including CCFLs), LEDs can change their state very rapidly.

LEDs are able to turn their output on and off rapidly.

The resulting flicker has a square profile. Unlike flicker from any earlier source, the abruptness of change in light output increased both the potential visibility of flicker as well as its negative effects on comfort and health (which extend well beyond the directly visible frequencies).

As these issues are gradually more recognized, backlight frequencies have been increased and pulse-width modulation is often partially or fully replaced by continuous current modulation to reduce flicker. The latter is still sadly not so common, while increasing the flicker frequency does not eliminate the problem (in fact, in some circumstances, higher frequencies can be more detrimental).

Flicker from the LCD

The refresh rate of an LCD simply defines how often the image can be changed. It is a common confusion to equate this with flicker and flicker rate, but the two are usually independent: the main source of flicker in LCDs is the backlighting (see above), while the display itself might only flicker to a small extent or basically not at all.

When the state of the LCD does not change (the displayed image remains the same between two frames), the light output can remain generally continuous, but some LCDs do add an extra source of flicker with shallow depth. The reason behind this is that current polarity between frames has to reverse and this can lead to a tiny difference in how light is passing through the pixel (a flexoelectric phenomenon).

Chen, H., Peng, F., Hu, M., & Wu, S. T. (2015). Flexoelectric effect and human eye perception on the image flickering of a liquid crystal display. Liquid Crystals, 42(12), 1730-1737.

This source of flicker was negligible with early TN type LCDs. However, more modern IPS/FFS technologies introduced better viewing angles and color stability, coming at a cost of greater flicker from flexoelectric origin between frames.

Temporal dithering

Meanwhile, a new source of flicker has also been introduced: temporal modulation of the LCD can further increase color range, adding a localized and backlight-independent component of flicker to modern display light. A pixel's color rapidly alternates between two adjacent values, creating the illusion of an intermediate color that the hardware cannot natively display, thus effectively increasing the perceived bit depth and smoothing gradients.

LCDs — Summary

In summary, LCDs replaced CRTs due to better packaging and less noticeable flicker, while their visual qualities were generally lower at first. Contemporary LCDs with LED backlighting featured bright and blue-rich light emissions, allowing for better visual performance but introducing concerns regarding sleep regulation, ocular surface health, retinal aging, and myopia progression.

Organic Light-Emitting Diode (OLED) Screens

A defining characteristic of OLED displays is their self-emissive nature; each individual pixel directly generates its own light. This eliminates the need for a separate backlight unit, a fundamental component of LCDs.

Image: American Optronics, Inc.

Their spectral characteristics are largely similar to those of LED/LCD screens, continuing the trend towards ever narrower bandwidth RGB spectra. However, the increased control over individual pixels allowed to decrease blue light emissions compared to LED-backlit LCDs. Further advantages of individually emissive pixels are deeper blacks, great contrast, and wide viewing angles.

On the negative side, flicker characteristics have worsened relative to LCDs. Many OLED displays use pulse-width modulation, where pixels "blink" to refresh with every cycle and the duty ratio is changed to achieve different perceived brightness levels.

This may sound similar to how CRTs operated, but in practice, flicker with OLEDs can be much more complicated. Due to each pixel being an OLED light source, they can theoretically stay on for a long period of time, and each pixel can be cycled on and off independently. To reduce potential burn-in (permanent image retention), many OLED panels periodically shift pixel positions or apply compensation pulses that temporarily adjust the luminance of static areas, preventing uneven degradation of the organic materials. This, combined with frequently applied temporal dithering techniques, make OLED display flicker a complex situation that is difficult to predict and that disturbs many users in novel ways — even if they were not bothered by earlier LCD screens.

Outlook

Most modern displays are variations of the above two technologies. Quantum dots (QD) have added further control over spectra through decreased bandwith, but QLED and QD-OLED (or even microLED) technologies work otherwise similarly to LCDs and OLEDs, so we do not discuss them in detail.

RGB spectra are getting narrower with newer technologies.

Image: PR Newswire

These technologies decreased the bandwidth of emission and allowed for the creation of more arbitrary spectra. This increased color saturation and gamut width (the range of possible colors). While such features are attractive at first sight, they also worsens the displays' health consequences. A narrower RGB spectrum is physiologically less natural, makes focusing on the display more difficult, and contributes to temporary myopia that likely translates to long-term changes in focus as well. Similarly, the unnaturally high saturation shifts our perceptual baseline for expected colors, rendering the real world more dull: a long-lasting effect with worrying developmental consequences.

OLEDs also allowed for getting better scores for displays in health-related metrics like the circadian action factor. These changes are more nominal in nature and their real-world effects are not so clear: when a metric is based on a certain threshold or a weight function, scores can be improved by going just slightly out of the measurement range while physiology does not work in such a distinct way. Nevertheless, this may have slightly reduced the tissue stressor effects of blue light emission by shifting the blue peak to somewhat longer wavelengths, relative to earlier LED-backlit LCDs.

While the lighting and display industry was first ridiculing then fighting any assertions about potential dangers of blue light, it had to cave in eventually. Since it is vital for any display to have sufficient emission in the blue range, the narrative focus is not to reduce overall blue emissions but to identify specific problematic wavelengths and to reduce nighttime blue exposure.

These narratives lack the understanding of why blue light from displays has greatly different effects from that of daylight. The difference stems from physiological interactions between various parts of the spectrum, so the same dosage of short-wavelength light will have a different tissue stressor effect depending on the concurrent presence and nature of long wavelengths (see related literature on our References page). Therefore, the real solution is not to create minute differences in wavelength, while continuously reducing bandwidth, but to increase bandwidth and make the spectral composition more balanced and natural.

In that regard, the sun is the gold standard and that is why Pixun channels daylight directly into the display. This way, from violet to the long end of near-infrared, the whole spectrum remains in the display's emission with all the natural dynamics of circadian and annual cyclicity, and of course without any flicker. No electric source can replicate that.

Google Sites

Report abuse