Kobo forma screen flickering/pwm issues

[Tokei]

Originally Posted by davidfor

Someone who took a video calculated the PWM frequency at about 1KHz. But, I have had discussion with someone at Kobo that confirmed that it is 1KHz.

Many thanks! The site notebookcheck.net which tests all devices in their reviews for PWM quotes a study that found that PWM frequencies of up to 400Hz can have detrimental effects (even though not visible). Link. So 1000Hz should really not pose a problem.

Mhm.... perhaps I can take off my tinfoil glasses now

[davidfor]

Originally Posted by Tokei

Many thanks! The site notebookcheck.net which tests all devices in their reviews for PWM quotes a study that found that PWM frequencies of up to 400Hz can have detrimental effects (even though not visible). Link. So 1000Hz should really not pose a problem.

Interesting article. But they don't say where they got the 400Hz figure from. They mention that tests have shown retina stimuli at 200Hz, but nothing higher. It sort of sounds like the doubled that for safety. Which I completely understand.

The image at the top of the article is, um, interesting. I had to scroll away from it quickly. I think if my wife saw it, it would be instant migraine. And for the record, my wife suffers severely from migraines. She will get one from just about anything. But not the Forma.

Mhm.... perhaps I can take off my tinfoil glasses now

[Tokei]

Originally Posted by davidfor

But they don't say where they got the 400Hz figure from.

I think they based that on the IEEE 1789 standard. "Even frequencies in the range of 100–400 Hz can have a negative influence on the human organism, as is explained in IEEE 1789"

I could find that standard, but I didn't see the 100-400hz anywhere in the document (though admittedly I didn't read the whole thing). The paper is a meta-study, drawing a low-risk / no-effect curve from collating various studies on the topic.

In their conclusions they actually use PWM as an example on page 46: "[...] the recommended practice for PWM dimming at 100% modulation depth is that the frequency satisfies f > 1.25 kHz. [...] This level of flicker could help minimize the visual distractions such as the phantom array effects. The recommended NOEL for PWM dimming is 3 kHz [...]." NOEL = "no observable effect level".

The paper itself is behind a paywall but the US Dep. of Energy has a nice summary here: https://www.energy.gov/sites/prod/fi...htfair2015.pdf

[davidfor]

Originally Posted by Tokei

I think they based that on the IEEE 1789 standard. "Even frequencies in the range of 100–400 Hz can have a negative influence on the human organism, as is explained in IEEE 1789"

I could find that standard, but I didn't see the 100-400hz anywhere in the document (though admittedly I didn't read the whole thing). The paper is a meta-study, drawing a low-risk / no-effect curve from collating various studies on the topic.

That's sort of what I meant. They gave examples up to 200Hz, but nothing for the 400Hz. If I was writing these sort of standards I'd pick the worst results than add a clear buffer.

In their conclusions they actually use PWM as an example on page 46: "[...] the recommended practice for PWM dimming at 100% modulation depth is that the frequency satisfies f > 1.25 kHz. [...] This level of flicker could help minimize the visual distractions such as the phantom array effects. The recommended NOEL for PWM dimming is 3 kHz [...]." NOEL = "no observable effect level".

The paper itself is behind a paywall but the US Dep. of Energy has a nice summary here: https://www.energy.gov/sites/prod/fi...htfair2015.pdf

That's an interesting summary. But, it just confirms to me that the 1KHz that the Forma uses is not going to be an issue.

[omnimodis78]

Originally Posted by davidfor

Someone who took a video calculated the PWM frequency at about 1KHz. But, I have had discussion with someone at Kobo that confirmed that it is 1KHz.

So this is the absolute confirmation that it uses PWM. Thanks for that. Did your contact provide some qualitative reasoning as to why Kobo chose to deliberately go with PWM? I'd love to understand the reasoning as, if for no other reason, I'd assume it's turning out to be a PR issue for them.

[PorcupineLTD]

Acording to this study that was used in defining the IEEE 1789 standard https://journals.sagepub.com/doi/abs...urnalCode=lrtd (paywall) humans can perceive flicker at up to 4.9 kHz (one participant in the study was able to) with the mean (of the study participants) being 2.47 kHz.

One important parameter that is not being discussed is the modulation depth/amplitude (difference between peek and low) which can compensate for the frequency.
For example incandescent light bulbs have a frequency of 100-120 Hz but because the modulation depth is less than 10% flicker is not usually perceptible compared to older fluorescent lighting that had a modulation depth of 30-50% at the same frequency.
So the higher the modulation depth the higher the frequency is needed to make flickering imperceptible. An estimated 4kHz would be needed for 100% modulation depth.

The 1kHz frequency of the Forma frontlight would be more informative were we to also know the modulation depth/amplitude.

[sherman]

Originally Posted by omnimodis78

So this is the absolute confirmation that it uses PWM. Thanks for that. Did your contact provide some qualitative reasoning as to why Kobo chose to deliberately go with PWM? I'd love to understand the reasoning as, if for no other reason, I'd assume it's turning out to be a PR issue for them.

