What is an acceptable loop time? (performance)

ESPHome
1

I’m having an ESP32-C3 doing decibel measurements with an I2S microphone ( GitHub - stas-sl/esphome-sound-level-meter ).

The debug log loop time shows it takes 6000+ ms and the web interface is responding slowly. (15+seconds before all values are loaded)

Another board (same ESP32-C3 from Olimex) runs the bluetooth Proxy for home assistant and has a loop time of 30ms, the web interface is fast there.

Is I2S audio so heavy or is the ESP32-C3 not performant enough? Would a regular ESP32 or an ESP32-S3 be faster?

This is the YAML of the “fast” ESP32-C3 that only does the bluetooth proxy, loop time 30ms:

substitutions:
  display_name: bleproxy

esphome:
  name: ${display_name}
  name_add_mac_suffix: true
  project:
    name: "ble_proxy"
    version: "1.0.0"
  platformio_options:
    board_build.mcu: esp32c3
    board_build.variant: esp32c3  
  includes:
    # should contain single line: #include <esp_task_wdt.h>
    - wdt_include.h
  on_boot:
    then:
      - lambda: !lambda |-
          // increase watchdog timeout
          esp_task_wdt_init(90, false);

preferences:
    flash_write_interval: 10min      

esp32:
  variant: ESP32C3
  board: esp32dev
  framework:
    type: esp-idf
    sdkconfig_options:
      CONFIG_BT_BLE_50_FEATURES_SUPPORTED: y
      CONFIG_BT_BLE_42_FEATURES_SUPPORTED: y
      # CONFIG_COMPILER_OPTIMIZATION_PERF: y      
      CONFIG_ESP_TASK_WDT_TIMEOUT_S: "90"
      CONFIG_COMPILER_OPTIMIZATION_SIZE: y    


status_led:  
  pin: 
    number: GPIO8
    inverted: true

logger: 
  # logs:
  #   component: ERROR
  #level: INFO

debug:
  update_interval: 1s

api:

ota:
  safe_mode: true
  password: !secret ota_password  

wifi:
  ssid: !secret wifi_ssid
  password: !secret wifi_password
  fast_connect: true
  reboot_timeout: 60s

  ap:
    ssid: !secret ap_ssid
    password: !secret ap_pass

esp32_ble_tracker:
  # scan_parameters:
    # interval: 1100ms
    # window: 1100ms
    # active: false
    # continuous: false

bluetooth_proxy:
  active: true

sensor:
  - platform: internal_temperature
    name: "${display_name} Internal Temperature" 
    update_interval: 1s 
  - platform: wifi_signal
    name: "${display_name} WiFi Signal"
    update_interval: 1s 
  - platform: uptime
    name: "${display_name} uptime"
    update_interval: 1s 
  - platform: debug
    free:
      name: "${display_name} Heap Free"     
    block:
      name: "${display_name} Heap Max Block"
    loop_time:
      name: "${display_name} Loop Time"

text_sensor:
  - platform: version
    hide_timestamp: true
    name: "${display_name} ESPHome Version"
  - platform: wifi_info
    ip_address:
      name: "${display_name} IP Address"
      icon: mdi:wifi      
    ssid:
      name: "${display_name} Connected SSID"
      icon: mdi:wifi-strength-2
    dns_address:
      name: "${display_name} DNS Address"    
  - platform: debug
    device:
      name: "${display_name} Device Info"
    reset_reason:
      name: "${display_name} Reset Reason"

button:
  - platform: safe_mode
    name: "${display_name} (Safe Mode)"

web_server:
  auth:
    username: !secret web_username
    password: !secret web_password

time:
  - platform: sntp
    id: sntp_time
    timezone: Europe/Amsterdam
    servers:
     - 0.pool.ntp.org
     - 1.pool.ntp.org
     - 2.pool.ntp.org

This is the YAML of the decibel meter, the slow ESP32-C3 (6000+ms loop time):

substitutions:
  display_name: decibelMeter

esphome:
  name: ${display_name}
  name_add_mac_suffix: true
  project:
    name: "decibel_meter"
    version: "1.0.0"
  platformio_options:
    board_build.mcu: esp32c3
    board_build.variant: esp32c3  
  includes:
    # should contain single line: #include <esp_task_wdt.h>
    - wdt_include.h
  on_boot:
    then:
      - lambda: !lambda |-
          // increase watchdog timeout
          esp_task_wdt_init(90, false);

external_components:
  - source: github://stas-sl/esphome-sound-level-meter  # add @tag if you want to use a specific version (e.g @v1.0.0)


esp32:
  variant: ESP32C3
  board: esp32dev
  framework:
    type: esp-idf
    sdkconfig_options:
      CONFIG_BT_BLE_50_FEATURES_SUPPORTED: y
      CONFIG_BT_BLE_42_FEATURES_SUPPORTED: y
      CONFIG_COMPILER_OPTIMIZATION_PERF: y      
      CONFIG_ESP_TASK_WDT_TIMEOUT_S: "90"


