Everything New In Home Assistant 2023.5!
HTML-код
- Опубликовано: 9 сен 2024
- Home Assistant 2023.5 has arrived - with new loads of new local voice control and assistant functionality as part of year of the voice!
If you'd like to support me and future projects:
Patreon:
/ everythingsmarthome
Home Assistant Release Notes:
Pending a few hours 👀
Buy Me a Coffee:
www.buymeacoff...
Written Article:
everythingsmar...
EVERYWHERE YOU CAN FIND ME:
Everything Smart Home Community Discord:
discord.everyt...
Twitter: / everysmarthome (@EverySmartHome)
Instagram: / everythingsmarthome
Facebook: / everythingsmarthomeoff...
#homeassistant #smarthome
________________________
Some of the links above are affiliate links, where I earn a small commission if you click on the link and purchase an item. You are not obligated to do so, but it does help fund these videos in hopes of bringing value to you!
I would love to see a video on enabling voice commands with home assistant
Okay gotcha! 🙏🏻
It would be terrific if, on top of voice commands, we could give our systems custom names. I'd definitely take "HAL9000" 😅
ruclips.net/video/TcNDiq_sM8A/видео.html please enable automatic translation
@@latrompetasonara just as long as you don't mind being called Dave in the future 😜
Also cover off some of the pitfalls. IE can you use it locally if you already have SSL enabled on HA, can you use google speakers locally?
Yes please, I'd love to see a guide on how to set this up.
Got it thanks!
The link to the article goes to the 2023.4 article 😉
What is the esp32 device with microphone you are using? Looks nice and compact
It's called the M5Stack Echo!
Yes please. A video on the esp32 and setup please
🙌🏻
My favourite new change is the one I developed. It is my first core addition. The Synology photos support
Amazing job, well done! 🙏🏻
Bedankt Lode, nicely done! I also have Synology photos so I'd be interested in knowing for which purposes you use this new feature yourself?
Any recommendations for which of those push-button mic/speaker devices are best? Thanks
Maybe we can get a video showing how to have multiple voice assistants for different things. For example one local assistant for controlling the home and one (Google or Alexa) for answering questions or one with openAI to just chat with.
"just chat with" i was literally thinking about this today.
i'm single. it would be interesting to just have a random conversation with ChatGPT. lol
Are there resources to guide building an ESP32 voice assistant device as you are using in the video?
Where do i find one of those little esp microphone/speaker/button things?
Yes!!! Where??
@@Jeppedy yes where????
It's called the M5Stack Echo, stock is difficult to find right now unfortunately
The new Android TV integration that replaces the (renamed) ADB bridge is quite nice. Anyone who has some Android TV devices that work with the Google TV app should definitely look into controlling their device using it's new remote service, it works much better than using the ADB.
Agreed!
Voice [recognition and synthesis] and AI are fine if *100% local* / non-cloud. Yes, it's perfectly possible to make that happen (case in point: Frigate AI NVR add-on, using Coral M.2 or USB AI accelerator chip)
I want to know more about openai. Looking forward to it.
So close. Can’t wait for the configurable always listening voice trigger feature. I’d like to see some DIY videos on how to build good to great speaker and Alexa/Google 360 microphone arrays. Also, how to integrate multiple of these and add the smarts where based on volume strength of the command, the system can figure out what room your in. Maybe 3 or more speaker/mic units hear you trigger a voice command, but one hears it louder than the others and assumes you are talking to it and the rest ignore the trigger word and command. For example, I’m in office and say “Computer, turn on lights”. I don’t have to tell it I’m in the office as that unit had heard my command the loudest and assumes I mean turn on the lights in just the office and not the whole house.
Just a couple small requests then 😅 haha no those would be cool, think there is a way to go!
where can I get those ESP microphone/speakers?
They are the M5Stack Atom, sold out right now unfortunately due to the announcement
@@EverythingSmartHome They're still available on a German webshop It seems. A little higher price tho
What button/mic combo are you using in the video to trigger voice commands?
Making a video on integrating voice makes sense only when the year of voice ends, and most of the features especially wake words, are developed. That way the tutorial would not be outdated in a few months. BTW, you have the best HA tutorials by far 🙂
Thanks, appreciate it!
A guide on activating the voice assistant would be super useful, thank you for doing that video in advance! :)
I'd love and am waiting for the assist function to be enabled as a call service or some other way to create a custom button on your dashboard which activates the assist functionality.