It's probably cheaper and easier to implement. No need to vary the current and/or voltage with PWM control. Especially if you're software controlling brightness.

[RAFS]

Very fascinating thread, actually. I say this cause I recently purchased the Aura H2O Edition 2, and found that while reading at night, with the light, something bothered me. It appeared to be pulsating (yes, I realize that this thread is about the Forma). I took a video (only 60FPS, cause I didn't read this thread until now), but it didn't pulsate during playback. Honestly thought I was just going a bit batty. But it bothered me enough to return the H2O, and pick up the Clara HD instead. The light on the Clara has been great, does not cause me the same issue. After reading this thread, I did the 240FPS video on the Clara, and it does not produce any of the flicker that I see in the videos here (it's very steady at different brightness levels and colours). I sure wish I would have done the same 240FPS video for the H2O, just to see if that was in fact the issue.

Again, I know it's not the Forma. I'm just glad that I'm not quite as batty as I think I may be.

[Tokei]

After some time reading with the Forma I am tending towards selling it . Cannot return thanks to Rakuten Japan’s medieval return policies, grrrr. I read in various settings from sun light to artificial light and in the dark and have the feeling that the text is not “calm” for a lack of a better word. I read screen to screen vs. my old Kindle Paperwhite that did not show this effect and had a calmer display. Light color, sharpness, light uniformity and basically everything else was much better (unsurprisingly) in the much more modern Forma.

I also did a “blind” test with my wife handing her both the Paperwhite and Forma asking how she likes the new reader. I didnt tell her about the whole PWM thing but she had the same impression.

Dunno if it is my mind playing tricks or not, but now that I am “tainted” it doesnt make much of a difference whether it is a real or just imagined irritation.

[DNSB]

Originally Posted by omnimodis78

So this is the absolute confirmation that it uses PWM. Thanks for that. Did your contact provide some qualitative reasoning as to why Kobo chose to deliberately go with PWM? I'd love to understand the reasoning as, if for no other reason, I'd assume it's turning out to be a PR issue for them.

For the most part, PWM gives more linear dimming and colour consistency than controlling the brightness by modulating the current.

My personal opinion about the fuss over PWM is much that of one university study. They attempted to test what effects were caused by PWM in a blind test. The results were that 99% of the people tested could not tell 200Hz PWM from continuous current modulated brightness as the red/green/blue LED light source was moved between 10% to 90% brightness. Not very surprisingly, people who complained about eye strain and headaches were within .1% equally split between the PWM and continuous light sources.

The paper was almost worth the cost for the description of the computer control of the modulated current LED lighting that was required to make it more linear with consistent colour.

Admittedly this was not tested in circumstances where saccades would be a common happening since the test images being illuminated were not text pages.

As for being a PR issue? Other than the discussions here, I haven't heard anything else negative or positive about PWM and Kobo ereaders.

[PorcupineLTD]

Originally Posted by DNSB

My personal opinion about the fuss over PWM is much that of one university study.

Could you please link said study?
Also good LED bulbs (like Philips ones) use DC and don't flicker at all.

[DNSB]

Originally Posted by PorcupineLTD

Could you please link said study?
Also good LED bulbs (like Philips ones) use DC and don't flicker at all.

I can't link the study since the only version I've seen was a hard copy in the corporate library and that was a cost item. One of the librarians tossed it at me one day as something I might want to take a quick look at while waiting for the cable guy to finish pulling some Cat 6a cabling. Who knows, one day we might do 10Gbps to the desktop.

As for the light source used in the study, it used individual red, green and blue LEDs probably to allow greater control over the colour and brightness.

[PorcupineLTD]

Originally Posted by DNSB

I can't link the study since the only version I've seen was a hard copy in the corporate library and that was a cost item.

Sorry but I am skeptical of conclusions drawn from unsourced studies. Can't you remember the title or some approximation of it? Also for 99% of participants to observe no difference there would need to be a quite high number of participants (over 100) surely such a big study on the matter can't be that obscure.

[DNSB]

Originally Posted by PorcupineLTD

Sorry but I am skeptical of conclusions drawn from unsourced studies. Can't you remember the title or some approximation of it? Also for 99% of participants to observe no difference there would need to be a quite high number of participants (over 100) surely such a big study on the matter can't be that obscure.

Sheesh.... I looked through that study a few months back. I consider myself lucky I remember as much as I did. The next time I have to visit the corporate library, I'll take a look but Ghod alone knows when that will be. The only reason for my last visit was a loss at rock/scissors/paper -- our preferred method of setting who gets the boring tasks.

[koboyedi]

The flickering is also visible to the camera when not filmed in slow-motion. While it's not visible while reading, your brain still processes these. Here's a short clip showing the issue:

https://vimeo.com/383801739