status_led:  
  pin: 
    number: GPIO8
    inverted: true

logger: 
  logs:
    component: ERROR # filters out these logs: Component api took a long time for an operation (424 ms).
  #level: INFO

debug:
  update_interval: 120s

api:

ota:
  safe_mode: true
  password: !secret ota_password  

wifi:
  ssid: !secret wifi_ssid
  password: !secret wifi_password
  fast_connect: true
  reboot_timeout: 60s

  ap:
    ssid: !secret ap_ssid
    password: !secret ap_pass

i2s:
  bck_pin: 4
  ws_pin: 5
  din_pin: 6
  sample_rate: 48000            # default: 48000
  bits_per_sample: 32           # default: 32
  dma_buf_count: 8              # default: 8
  dma_buf_len: 256              # default: 256
  use_apll: true                # default: false

  # right shift samples.
  # for example if mic has 24 bit resolution, and
  # i2s configured as 32 bits, then audio data will be aligned left (MSB)
  # and LSB will be padded with zeros, so you might want to shift them right by 8 bits
  bits_shift: 8                 # default: 0

sound_level_meter:
  id: sound_level_meter1

  # update_interval specifies over which interval to aggregate audio data
  # you can specify default update_interval on top level, but you can also override
  # it further by specifying it on sensor level
  update_interval: 5s           # default: 60s

  # you can disable (turn off) component by default (on boot)
  # and turn it on later when needed via sound_level_meter.turn_on/toggle actions;
  # when used with switch it might conflict/being overriden by
  # switch state restoration logic, so you have to either disable it in
  # switch config and then is_on property here will have effect, 
  # or completely rely on switch state restoration/initialization and 
  # any value set here will be ignored
  is_on: true                   # default: true

  # buffer_size is in samples (not bytes), so for float data type
  # number of bytes will be buffer_size * 4
  buffer_size: 1024             # default: 1024

  # ignore audio data at startup for this long
  warmup_interval: 500ms        # default: 500ms

  # audio processing runs in a separate task, you can change its settings below
  task_stack_size: 4096         # default: 4096
  task_priority: 2              # default: 2
  task_core: 1               # default: 1

  # see your mic datasheet to find sensitivity and reference SPL.
  # those are used to convert dB FS to db SPL
  mic_sensitivity: -26dB        # default: empty
  mic_sensitivity_ref: 94dB     # default: empty
  # additional offset if needed
  offset: 0dB                   # default: empty


  groups:
    - filters:
        # for now only SOS filter type is supported, see math/filter-design.ipynb
        # to learn how to create or convert other filter types to SOS
        - type: sos
          coeffs:
            # INMP441:
            #      b0            b1           b2          a1            a2
            - [ 1.0019784 , -1.9908513  , 0.9889158 , -1.9951786  , 0.99518436]

      # nested groups
      groups:
        # group 1.1 (no weighting)
        - sensors:
            # 'eq' type sensor calculates Leq (average) sound level over specified period
            - type: eq
              name: ${display_name}_LZeq_1min
              id: ${display_name}_LZeq_1min
              unit_of_measurement: dB(Z)

            # 'max' sensor type calculates Lmax with specified window_size.
            # for example, if update_interval is 60s and window_size is 1s
            # then it will calculate 60 Leq values for each second of audio data
            # and the result will be max of them
            - type: max
              name: ${display_name} LZmax_1s_1min
              id: ${display_name}_LZmax_1s_1min
              window_size: 1s
              unit_of_measurement: dB(Z)

            # same as 'max', but 'min'
            - type: min
              name: ${display_name} LZmin_1s_1min
              id: ${display_name}_LZmin_1s_1min
              window_size: 1s
              unit_of_measurement: dB(Z)

            # it finds max single sample over whole update_interval
            - type: peak
              name: ${display_name} LZpeak_1min
              id: ${display_name}_LZpeak_1min
              unit_of_measurement: dB(Z)

