When the project was initially announced, Pine was claiming a 60Hz refresh rate, which sounded amazing but is essentially impossible for e-ink at full dynamic range. They've since clarified this:
>Edit August 16 00:09 UTC: A previous version of this post listed the e-ink panel’s refresh rate at 60Hz. This number requires much more context. It takes multiple frames to display most images on an e-ink panel. The visual performance of the panel also depends on the method of converting the screen image to data the panel understands. We will be unable to make estimates of the panel’s true performance in frames per second without much more testing and development. So we’ve removed the 60Hz figure for now. We apologize for any misunderstanding.
Realistically, the best you can do with e-ink is 1-bit with ~120ms refresh (just over 8 frames per second). If you want shades of grey it slows considerably. Dasung and now Boox Mira both support alternative modes that compromise on dynamic range, number of shades of grey, and amount of ghosting to be able to display video somewhat adequately.
> Realistically, the best you can do with e-ink is 1-bit with ~120ms refresh
I'm curious, why is that so? Can't the control algorithm boost the speed by sending higher voltages when switching, then going back to "holding" voltage?
No it can't. (voltage is applied to entire row - select cells - at once) The algorithm can realistically only affect timing. (have the voltage applied to the entire row longer, but that's not very useful if you have varying shades of gray within the row)
How about a PWM-like approach? E.g. select one row with high voltage, and activate the columns with a much higher frequency (where "longer on" results in more effective power into the cell)?
