--- title: "Let's use Google Speech to control the device" url: "https://maker.wiznet.io/louis_m/projects/lets-use-google-speech-to-control-the-device/" markdown_url: "https://maker.wiznet.io/louis_m/projects/lets-use-google-speech-to-control-the-device/md" type: "WCC: WIZnet Created Content" author: "Louis" author_url: "https://maker.wiznet.io/louis_m/" original_author: "Louis" original_url: "https://www.hackster.io/louis_m/let-s-use-google-speech-to-control-the-device-bd58b3" published: "2022-12-15" language: "en" hardware: ["Raspberry Pi 4 Model B", "WIZnet W5100S", "Adafruit NeoPixel Ring: WS2812 5050 RGB LED"] likes: 8 views: 706 comments: 2 source: "WIZnet Makers (https://maker.wiznet.io/)" --- # Let's use Google Speech to control the device > Use Google Speech-to-Text API to convert voice to text and then control the device. Original author: Louis (source: https://www.hackster.io/louis_m/let-s-use-google-speech-to-control-the-device-bd58b3) ## Components - **Raspberry Pi 4 Model B** x 1 ([docs](https://datasheets.raspberrypi.com/rpi4/raspberry-pi-4-product-brief.pdf)) - **WIZnet W5100S** x 1 ([docs](https://docs.wiznet.io/Product/Chip/Ethernet/W5100S)) - **Adafruit NeoPixel Ring: WS2812 5050 RGB LED** x 1 ([docs](https://www.adafruit.com/product/1586)) ## Documents and links - [code](https://maker.wiznet.io/_lib/download.asp?path=project/3921&downfile=wiznet%5Fspeak%2Epy) (code) - [code_pico](https://maker.wiznet.io/_lib/download.asp?path=project/3921&downfile=code%2Epy) (code) ## Article ### **OverviewGoogle's Speech-to-Text** is a simple API based on machine learning that applies a powerful neural network model. Convert your voice to a string using the google.cloud.speech library. The converted string communicates with the **W5100S** via MQTT and controls the **NeoPixel**. ![](https://maker.wiznet.io/upload/ckeditor5/298547281%5F1671136746%2Egif) ### **Prepare** > **Google** Speech-to-Text Speech-to-Text is an API that is powered by Google's artificial intelligence (AI) technology. You send your audio data to Speech-to-Text, then receive a text transcription of your audio data in response. For more information about how to construct a Speech-to-Text request, see the [requests](https://cloud.google.com/speech-to-text/docs/speech-to-text-requests)page. Enable Speech-to-Text on a project. - Make sure billing is enabled for Speech-to-Text. - Make sure your project has at least one service account. - Download a service account credential key. ![](https://hackster.imgix.net/uploads/attachments/1536567/image_31VH2FiEV2.png?auto=compress%2Cformat&w=740&h=555&fit=max) > **Connect Raspberry Pi** ![](https://hackster.imgix.net/uploads/attachments/1536518/image_PNyVJQmVff.png?auto=compress%2Cformat&w=740&h=555&fit=max) ![](https://hackster.imgix.net/uploads/attachments/1536540/image_c4R1Ncqd3E.png?auto=compress%2Cformat&w=740&h=555&fit=max) Check out our step-by-step guide: [How to set up your Raspberry Pi 4 for the first time](https://magpi.raspberrypi.com/books/beginners-guide-4th-ed) > **Connect W5100S-EVB-Pico** ![](https://hackster.imgix.net/uploads/attachments/1536521/image_rxHCXipAjC.png?auto=compress%2Cformat&w=740&h=555&fit=max) W5100S-EVB-Pico is a microcontroller evaluation board based on the Raspberry Pi RP2040 and fully hardwired TCP/IP controller W5100S – and basically works the same as Raspberry Pi Pico board but with additional Ethernet via W5100S. ### **Setup Google API service** ![](https://hackster.imgix.net/uploads/attachments/1536574/image_Ryj7NWVxaB.png?auto=compress%2Cformat&w=740&h=555&fit=max) In this tutorial, you'll learn how to use the Speech-to-Text V1 API from a Python application by following these steps: Select a Google Cloud project and enable the Speech-to-Text API. - Select a Google Cloud project and enable the Speech-to-Text API. Set up authentication by creating a service account and downloading a key. - Set up authentication by creating a service account and downloading a key. Create a sample application that transcribes an audio file to text. - Create a sample application that transcribes an audio file to text. Estimated time to complete: To get started, click [**Start**](https://console.cloud.google.com/freetrial/signup/tos?