Ok, well I set everything up, but you can't use the Assist mic unless you are over https. I am using NGINX Home Assistant SSL proxy so that I don't have to use SSL inside of my network (since HA App won't work with an untrusted SSL cert).
Excellent - Thanks again for another great update
Would love to see a full guide on all the new assist features!
Thanks Michael!
Very cool.
I'd love to have a battery-powered dongle around my neck wit PTT voice assistant. Maybe using the M5. I've wanted it for a while because I use voice assistant in a school. The children tend to hijack it and I'd love it to only work when a button is pushed.
I think I'm going to set up the old school phone one, though. That would be cool to pick up a phone, give commands and put it down again.
Yaaaaaaas! Please do a video on the new voice commands and functionality. Love your videos, super helpful, keep up the great work!
Thank, appreciate it!
Would have been real fun, if they wouldn't have deleted all configurations for Alexa. That costs me 2hrs to redo the whole configuration, as some entities weren't available for Alexa anymore and it was a pain in the but to find them.
It is already really anyoing that with every big upgrade they reset the menu bar, something I can't understand. Who the f... thinks it is a good idea to reset that every time?
The VOIP feature sounds good, but when I need to press a button first or unhook a phone, I can do it as fast with my tablet on the wall or even my smartphone. The ability to setup an activation code (yes code, not word, perhaps someone would like to have more than one word to reduce the false activation. A single word is far too prune to fail like Alexa or the other hard coded words for it.) is in my opinion necessary to make that functionality interesting. Before that it is in my eyes only a nice thing to play around with, but nothing really useful.
i think the main application of this realistically is the integration of this with your current smart speakers/mics in the house. It would be good to see a setup tutorial if that is possible and if so is it possible to do with and without nabu casa setup for Alexa/Google?
I like the voip idea, a used conference speaker would be perfect for that.
One slight disappointment is that I *just* got rhasspy working pretty well. So I'll probably eventually move over to the new built in voice stuff, but until their wake words work I'll have to leave it. (I'm using porcupine for wake word, kaldi for speech recognition, and a coqui-server for text to speech)
Since the developer of rhasspy is also the one who made these HA integrations, it's also a big step on the way to rhasspy 3. Piper is even on the rhasspy repository. So hopefully the migrations will be painless :)
Yeah I think both projects will benefit from each other!
Thank you !
A fellow Fouad fan I see! Love the shirt.
I turned on the Voice Assistant, but it's a bummer it can only run over https, right after I just set everything up with nginx so I could have http control at home 😅
What is the ESP you were using for the PTT? Looks super tiny!
M5Stack Echo
To do the voice thing would be cool to have a keyboard shortcut on my pc (with HASS agent for example) and trigger the mic on my bluetooth headset and do the command... or a button widget on my phone and use the mic on the phone to process that command. I don't have mics lying around the house...
Can you do a custom built esp device with a mic? And another one with a speaker? Or do they need to be the same device?
Would be even better if the speaker output could be on integrated speakers used for tts already.
I use the homekit integration and Siri just works, no extra config or cloud needed compared to Alexa. I use iPhone, watch, and HomePod mini. Nice thing is I don't have to specify entities to include for Siri voice, but instead all the entities in HomeKit automatically just work with Siri.
What, how. Ever since an update, Siri on homepod does not recognize my voice, forcing me unlock my iphone for most tasks, disconnecting from wifi all the time. I had big hopes for Siri, but it honestly sucks in my experience.
Yes please an in depth tutorial
Great intro video! Of course an in depth follow up would be very nice. Also a quick question to anyone that tried that already. it seems that I am only able to setup one language for Piper and Whisper (unless I use whisper's "auto" mode). Any pointers to how to create two local pipelines for two different languages? (both for TTS and STT)
Thanks a lot!
How long before we get a mechanical keyboard with built in push to talk button? :D
Don't encourage them :) The time the developers spend at the keyboard may make a physical button seem reasonable.
😂
I was kind of hoping the VOIP integration could also route through audio from the reolink doorbell to a phone or have ChatGPT answer the door when I'm not there and send me a text on whose there and what they want. Or go even further and recognize the postman and tell him where to leave the package.
Waiting for three more features:
1) wakewords
2) custom commands so I can say "lumos" instead of "turn on the living room lights"
3) context; knowing which mic took the command so if I say "lumos" in a room, it turns on those lights, not the lights somewhere else.
had to remove the workday sensor from the configuration.yml now it's an integration...
