Real-time Speech Recognition in 15 minutes with AssemblyAI
Вставка
- Опубліковано 11 лис 2021
- Get your free speech-to-text API token 👇
www.assemblyai.com/?...
Transcribing in real-time is a super skill only court reporters can brag about. But luckily, we don’t need to learn how to type fast to get transcriptions of audio quickly. Thanks to Assembly AI’s Streaming Speech-to-Text model (previously real-time speech recognition), it is very simple to set up a python script that can listen for audio and turn it to text.
In this video, we will see how to create this script on Python with the help of pyaudio, web sockets and asynchronous functions. The app will have the power to listen to audio input through a microphone and display the transcription in real-time. We will integrate this code into a simple Streamlit application to showcase the real-time speech recognition with a touch of interactivity.
If you’d like to follow along, don’t forget to get your own AssemblyAI API token for free at assemblyai.com
You can find the code from this tutorial in this GitHub repository: github.com/misraturp/Real-tim...
Find the written form of this tutorial here: www.assemblyai.com/blog/real-...
AssemblyAI Streaming STT docs: www.assemblyai.com/docs/speec...
The accuracy and formatting is quite interesting here.
Thank you! - Mısra
@@AssemblyAIhi what if I want the input to be not from microphone and i want it from my speaker or laptop speaker how do I do it then.
Most instructional and useful video. Thank you.
You're very welcome!
Thanks for everything :)
that is sick! goodjob
Thank you Amine! - Mısra
Very sympathic and engaging presentation. Maybe the most beautiful side of Artificial Intelligence 😊
Thank you!
Is there a JavaScript version of this video please? Both service side and front end…
Thank you for this nice tutorial
Glad you liked it
Very kewl it works. However for the live service a $8 minimum is required. - but totally worth it
This is really awesome! I would prefer though to see the final result as a short highlights at the beginning of your videos before you go into the details of how to. But thanks anyway for the effort and the time! Hugs!
Thanks for the feedback! It's definitely a good idea to give an impression of the app that is being built. With the newer videos we do a preview at the beginning of the videos indeed. - Mısra
@@AssemblyAI 🤗🤗♥️♥️
How can I use this for transcribing particular chrome tabs for online meetings? Can you help me with that?
Very nice and easy setup! Took me ages to achieve the same with whisper. However, are you planning to support other languages, german, french or even a multilinugal model?
Can I ask how you were able to do it?
Awesome 👍
Thank you!
Incredible video. Would I be able to use chatGPT, as an intermediate, to correct the sentences based on vocabulary and grammar, and after that, get the response from that chatGPT "reviewed" sentences?
Yup i tried to make it but it worked only text cant use speak 😅
I've this installed and worked on windows. My question is that Real-time Speech Recognition only recognize english ? Does it support other languages ? Thank you.
How can I introduce um, ahs, and other filler words in to the Realtime transcription? I see it is an option for uploading an audio file
Could you please instruct me on how I can include this in my Angular project?
As of now, only the English language is supported in the live transcription feature. when can we expect more language support ?
when i tried to install pyaudio terminal displays this error: "Cannot open include file: 'Python.h': No such file or directory "
Does it also offer diarization in real time?
hi great works but what about smartphone ?
I want to implement a realtime app using voice recognition on python 3 / android 11 on my samsung s22. It's my understanding portaudio is NOT supported on Android 11. Is portaudio your only way to get to the mic?
Can this work in Unreal Engine 5
Can you please let me know how can I integrate this in my JAVA Maven project?
Hey Moncef, unfortunately I also don't have experience on that. -Mısra
can you pls let me know how to save that text
is it possible to transcribe in real-time in other languages, such as spanish?
I am here for Freddie.
Live transcript is paid version?
Assembly ai real time transcription doesn't have a free trial ?
Is it possible to make it faster ? The transcription takes a little time
Can we setup Bangla language for this system?
Is it possible to use this service in other languages apart from English?
im also trying to find an answer to this question
@@tiagofyhnesteves74 they answered to me privately. It's not possible yet. I switched to Azure cognitive services, which is multi-language.
@@GiulianoGolfieri I had taken it for granted that it was a multilingual service, a fucking morning's work wasted. Grazie della info, Giuliano
how to add custom ui?
I know this is a little old, but I get a 4002 error, what might cause that?
Did you ask ChatGPT?
I'm on windows, When I try to run it it does nothing, just creates a new line on the terminal. when I cloned the github repo, it was working, hmmm
Have you tried speaking while the code is running? It might be that you don't have a microphone connected to the computer.
can we speak any indic languages , will this endpoint will transcribe or not ?
Yes AssemblyAI's API supports Hindi Transcription, check out this tutorial: ua-cam.com/video/3WlNOCeyyjQ/v-deo.html
buut.. if i have no money to buy the pro key?😕
You can get started for free!
Hi. how can i change the lenguage to transcript in spanish?
Hey Angel, here is the documentation on transcribing in languages other than English. docs.assemblyai.com/walkthroughs#specifying-a-language
@@AssemblyAI But do different languages work with realtime transcription on this specific endpoint? The documentation, regarding changing the language, is not clear about this.
@@dirtydevil81 do you find a solution?
Should have mentioned cost to go to pro version.... why hide it?
does this work for song transcription ? in real time ?
Hey Siam, depending on the amount of background music and clarity of pronunciation you'd get varying levels of success with transcribing songs.
is assemblyAI support arabic language with vowelization?
We are launching support for Arabic in late January!
Aha Türk !
Ahahahha
THIS IS ONLY AVAILABLE ON MAC/LINUX, RIGHT?
No, it is available independent of the operating system.
received 4001 (private use) Not authorized; then sent 4001 (private use) Not authorized
Türk müsün?
When i added this line
if json.loads(result_str)['message_type'] == 'FinalTranscirpt':
it wouldnt transcript anything anymore