_ga=2.193169323.722706232.1670989453-1493226483.1670219256&_gac=1.19213836.1670310478.EAIaIQobChMIpYTSyePh-wIVCqqWCh1tHwffEAAYASAAEgIELPD_BwE&facet_utm_source=google&facet_utm_campaign=%28organic%29&facet_utm_medium=organic&facet_url=https:%2F%2Fcloud.google.com%2Fspeech-to-text&facet_id_list=%5B39300012,%2039300022,%2039300118,%2039300191,%2039300195,%2039300241,%2039300319,%2039300320,%2039300325,%2039300333,%2039300346,%2039300354,%2039300364,%2039300373,%2039300412,%2039300422,%2039300434%5D). > **Creating a project on the Google Cloud platform<step-1>Menu -> IAM & Admin -> Manage resources** ![](https://hackster.imgix.net/uploads/attachments/1537149/image_88Y7306VBj.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-2>** Go to the page **[Manage Resources]** and **[+CREATE PROJECT]** select. ![](https://hackster.imgix.net/uploads/attachments/1537105/image_Angg9jmXkO.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-3>** Enter a name for the project and press **CREATE.** This creates a project. ![](https://hackster.imgix.net/uploads/attachments/1537148/image_dfMcQD6ONH.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-4>Menu -> APIs & servuce-> Library** ![](https://hackster.imgix.net/uploads/attachments/1537146/image_9KmizvlAUF.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-5>[Cloud Speech-to-Text API]** Find and click. **[MANAGE]** click ![](https://hackster.imgix.net/uploads/attachments/1537142/image_7oLfuR2jA7.png?auto=compress%2Cformat&w=740&h=555&fit=max) ![](https://hackster.imgix.net/uploads/attachments/1537144/image_1GZIjzrgMb.png?auto=compress%2Cformat&w=740&h=555&fit=max) > **Create User Credentials<step-1>[+CREATE CREDENTIALS]** select and **Service account** click ![](https://hackster.imgix.net/uploads/attachments/1537154/image_jqgBGF63wu.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-2>** Enter a name for the account and press **CREATE** ![](https://hackster.imgix.net/uploads/attachments/1537158/image_WkK3CsSP6L.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-3>** Verify that the service account has been created as follows: ![](https://hackster.imgix.net/uploads/attachments/1537159/image_enIrhUmbpe.png?auto=compress%2Cformat&w=740&h=555&fit=max) > **Create APIkey<step-1>Manage service accounts** click. Enter the service account you just created. ![](https://hackster.imgix.net/uploads/attachments/1537162/image_XpvoQaBg3d.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-2>KEYS -> ADD KEY -> Create new key** ![](https://hackster.imgix.net/uploads/attachments/1537163/image_4O7v22krnW.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-3>** Press Create while selected as the **JSON** format. ![](https://hackster.imgix.net/uploads/attachments/1537164/image_PhgnOFVMWI.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-4>** Move the created **JSON file** to the RPi directory. ![](https://hackster.imgix.net/uploads/attachments/1537166/image_tB9IC6kl7K.png?auto=compress%2Cformat&w=740&h=555&fit=max) **<step-5>** Now it's time to **export the JSON** file. ```plaintext example) pi@raspberry:~ $ export GOOGLE_APPLICATION_CREDENTIALS="KEY_STRING" ``` ![](https://hackster.imgix.net/uploads/attachments/1537168/image_cUsefAWJMM.png?auto=compress%2Cformat&w=740&h=555&fit=max) ### **Getting Started** > **Raspberry Pi Setup** - Connect the microphone to the RPi USB port and download the Google.spech library for the Google Speech-to-Text service. ```plaintext pi@raspberry:~ $ pip install --upgrade google-cloud-speech or pi@raspberry:~ $ python3 -m install --upgrade google-cloud-speech ``` ![](https://hackster.imgix.net/uploads/attachments/1537171/image_y42zPs9wSp.png?auto=compress%2Cformat&w=740&h=555&fit=max) - Next, download