I would buy a Sofirn C01s with

what ramp calculator are you talking about?

wle

but that is not what they are offering… it is 100-3… or maybe that is what you mean, you want the modes as is, just LH

maybe you could ask Sofirn to go LH?

but that is not what they are offering… it is 100-3… or maybe that is what you mean, you want the mode brightness levels as is, just LH

maybe you could ask Sofirn to go LH?

or, did you mean youre fine with HL, and are one of the 9% of the votes in favor of HL?

H-L is only usable for keychain light for me.
For EDC-light I prefer L-(M)-H with a sublumen mode.

This one.

Why did Sofirn go with H-L instead of L-H with these survey results? (C01S BLF on ali)

AFAIK there were no “surveys” prior to release.

> why …?

My guess is, the nghts in urban China are never dark enough to appreciate why some of us want low modes.

Sofirn does not know I started this poll.

I started this poll after the light was already announced.

.

I think it is too late to change the Sofirn C01s driver, this poll was just for my curiosity.

I think that high-low is better for the general public who just wants light when a flashlight is switched on. People may never discover the second mode.

Which makes it just a bit nicer if we manage to arrange a BLF edition with low-high :slight_smile:

Sort of off-topic for this thread, I’d like ask about the “stepped ramp” modes in Anduril — which one does it use to calculate the spacing for the n# of levels?

low first, nothing worse then having a small light that starts with a 100mode lumens first… ughh

Is there a reason that both HL and LH mode sets can’t be implemented in cases like this? I.E, go to Low mode, wait 10s, output blips, power cycle —> mode order is reversed?

how is this ‘squared’?

0.30 / 24.9 / 88.9 / 192 / 335 / 517 / 739 / 1000 lm: Square

i looked at the code and still don’t get it

obvs the .3->24.9 jump is something else, but how is 25-89-192… related to squaring anything?
or even square root? (code seems to do X to the 1/2 power in it)…

wle

Square of Voodoo, everybody knows that :slight_smile:
Some reference to your vision perception maybe or just Voodoo.

Exactly, sales to the general public are what makes the company money. “Wow 100 lumens from an AAA light?!? and a low mode so if I’m stuck in an elevator, I can have light for a couple of days! Cool!!! “

For a small twisty light, I think 2-3 modes are enough with the highest mode last and no mode memory. I like the idea of 8-40-100. It gives you a useable low, with higher modes available.

I did order one as is. This will be my keychain backup (or gifted if I don’t like it)

It uses this to calculate the spacing for the stepped ramp.

The ramp is generated using the level_calc.py tool, and the steps along that ramp are generated using code similar to steps.py.

It’s the shape of the curve. The user sets the bottom, the top, and the number of steps… and the code picks points along that curve. The algorithm is like this:

  1. Draw a curve. For example, “y = x squared”, or “y = log(x)”.
  2. Pick several points along the X axis. These points must have the same amount of space between them.
  3. Record the Y value at each of those points. These are the number of lumens the light should emit to make an “evenly spaced” set of levels.

For example, let’s use the log curve as suggested earlier. For a formula, it’ll be 10X , and the values suggested earlier place X 0.5 units apart starting at –0.5:

  • 10–0.5 = 0.31 lm
  • 100.0 = 1.0 lm
  • 100.5 = 3.1 lm
  • 101.0 = 10 lm
  • 101.5 = 31 lm
  • 102.0 = 100 lm

These are evenly spaced according to a log curve, because they are sampled at equally-spaced X values.

Now the same thing, six evenly-spaced levels from 0.3 to 100, using X2 as the curve shape. Start with X=0.55, and the step size is 1.89:

  • 0.552 = 0.30 lm
  • 2.442 = 5.94 lm
  • 4.332 = 18.7 lm
  • 6.222 = 38.7 lm
  • 8.112 = 65.8 lm
  • 10.02 = 100 lm

To see the evenly-spaced inputs, just reverse the formula. Instead of x2, use sqrt(x). Put that sequence through sqrt(x), and “0.30 / 24.9 / 88.9 / 192 / 335 / 517 / 739 / 1000 lm” becomes “0.55 / 4.99 / 9.43 / 13.86 / 18.30 / 22.74 / 27.18 / 31.62”. They are each ~4.44 units apart.

Does this make sense?

Sometimes people use log() as an approximation for what looks evenly-spaced, but it’s only an approximation. It looks good for a fairly narrow range of output, like from 3 lm to 100 lm… but it kinda gets off track outside that range. Lower modes end up too close together, and higher modes end up too far apart.

The curve which most research recommends is x3 , because it generally produces good results across a wide range of brightness. It seems to match human perception of light pretty well, in terms of being visually “linear”.

However, I don’t always use the cubic shape. For lights which get really bright, sometimes I’ll use a steeper curve because it provides more resolution toward the bottom end of the ramp. Most of what people seem to use is under 1000 lumens, and anything above that tends to be treated as “turbo” or a “burst” mode. So I try to make sure there is a good amount of resolution under 1000 lumens, even if it means the upper levels are spaced a bit far apart.

I would buy a Sofirn C01S with Andúril or at least with RampingIOS V2.0 :heart_eyes: :stuck_out_tongue:
(just kiddin’)

OK, TK.
IC.
tx.
:slight_smile:

wle

If you search C01S in the google search side bar it will show every single thread because of the side bar poll. This is worse than pulling in results from signatures!