Change colour depth (pwm bits)

[arahasya]

I don't know the relation between colour depth and pwm bits but from results I know from rpi-rgb-led-matrix using --led-pwm-bits=1 gives the best refresh rate I want to display text only without any shade so it is suitable me it. Is it possible to do the sane using this library?

Thanks

[rjrouquette]

Color depth and PWM bits are independent. The color depth is fixed at 24-bit RGB. The PWM bits control the fidelity of the color rendering.

For your case, you'll need to set the PWM bits to 1, and reconfigure the DPI with 1 PWM bit plane. If your panels have less than 16 physical scan lines, you might need to increase the number of bit planes to make the DPI configuration valid. I don't recall the minimum supported dimensions offhand.

You'll still need to use 24-bit RGB with setPixel(), but the actual output will be binary.

[rjrouquette]

Also of note, is that if you reduce the number of PWM bits, you'll also limit the usefulness on the brightness setting. The brightness control relies on the PWM resolution.

[arahasya]

Really don't understand this formula for Pwm bit planes

22 = 7+(1+2+4+8)

Can you please explain this again?

[rjrouquette]

22 = 7 + (1 + 2 + 4 + 8)

7 = number of PWM bits (1 + 2 + 4 + 8) = BCM bits

There are 7 PWM bits and 4 BCM bits which yields the 11-bit overall depth.

• [Overall Depth] = [PWM bits] + [BCM dits]
• [Bit Planes] = [PWM bits] + ((2 ^ [BCM bits]) - 1)

This is why for 1-bit overall depth you get 1 bit plane, and no brightness control:

• 1 = 0 + 1
• 1 = 0 + ((2^1) - 1)

[arahasya]

So the framebuffer_height = 1 + 1
Only 2? That will give me a very high refresh rate don't think driver IC supports that high

[rjrouquette]

No. The frame buffer height is derived as follows:

• height = ([scan rows] * [bit planes]) + 1

So with a 3-bit row address it would be 9:

• 9 = (8 * 1) + 1

I suspect that might be too small for the DPI driver, so you may need to add more bit planes and color depth. With 3-bit row select and 2-bit depth it would be:

• 25 = (8 * 3) + 1

[arahasya]

Oh yes sorry forgot to add 8 I'll run the tests tomorrow and get back with results

[arahasya]

Working. Had to reduce avr clock