* added bitwise operation based sqrt16
- replacement for fastled, it is about 10% slower for numbers smaller 128 but faster for larger numbers. speed difference is irrelevant to WLED but it saves some flash.
* updated to 32bit, improved for typical WLED use
- making it 32bits allows for larger numbers
- added another initial condition check for medium sized numbers
- increased the "small number" optimization to larger numbers: the function is currently only used to calculate sqrt(x^2+y^2) which even for small segments is larger than the initially used 64, so optimizing for 1024 makes more sense, although the value is arbitrarily chosen
- Implement vector in bus manager
- Memory calculation according to explanation from @Makuna
- Prefer 8 RMT before 8 I2S on ESP32 (fixes#4380)
- speed improvements in ABL
- verbose debugging
- get bus size from NPB (prototype)
- Parallel I2S output bugfix
- automatic selection of appropriate I2S bus (`X1xxxxxxMethod`)
- removed I2S0 on ESP32 (used by AudioReactive)
- renumbered internal bus numbers (iType)
- added buffer size reporting
This is purely a "clean code" thing, no impact on function - it helps to avoid confusion when reading the code.
C++ allows declaration and implementation to use different variable names.
* changed some parameters to "pointer to const", so compiler can better optimize code size and performance - because data behind a const pointer will never be modified by the called function.
* made setPixelColor `const`
* fixed a few potentially uninitialized local vars (the may have random values if not initialized)
* avoid shadowing "state" in handleSerial()
* plus a few very minor improvements
Both ESP8266 and ESP32 have a hardware random register. This update makes use of that. It is slightly faster than the fastled variants but mostly it is truly random, even when the timing limitations stated in the datasheet are disregarded. Also saves a bit on code size.
- Replaced all random8() and random16() calls with new hw_random() versions
- Not replaced in FX where PRNG is required
Removing the bool saves on code size and makes the function a tiny bit faster. Also this is a cleaner solution IMHO.
-updated blend function to optimized 8bit calculation
- efficient color blend calculation in fews operations possible
- omitting min / max checks makes it faster on average
- using 8bit for "blend" variable does not significantly influence the resulting color, just transition points are slightly shifted but yield very good results (and better than the original 16bit version using the old fastled math with improper rounding)
- updated drawCircle and drawLine to use 8bit directly instead of 16bit with a shift
Processing of received button command is no longer processed in the callback, instead the value is saved to a variable and processed in the main loop.
The actual fix is to not access the file system while data is being sent out: even just trying to open a non-existing file causes glitches on the C3. Waiting for the bus to finish fixes this BUT it causes a frame-delay which is the lesser evil than random color flashes.
- increase WLED_MAX_BUSSES for C3 (fixes#4215)
- fix for #4228
- fix for very long running effect (strip.now, strip.timebase)
- C++ API change to allow `seg.setColor().setOpacity()`
- `sin16_t() / cos16_t()` are faster and more accurate than fastled versions
- `sin_approx() / cos_approx()` are float wrappers for `sin16_t() / cos16_t()` and are accurate enough to replace `sinf()/cosf()`
- `atan2()` is used only in octopus to calculate center offset, new approximated version saves flash
- `tan(), atan(), asin(), acos(), floor(), fmod()` are used only for sunrise/sunset calculation, using wled_math version saves flash
- `beatsinx()` replacements are to make use of new `sin16_t()/sin8_t()` functions to reduce flash size
- Extensively tested surnise/sunset calculation: deviation is 1min. max
- Tested some of the relevant FX and found no visual difference: Julia, 2D Drift, Drift Rose, Ghost rider, Rotozoomer, Palette, Arc 1D expansion
- total flash savings: 7.4k
uint16_t to unsigned to make it consisten throughout the hand-down.
colorFromPaletteWLED now returns uint32_t which saves the conversion to CRGB and back to uint32_t (in most uses at least).
also added (preliminary) CRGBW struct. I tried to use it in place of uint32_t colors but it adds a lot of overhead when passing the struct so reverted to uint32_t in most places.
updated a few FX to use the CRGBW struct and also cleaned some code to improve flash useage.
- also added a struct to handle HSV with 16bit hue better (including some conversions, can be extended easily)
- the functions are optimized for speed and flash use. They are faster and more accurate than what fastled offers (and use much less flash).
- replaced colorHStoRGB() with a call to the new hsv2rgb() function, saving even more flash (new function is untested!)
- the 16bit hue calculations result in an almost perfect conversion from RGB to HSV and back, the maximum error was 1/255 in the cases I tested.
- blurring now uses desaturated adding: it is faster most of the times and blurring adds scaled colors so should rarely (ever?) saturate, I saw no visual difference in tests.
- formatting
Remove the large stack buffer as we're just going to copy it in to a
heap buffer anyways. Later we can refine the length estimation or use a
rope-style dynamic data structure like DynamicBufferList.
- optimized color_add() again: now it is as fast with preserved ratio scaling than the "fast" variant was before (if no scaling is needed, it is even faster). plus it saves 250 bytes of flash
- bugfix in `color_fade()`
- removed a lot of whitespaces
Blaž Kristan