Yes please. How-to would be great
I'd love to see a guide on how to set this up. Really interesting :)
Any info on the esp32 with button and speaker?
it's an Atom Echo mini esp device. Compatible with Rhasspy from what I've seen. Good luck finding one at the moment though
Yes thats right M5Stack Atom
Is there any way to interact with the Assist voice assistant outside of wiring up a VoIP phone or getting those ESP32 buttons? Wasted last weekend trying to get Rhasspy set up, but looks like HA will eventually replace that.
What is the sense in pressing a button and speak a command to turn on the light instead of turning the light on with the button directly?
What if you want that button to handle 10 different tasks? Turn lights on, set light colours, turn your heating on, set the temperature, get the temperature in another room, get the weather for the day, etc etc
@@EverythingSmartHome Would be great when it gets an "always listening" mode with keyword.
A video on enabling voice commands would be really nice :)
Hello, Great video as always. Could you tell me what is the device that you're using for voice commands please ? I'm also interesting for a dedicated video about this subject.
Thanks for the video Lewis!
Thanks for watching and commenting Jamie!
Someone should do a video of setting up aliases to solve for devices that don't work. And for what phrases can be used... I have a master bathroom night light... I named it 'nite lite', so assist fails because it gets night light... I have a rear porch table lamp that doesn't respond but works if I call out rear porch lights... Some help getting folks started would be nice...
All Aqara devices drop and become unavailable. Repairing the devices only lasts a little while and they drop again. So far the only solution I’ve been able to find is to revert back to 2023.4.
suck EP1's couldn't have had this built in..... now gonna have to be separate devices (more usb cables) or scrap fully functional, expensive ESP1's to buy the new stuff and reconfigure the whole house.
unlessss.... we can siphon power off of the EP1? and have new cases printed for voice-specific board attachments like a custom Everywhere Voice One (EV1) board. Wouldn't technically be a daughter board, but would eliminate extra cord runs.
Yes, a guide to the voice additions would be cool.
Keep doing what you doing, and do subtitles plz
Yes such a tutorial video would be awesome!
🙏🏻
Did you make that button or is that something that can be purchased?
Hey Lewis, Great content as always, I would like to know how you set up the M5 atom, thanks
Cool. Thank you! Do you know of an alternative for the M5Stack Echo? It's hopelessly sold out everywhere.
Push To Talk should be available on cell/smart phones. I'm sure I'm not the only one that has their smart phone next to them 99% of the time.😢
BTW the link to the written article is for last month, and I don't see a new one on your website... I want to know what that little cube you kept referring to as an ESP32 is...
M5Stack echo!
I would love to see a voice guide!
Where did you get that esp32 with the mic and speaker?
I run HA docker version no add-ons. Are we doomed?
Please do the voice setup videos
🙏🏻
With the year of the Voice, is there a simple way to make announcements on Alexa? I want to make announcements easy through routines or NodeRed, etc
Or has this become the "Year of Subscriptions"
Wake me up once we have wake word.... Good to see some hardware options coming to go with the software side, but it seems we're still a way off anything the rest of my family would use.
Is anyone else frutrated by absolutely crazy way of installing HA? Why cannot it be installed by distro's package manager? And by installing as docker i get limited and not able to install add-ons. KVM image also was struggling to get network running. Yes, i know i can buy Raspberry for that, but i don't want to.
that's awesome, think I need to dust off my old VOIP phone and try this out
I'm old enough that "dusting off an old VOIP phone" seems odd. I'm dusting off a mechanical phone.
Wonderful ... But my biggest question is: are you OK, Lewis??? You seem pretty tired - hope you are well, and just having too much fun staying up late, playing with toys ... :)
I'm doing good thanks for asking 🙏🏻💙
It's a coin toss on whether home assistant uodates are going to ruin my zigbee connections. Terrifies me to update lol
Yes, a 'How to" would be most helpful
🙏🏻
Great summary as always!
So, when I go to the Voice Assistants page, I don't see any of the configuration options, just a page that says 'Get the best assistant with Home Assistant Cloud'. Do you have to have a Home Assistant Cloud subscription to add other assistants? I currently have Alexa enabled using the Amazon Lambda approach in order to avoid the £6.50 a month fee.
