After upgrading HA to the latest release, it works!
did anyone get a display up and running?
Dan333! You’re a gentleman and a scholar! After fighting the battle of battles getting a RPI zero 2w with an audio hat programmed-connected-working with Home assistant (literally took me days), I had this bad boy up and running within 2 hours of learning how to flash an esp32 board. Amazing work!
Captured audio from this device is low. anyone else notice this?
Any ideas to improve input volume?
I place unit into debug
assist_pipeline:
debug_recording_dir: /share/assist_pipeline
when listening to voice command capture this audio input is very low. I compare with HAV from nabu and difference is big. The command captured by waveshare is barely audible at full volume. I try adjust gain in UI and also by yaml.
I think at 4’ audio is captured OK at just above normal speaking level.
At 7’ I feel it is necessary to, not quite yell, but more forcefully speak. In quite room it is basically yelling. It would be weird to do for every use or if I saw someone else make command in this manner.
Most ASRs have preprocessing, so don’t worry about the amplitude.
Normalize the volume level of recorded samples in an audio editor., and if the sound is not distorted, then everything is fine.
If you still want amplification at the capture stage, add a multiplier to the configuration.
microphone:
microphone: your_mic_id
gain_factor: 16
voice_assistant:
microphone:
microphone: mic
gain_factor: 20
micro_wake_word:
microphone:
microphone: mic
gain_factor: 10
made no difference. results same.
In many case the audio is so low it can barely be distinguised from no audio. its is unintelligible to human ear. I think for piper I just get incorrect command but with speech-to-phrase command are not recognized at all.
EDIT
I forgot, gain_factor for voice assistant actually failed. is “auto_gain” same adjustment. I used that instead.
Hi! Great project!! I wanted to know if it was possible to hear the voice output from Home Assistant when I send commands.
For example, when I tell my Assistant “Turn on the desk lamp,” it responds, “I turned on the desk lamp.” This happens when I send voice commands from my smartphone.
Is it possible to also receive audio from the Waveshare-s3-audio when I command it with wakewords? From the logs, I see that the response from HA is received.
Thanks for your great work!
I think what you want is seperate from waveshare device.
HA made change so command with obvious result does not give voice response. “turn on light” doesnt get response because you can see light turn on.
look at dubug to see if there is response generated. If it is there will be a way to play it back but my guess there is not response
OK, after some testing this morning, I figured out what’s happening, and I don’t think it’s a problem, but I’m writing it here for future reference:
- If I send a direct power command to a device, the Waveshare performs the operation but responds with a sound.
- If I say something else, specifically to avoid being understood, or set a timer, it responds with its voice, as it should.
So, no problem for me, just knowing.
In the meantime, is there a specific (or configurable) phrase to find out how much time is left until a set timer ends?
Thanks for your support.
What STT provider are you using? Command may differ between them.
Speech to phrase STT use these commands
I installed this with the camera module and everything works great except my camera image is sideways and I have not been able to find any way to rotate the image 90 degrees. Has anyone else had this problem?
Did you try camera options below
vertical_flip (Optional, boolean): Whether to flip the image vertically. Defaults to true.
horizontal_mirror (Optional, boolean): Whether to mirror the image horizontally. Defaults to true.
When i am pressing manual download process is stucking at : “Library Manager: Installing kahrendt/ESPMicroSpeechFeatures @ 1.1.0
INFO Installing kahrendt/ESPMicroSpeechFeatures @ 1.1.0
Downloading [------------------------------------] 0%”
I’ve put together a complete ESPHome YAML for the ESP32-S3 Touch AMOLED 1.75" , including display, audio, and a Home Assistant voice assistant setup.
It’s hardware-specific but fully working, and might be useful as a reference (or starting point) for similar Waveshare / AMOLED S3 boards.
1 Like
Yes I did but if you look through the code he just basically sets both of these to false to rotate the image 180 degrees. I just need 90 degrees.
I’ve done the same for the Waveshare ESP32-S3 Touch LCD 1.83" and I created a component for the QMI8658 IMU that it uses: GitHub - Djelibeybi/esphome-qmi8658: ESPHome external component for the QMI8658C IMU found in Waveshare devices. I only use it to rotate the display so that it’s always upright, but it has some pretty nifty other features.
I just picked up a couple of these and they work great thanks to @Dan333 .
Anyone else think they work/look better upside down? Having the USB connector at the top just causes them to tip over frequently.
work better? maybe. it may be same or at least comparable
Look better? yes
USB at bottom better? yes
i took the rubber feet from speaker grill and placed them on the plexiglass and flipped it as you have in photo.
Unfortunately, cannot get sendspin working with latest version of esphome.
worked in previous versions with the sendspin components added, I think 2025.11, but cant get it working with 2025.12.5.
EDIT
I was able to get sendspin working
Wish I could get my atom echos working
Try cleaning the build files as the sendspin component was upgraded a few days ago to allow building with 12.5.