        # group 1.2 (A-weighting)
        - filters:
            # for now only SOS filter type is supported, see math/filter-design.ipynb
            # to learn how to create or convert other filter types to SOS
            - type: sos
              coeffs:
                # A-weighting:
                #       b0           b1            b2             a1            a2
                - [ 0.16999495 ,  0.741029   ,  0.52548885 , -0.11321865 , -0.056549273]
                - [ 1.         , -2.00027    ,  1.0002706  , -0.03433284 , -0.79215795 ]
                - [ 1.         , -0.709303   , -0.29071867 , -1.9822421  ,  0.9822986  ]
          sensors:
            - type: eq
              name: ${display_name} LAeq_1min_1s
              id: ${display_name}_LAeq_1min_1s
              unit_of_measurement: dB(A)
              update_interval: 1s
            - type: eq
              name: ${display_name} LAeq_1min
              id: ${display_name}_LAeq_1min
              unit_of_measurement: dB(A)
            - type: max
              name: ${display_name} LAmax_1s_1min
              id: ${display_name}_LAmax_1s_1min
              window_size: 1s
              unit_of_measurement: dB(A)
            - type: min
              name: ${display_name} LAmin_1s_1min
              id: ${display_name}_LAmin_1s_1min
              window_size: 1s
              unit_of_measurement: dB(A)
            - type: peak
              name: ${display_name} LApeak_1min
              id: ${display_name}_LApeak_1min
              unit_of_measurement: dB(A)

        # group 1.3 (C-weighting)
        - filters:
            # for now only SOS filter type is supported, see math/filter-design.ipynb
            # to learn how to create or convert other filter types to SOS
            - type: sos
              coeffs:
                # C-weighting:
                #       b0             b1             b2             a1             a2
                - [-0.49651518  , -0.12296628  , -0.0076134163, -0.37165618   , 0.03453208  ]
                - [ 1.          ,  1.3294908   ,  0.44188643  ,  1.2312505    , 0.37899444  ]
                - [ 1.          , -2.          ,  1.          , -1.9946145    , 0.9946217   ]
          sensors:
            - type: eq
              name: ${display_name} 4LCeq_1min
              id: ${display_name}_LCeq_1min
              unit_of_measurement: dB(C)
            - type: max
              name: ${display_name} LCmax_1s_1min
              id: ${display_name}_LCmax_1s_1min
              window_size: 1s
              unit_of_measurement: dB(C)
            - type: min
              name: ${display_name} LCmin_1s_1min
              id: ${display_name}_LCmin_1s_1min
              window_size: 1s
              unit_of_measurement: dB(C)
            - type: peak
              name: ${display_name} LCpeak_1min
              id: ${display_name}_LCpeak_1min
              unit_of_measurement: dB(C)

# automation
# available actions:
#   - sound_level_meter.turn_on
#   - sound_level_meter.turn_off
#   - sound_level_meter.toggle
switch:
  - platform: template
    name: "${display_name} Sound Level Meter Switch"
    # if you want is_on property on component to have effect, then set
    # restore_mode to DISABLED, or alternatively you can use other modes
    # (more on them in esphome docs), then is_on property on the component will
    # be overriden by the switch
    restore_mode: DISABLED # ALWAYS_OFF | ALWAYS_ON | RESTORE_DEFAULT_OFF | RESTORE_DEFAULT_ON
    lambda: |-
      return id(sound_level_meter1).is_on();
    turn_on_action:
      - sound_level_meter.turn_on
    turn_off_action:
      - sound_level_meter.turn_off
  - platform: restart
    name: "${display_name} Restart"      

sensor:
  - platform: internal_temperature
    name: "${display_name} Internal Temperature" 
    update_interval: 300s 
  - platform: wifi_signal
    name: "${display_name} WiFi Signal"
    update_interval: 300s 
  - platform: uptime
    name: "${display_name} uptime"
  - platform: debug
    free:
      name: "${display_name} Heap Free"   
    block:
      name: "${display_name} Heap Max Block"
    loop_time:
      name: "${display_name} Loop Time"

text_sensor:
  - platform: version
    hide_timestamp: true
    name: "${display_name} ESPHome Version"
  - platform: wifi_info
    ip_address:
      name: "${display_name} IP Address"
      icon: mdi:wifi      
    ssid:
      name: "${display_name} Connected SSID"
      icon: mdi:wifi-strength-2
    dns_address:
      name: "${display_name} DNS Address"
  - platform: debug
    device:
      name: "${display_name} Device Info"
    reset_reason:
      name: "${display_name} Reset Reason"

button:
  - platform: safe_mode
    name: "${display_name} (Safe Mode)"

web_server:
  auth:
    username: !secret web_username
    password: !secret web_password

time:
  - platform: sntp
    id: sntp_time
    timezone: Europe/Amsterdam
    servers:
     - 0.pool.ntp.org
     - 1.pool.ntp.org
     - 2.pool.ntp.org
2

I can’t comment much on low level performance details but I did notice the docs suggests loop time should be about 16ms.

It seems like a pretty demanding application so it wouldn’t surprise me if beefing up to an esp32 or S3 resolves things.

Can’t offer much more as above my pay grade;)

1 Like
3

Thanks for the documentation link. 16ms is quite a bit faster than the times I get.

I do notice after some trial and error that setting the log level to NONE shaves of about 2000ms, then the loop time is between 3100 and 4000ms.

4

After some more tweaking, disabling OTA and the built in web server shaved off some more seconds, now the loop timer is around 1200 to 2000 seconds.