Hm, @homeassistant is now offering a bevy of different voices for TTS, including some that shout and some that whisper. I added some quick automation to change to a whispering voice after a specific time, but I should probably do that with a light sensor instead. That way it'll whisper when the lights are low, but speak normally if the lights are on full brightness. Seems like a logical approach.
can it be set to respond to noise levels? That's genuinely intuitive. If I have to shout my question it's probable I need it to shout back to be heard. If I whisper it's because I need to be quiet. If it shouts back instead and wakes the baby that speaker is meeting it's destiny with the patio slabs outside the window.
@Iwillyeah Hm, that's a good idea. The device I use (Home Assistant Voice PE) doesn't have an ambient noise sensor exposed. At least not yet, it is a Preview Edition, so that might show up later. That doesn't mean a 3rd party device couldn't be deployed to test noise levels. Of course Home Assistant is extensible to the point of being ridiculous, so it's theoretically possible to use a device to detect noise levels and adjust voice and volume accordingly.
well, I'm not really much of an ideas person, and even less of a solutions person for anything technical.
Still, I imagine it's actually a bit tricky to build something that knows the difference between a sound that is quiet and a sound that is fair away, so it can respond with the right volume. How do we do that with our hearing? Is it something do with having two ears, or can you trick our perceptions of loudness and distance if we can't see the sound source?
@Iwillyeah I think humans are pretty bad at that too. Of course, with an automated solution you can have multiple sensors scattered over a wider range. If those sensors are detecting volume and you're aware of where each sensor is in relation to the speaker, you can determine it's source of origin and the required volume to have a response heard. It's worth noting that I don't think Home Assistant will do that at present, but it's not something technically impossile.
that's very clever. Thanks for taking my suggestions seriously. Having a poorly sleeping baby was the greatest learning experience of my life, though mostly what I learned was how ANNOYING every single thing and living person in the entire world was.
@Iwillyeah Yea, I know what you mean. There was a month or so when my youngest was a baby and she refused to sleep unless she was being rocked in a rocking chair. We had to hold her and rock her to sleep every night. It felt like forever before we got her sleeping in a crib, and I seriously consider felonies if someone did something to mess with her when she was sleeping.
I was just playing with that literally minutes ago.
I’m setting up som Zigbee water leak sensors and I’m broadcasting “Sh*t! There’s a water leak! Check the [entity_id] sensor” on all my speakers (plus a phone notification).
It was certainly more urgent in the “scared” voice but:
1) it was only available in the US male voice and
2) it was only in the Google TTS. I prefer to use the local Piper TTS especially for important messages. It’s actually surprisingly good quality.
@ianjs It's smart to use local TTS when things are important, like water leaks. You wouldn't want to wake up up to your eyeballs because your WiFi was out. I use the Home Assistant Cloud for most of mine, but I can only get the original "Mycroft" from Piper these days. The closest I've found is "Thomas", but it also doesn't come with whispers and shouts. I change over to an American sounding accent that whispers after a certain time of the night.
Sollte Bundeskanzler Friedrich Merz (CDU) den israelischen Ministerpräsidenten Benjamin Netanjahu trotz des internationalen Haftbefehls in Deutschland empfangen, werde ich eine Anzeige inklusive Strafantrag gegen ihn stellen. Das Recht gilt für alle, auch für Kanzler! #EinfacheSprache
Zagreb Airport is flying high after clocking up a whopping 419,675 passengers in April alone – that’s a 13.6% jump compared to the same month last year.
Last night, President Rumen Radev submitted a proposal to the National Assembly for a referendum with the question: "Do you agree that Bulgaria should adopt
The EU destroyed Moldovan agriculture, and the integration process deepens the crisis, a Russian website writes. In fact, Moscow imposed embargoes on Moldovan
To provide the best experiences, we use technologies like cookies to store and/or access device information. Consenting to these technologies will allow us to
"CERN researchers achieved the feat by aiming beams of lead at each other, travelling at close to the speed of light. The ions occasionally glance past each other, rather than hit head on. When this happens, the intense electromagnetic field around an ion can create a pulse of energy that triggers an oncoming lead nucleus to eject three protons — turning it into gold."
A documentary about Malta’s street cats by husband-and-wife filmmakers Sarah Jayne Portelli and Ivan Malekin has been enjoyed by hundreds of filmgoers in
I'm never going to forget how last time sharing this clip of the neo-nazi Richard Spencer being punched in the face prompted SOMETHING to set the start & loop points of the clip to just after the knuckles had met their rightful target. For experimental value, here it is again: A nazi being punched in the face, with appropriate soundtrack from The Who's : Won't Get Fooled Again #punch_a_nazi #fcknzs
A court in Cyprus sentenced two Hungarian nationals to prison on Friday for brokering sales of properties in the Turkish-occupied north of the island seized
Iwillyeah
in reply to Mike Stone • • •Mike Stone
in reply to Iwillyeah • • •@Iwillyeah Hm, that's a good idea. The device I use (Home Assistant Voice PE) doesn't have an ambient noise sensor exposed. At least not yet, it is a Preview Edition, so that might show up later. That doesn't mean a 3rd party device couldn't be deployed to test noise levels. Of course Home Assistant is extensible to the point of being ridiculous, so it's theoretically possible to use a device to detect noise levels and adjust voice and volume accordingly.
@homeassistant
Iwillyeah
in reply to Mike Stone • • •well, I'm not really much of an ideas person, and even less of a solutions person for anything technical.
Still, I imagine it's actually a bit tricky to build something that knows the difference between a sound that is quiet and a sound that is fair away, so it can respond with the right volume. How do we do that with our hearing? Is it something do with having two ears, or can you trick our perceptions of loudness and distance if we can't see the sound source?
Mike Stone
in reply to Iwillyeah • • •@Iwillyeah I think humans are pretty bad at that too. Of course, with an automated solution you can have multiple sensors scattered over a wider range. If those sensors are detecting volume and you're aware of where each sensor is in relation to the speaker, you can determine it's source of origin and the required volume to have a response heard. It's worth noting that I don't think Home Assistant will do that at present, but it's not something technically impossile.
@homeassistant
Iwillyeah
in reply to Mike Stone • • •Mike Stone
in reply to Iwillyeah • • •ianjs
in reply to Mike Stone • • •I was just playing with that literally minutes ago.
I’m setting up som Zigbee water leak sensors and I’m broadcasting “Sh*t! There’s a water leak! Check the [entity_id] sensor” on all my speakers (plus a phone notification).
It was certainly more urgent in the “scared” voice but:
1) it was only available in the US male voice and
2) it was only in the Google TTS. I prefer to use the local Piper TTS especially for important messages. It’s actually surprisingly good quality.
Mike Stone
in reply to ianjs • • •@ianjs It's smart to use local TTS when things are important, like water leaks. You wouldn't want to wake up up to your eyeballs because your WiFi was out. I use the Home Assistant Cloud for most of mine, but I can only get the original "Mycroft" from Piper these days. The closest I've found is "Thomas", but it also doesn't come with whispers and shouts. I change over to an American sounding accent that whispers after a certain time of the night.
@homeassistant