You can only expose entities to Google Home and Alexa with nabu casa currently. However you can control the exposed entities through yaml config to both those platforms if you are doing it the free way, it's been in the docs for years 😅
@@EverythingSmartHome thanks for taking the time to reply, much appreciated 👍
And yes, I basically have been doing it via yaml for years and tend to do explicit includes to stop Alexa picking up hundreds of entities!!
I see the roborock integration is out, probably going to stay with valetudo for now given it does't require a login of any kind
And it's local so better!
When you beat the launch party by 2.5 hours 😂
Why look for wake word detection on HA over using Rhasspy? Rhasspy can currently accept rtp streams, so you can have your esp32 always streaming to a centralized Rhasspy instance listening for your wake word there. Having the HA devs spend cycles trying to catch up to what other open source devs have already done well seems... less than optimal. What do you think?
That's not the same thing though, they are trying to have the work word on the device so you don't have to constantly stream everything across the network all the time, the streaming would only happen after the work word is detected.
Having the creator of Rhasspy employed as the main Dev for voice on Home Assistant should mean that Rhasspy also benefits from this in the long run
@@EverythingSmartHome With $20 amazon echo devices clearly the problem handling the wake word on inexpensive hardware has been solved. What was proprietary becomes open source.
is there or will there be any form of use the home assistant voice for example on alexa or google devices?
Would very much doubt it
Multi-language households? I'm in. Or rather, I'll be in once I get something with more RAM, that poor pi3B can barely handle what i already got.
The Voice Assist feature is pretty dumb. I have a single lock in my smart home called "Front Door" and when I say "lock the front door" it doesn't understand. In fact it doesn't seem to understand 80% of my commands. The Assist feature needs natural language processing.
Thats one of he points of this update, you can change the model to a more accurate but more intensive model
I hope this doesn't phase out Rhasspy
So you can't use your Google speakers if you don't have nabucasa?
Nope you can, made a video on it years ago
@@EverythingSmartHome no i mean for the voice
Glad for the people that use voice commands but the year of the voice isn't the year for me.
I just don't like talking to "nobody". I have Google Assistant disabled on my smartphone, I don't have an smartspeakers, I do have an old Sonos Play 5 (an old one)
Am I the only one who's not exited about these updates?
You’re not alone 👍
I'm not a huge voice command person either, like you I don't really like talking to machines. I'm still excited though just to see advancements in HA for people who are super into voice and because it's something that's never really been done before so I love that it's HA doing it. I also try and give voice more of a chance since it's local and private now :)
So how do you update to latest version ?
Should be out within the hour
👍👍👍
When will this be available ?
Should be within the hour
Shame it broke my zigbee and messed up my energy recording, had to roll back
I have no interest in voice anything so 2023 will be a year of no upgrades for me!
Sorry but "unlike anything we've seen in voice assistants so far" is just not true and it makes me a bit sad to hear this 😢. For example Michael's own Rhasspy system and SEPIA have been doing it for a long time. I'm using SEPIA for over 2 years now doing the very same thing fully offline 😑.
If you make a video on enabling voice commands... I've heard that https is a must, otherwise it doesnt work. I followed your other video first ruclips.net/video/xXAwT9N-7Hw/видео.html
in which you make a connection with cloudflare, is it possible to include how to setup https too in that video based on the first movie so its compatible? thx
Really not interested in using voice. Not saying people who do are wrong, just making it known that there are some of us out here who hope that we’ll still be able to use all functionality going forward without having to speak commands.
It’s not replacing any existing functionality just adding the ability to use voice
Yep you will be all good, they wouldn't restrict functionality like that
Sticking with Rhasspy for now, no wake word is pointless, push to talk is as much effort as tapping a on\off button, pointless.
Insert thoughtful comment here
🧠
The whole year of voice is a dead-end direction..
who the f*ck want to say keywords and aliases to turn on or off anything?
In an ideal smarthome things should work WITHOUT user interaction..
but our world is not ideal.. so the second best case is: almost everything work automaticaly without user interactions, and the rest works with interaction over wall-tablet or mobile phone or buttons...
voice assistant things looks great for some geeks for a short period of time, but this method of using out smarthome is not convient in everyday living..
P.S.
Please let me know if we get some really useful updates.. until then, i stop updating my system...
Yes please, I'd love to see a guide on how to set this up.
🙏🏻
Voice commands and a guide on it would be amazing
Got it, thanks!