The previous code was scaling down fftResult[] to matrix height first, then computing smoothed bars.
We now to calculate smooth bars at full fftResult[] resolution first, and afterwards map the result to Matrix height. This improves accuracy and preserves a few more bits of real detail.
index.htm and index.js and peek.js:
- add expand and peek function for peek expand to removeEventListener if not expanded
- add peekexp and segvexp in settings to save status (wip)
- add File System in settings
ledmap: no lm prefix
jmap:
- also json extension (instead of jmap)
- also support json object with json array embedded:
{"n":"snake32"
,"width":32
,"height":32
,"map": []
}
usermod_mpu6050_imu.h and usermod_temperature.h:
- remove bool enabled = false/true (now default false)
- remove static const char _name[] and _enabled[]
- add constructor which calls superclass (temp?)
- replace _enabled with "enabled"
- remove const char PROGMEM init for _name[] and _enabled[]
settings_um.htm:
- Add usermod table with class properties (name and enabled) -> WIP as not all UM's now and enable will be checkbox ? (compare with info tab...?)
fcn_declare.h
- add _name and enabled to Usermod class and add name to constructor
usermods_list.cpp
- add Usermod name to constructor for mpu6050 and temperature)
- to do add enabled is true to constructor
* added usermod config information (so it can be disabled)
* check that "INTERRUPT_PIN" does support interrupts (MM specific)
* expose "dmpReady" in public interface
* adjusted games usermod to switch to "simulation" in case mpu6050 is not active
Still toDo:
1. add interrupt PIN and errormessages to settings page.
2. de-alloc intPin when disabled (and disable interrupt)
3. usermod is still destroying DEBUG_PRINT macros.
4. temporarily disable MPU interrupt when onUpdateBegin(true)
* rotary ALT: changing brightness did not have an effect (stateUpdated missing)
* four-line-display ALT: allow clockmode without NTP (network time)
- there are other possible time sources, like RTC.
useful for small -C3 boards with 2MB flash memory only. Based on suggestions from https://github.com/Aircoookie/WLED/pull/2951 .
Needs `-D WLED_DISABLE_OTA` as OTA is not possible with 2MB.
RAM: [== ] 22.5% (used 73724 bytes from 327680 bytes)
Flash: [========= ] 94.9% (used 1492020 bytes from 1572864 bytes)
the main difference is that "esp32-poe" is used as variant, instead of "esp32dev". Seems that this is needed to define a few special purpose pins on ethernet boards.
Might help to fix https://github.com/MoonModules/WLED/issues/23
optional 4MB partition layout: 1.8MB firmware (+400KB), 256KB LittleFS (-800KB); based on "min_spiffs.csv" from arduino-esp32, but increased filesystem partition from 128KB to 256KB.
FX.cpp: GEQ: even distribution of bands
FX_fcn: tiny change
wled.cpp: setupMatrix also if loadedledmap != 0
ws.cpp: replace getLengthTotal by width * height (for ledmaps)
* some typos fixed
* Added semaphores around display calls; should fix https://github.com/Aircoookie/WLED/issues/3074
-- changes still need some cleanup to improve readability...
Like DDP, this allows WLED to address network systems using the Art-Net protocol.
Universe starts at zero, because that's the first universe in Art-Net.
Works with RGB. It's coded to also work with RGBW, but I couldn't find a great place to enable it without mucking with things I don't understand.
* This is basically "app3M_fat9M_16MB.csv" from arduino-esp32 2.0.7 - just changing the filesystem type from FAT to spiffs.
* deleted the older, broken "Soundreactive 16MB" partition file.
* minor fixes for pio.ini
* adding wled.ini.cpp to .gitignore (was overdue ...)