and run mosquitto because you will control the device through mqtt. To install the Mosquitto Broker enter these next commands: ```plaintext pi@raspberry:~ $ sudo apt install -y mosquitto mosquitto-clients ``` To make Mosquitto auto start when the Raspberry Pi boots, you need to run the following command (this means that the Mosquitto broker will automatically start when the Raspberry Pi starts :) ```plaintext pi@raspberry:~ $ sudo systemctl enable mosquitto.service ``` Now, test the installation by running the following command: ```plaintext pi@raspberry:~ $ mosquitto -v ``` ![](https://hackster.imgix.net/uploads/attachments/1537175/image_7WxVaFFYU2.png?auto=compress%2Cformat&w=740&h=555&fit=max) - We will issue the data converted from Google STT to Text again in MQTT, so we will add the following command to the code. The Speech-to-Text and MQTT integrated source code is attached below. ```plaintext mosquitto_pub -d -h mqtt_host -t mqtt_topic -m mqtt_msg ``` > **WIZnet W5100S-EVB-Pico setup** - **Installing CircuitPython** Install `**CircuitPython**` on Raspberry Pi Pico by referring to the link above. - **Download the below uf2 file link on your computer** CircuitPython 7.3.3 link - [DOWNLOAD](https://downloads.circuitpython.org/bin/raspberry_pi_pico/en_US/adafruit-circuitpython-raspberry_pi_pico-en_US-7.3.3.uf2) - **Setup WIZnet Ethernet Library** Copy the [**WIZnet library**](https://github.com/Wiznet/RP2040-HAT-CircuitPython)[In](https://github.com/Wiznet/RP2040-HAT-CircuitPython)to the lib folder inside the Raspberry Pi Pico. Before continuing, make sure your board's lib folder has at least the following files and folders copied over: https://github.com/Wiznet/RP2040-HAT-CircuitPython ```plaintext adafruit_bus_device adafruit_wiznet5k adafrui_minimqtt ``` ![](https://hackster.imgix.net/uploads/attachments/1537180/image_T8Qu48gc5Z.png?auto=compress%2Cformat&w=740&h=555&fit=max) - **Setup NeoPixel Library** Copy the [CircuitPython NeoPixel](https://learn.adafruit.com/circuitpython-essentials/circuitpython-neopixel) into the lib folder inside the Raspberry Pi Pico. ```plaintext adafruit_pixelbuf.mpy neopixel.mpy ``` ![](https://hackster.imgix.net/uploads/attachments/1537199/image_ABXt0qF9we.png?auto=compress%2Cformat&w=740&h=555&fit=max) - **Install Programs** The following serial terminal program is required for **MQTT Publish** test, download and install from below links. ✔[Tera Term](https://osdn.net/projects/ttssh2/releases/) Before this, you should find the Raspberry Pi Pico COM port number from "**Device Manager**" on your PC. ![](https://hackster.imgix.net/uploads/attachments/1537185/image_mKSU46vpf5.png?auto=compress%2Cformat&w=740&h=555&fit=max) > **Let's start Speech-to-Text control device** - Paste code into **code.py **on your Raspberry Pi Pico and Save it. Press "**Ctrl+D**" on the terminal window. you can see the messages as follows ![](https://hackster.imgix.net/uploads/attachments/1537209/image_T8o8ZHfl4M.png?auto=compress%2Cformat&w=740&h=555&fit=max) - Execute wiznet_speak.py Python code. I have attached the sauce below ![](https://hackster.imgix.net/uploads/attachments/1537210/image_8l0w2eVmTo.png?auto=compress%2Cformat&w=740&h=555&fit=max) - Let's shout "Hi Pico" into the mic ![](https://hackster.imgix.net/uploads/attachments/1537220/image_hkjG8oYgeQ.png?auto=compress%2Cformat&w=740&h=555&fit=max) > **Rainbow Color** ![](https://hackster.imgix.net/uploads/attachments/1537222/rainbow_xBFA4EOfUr.jpg?auto=compress%2Cformat&w=740&h=555&fit=max) - Let's put the color we want into the microphone ![](https://hackster.imgix.net/uploads/attachments/1537223/image_pMZAppwprr.png?auto=compress%2Cformat&w=740&h=555&fit=max) > **Color LED on** ![](https://hackster.imgix.net/uploads/attachments/1537224/image_0L8AOwow93.png?auto=compress%2Cformat&w=740&h=555&fit=max) --- Source: https://maker.wiznet.io/louis_m/projects/lets-use-google-speech-to-control-the-device/