Voice Control

asked 2014-01-02 23:18:35 +0200

leon.anavi
1421 ●5 ●8 ●10

updated 2015-03-24 08:03:14 +0200

jaacoppi
591 ●5 ●16 ●16

Are there any plans for integration of voice control application for Sailfish OS?

I would like an app that recognizes speech and supports text to speech with features such as dialing a contact, searching Internet, writing a email or text message, updating status of a social network (facebook, twitter, LinkedIn), searching for directions using the Maps app showing a weather forecast, etc. Something like Siri for iOS :)

Thanks, Leon

EDIT

Could someone (official) answer this question: Are there any plans of a voice control application?

EDIT 2015-03-12

From TMO by taixzo - 2015-03-09:

Progress update: I have successfully got the Saera backend up and running on the Jolla with my new library. This means I can type things into the box and get generic replies back.

Next steps:

Get code on Github - the project structure has been completely reorganized which is why it's not up yet
Fix variables in intents (they are being calculated for recognition but not being reported back yet)
Get voice recognition (i.e. pocketsphinx) up and running
Finish porting old Saera actions to the new system
And, of course, find out how to talk to the jolla phone (is there documenation anywhere on how the Jolla alarm system works?)

edit retag flag offensive close delete

Comments

2

Onboard or offboard? As far as I know, Google voice and Apple's Siri are offboard, i.e. all voice is transmitted to their servers. Do we want this? Or is it something, people accept (for certain texts) as trade in for the convenience it brings? (Voice recogntn. onboard is probably not yet in reach.)

jgr ( 2014-01-03 00:56:01 +0200 )edit

2

As a user I would like to have such an app on my Jolla. As a developer I would like to be able to use TTS and speech recognition APIs. Most probably the voice recognition at the moment will be done on remote servers so some users might not use the app due to security issues but it's OK for me :)

leon.anavi ( 2014-01-03 01:18:51 +0200 )edit

13

Offboard recognition is an instant dealbreaker for me. There are a number of open source voice recognition systems available. On the N9(00) there was the community-made Siri clone called Saera iirc.

Fuzzillogic ( 2014-01-03 01:22:01 +0200 )edit

What about XOWi ? http://www.xowi.me/

FJVA ( 2014-01-10 00:18:36 +0200 )edit

5

limited onboard speech to text and control would be good. Offboard depends on a very good connection to hosted servers and is not always available in areas of poor signal strength. Richard

richardski ( 2014-01-20 10:28:04 +0200 )edit

22

answered 2014-04-05 19:36:05 +0200

This post is a wiki. Anyone with karma >75 is welcome to improve it.

updated 2015-07-21 23:37:52 +0200

raketti

2837 ●44 ●58 ●65

Heads Up.

Saera -a siri like app- is coming to sailfish.

https://twitter.com/taixzo/status/452300825646211072/photo/1

EDIT: It's here! (Although a bit ... sort of ... in a beta / early stage) But it's ready for grabs in OpenRepos.

edit flag offensive delete publish link

Comments

Also nothing new here?

torcida ( 2014-10-14 22:32:10 +0200 )edit

Current status: phone control is working, now trying to find a build of pocketsphinx for SailfishOS. Updates at http://talk.maemo.org/showthread.php?t=84753

taixzo ( 2015-04-09 20:30:19 +0200 )edit

1

Nice.

I'm waiting for it to hit the store.

ApB ( 2015-04-09 20:50:42 +0200 )edit

It may not be able to be submitted to the store due to Jolla policies on what applications can do. If that is the case I will upload it to Openrepos instead.

taixzo ( 2015-04-09 20:57:06 +0200 )edit

In case this happens maybe we can bring it up in community meeting. We need stuff like that tightly integrated in the OS. For a better experience.

ApB ( 2015-04-09 21:15:58 +0200 )edit

9

answered 2015-03-13 10:15:53 +0200

This post is a wiki. Anyone with karma >75 is welcome to improve it.

updated 2015-03-24 01:18:57 +0200

Igorion

1019 ●18 ●24 ●24

How about Sirius as voice assistant / intelligent personal assistant (IPA) in Sailfish OS? It's FOSS project under BSD license, source is here https://github.com/jhauswald/sirius. Their website is here - http://sirius.clarity-lab.org/.

edit flag offensive delete publish link

Comments

I also took a look the Sirius project and while I quicky read what it was about I got the idea it's designed more towards desktop use, correct me if I'm mistaken. But if it's easy (read: if someone is up to the task) to port to Sailfish, then why not, the more the merrier! :) But to be honest, in some cases, less is more. :)

raketti ( 2015-03-13 10:28:50 +0200 )edit

raketti, I hope it's easy :) As can I see here http://sirius.clarity-lab.org/index.html%3Fp=9.html they planning to use their project not only on the desktop / notebooks, but on mobile devices as smart watches too.

Igorion ( 2015-03-24 01:44:43 +0200 )edit

Oh, cool. Didn't find that piece of text. This seems really promising, and personally I hope this will be implemented to Pebble (just got mine - super cool!), since it's also a quite open device and community.

<off-topic> I'm planning to write them some feedback, since I can't update the firmware without Android or Apple phone (but that's a nother topic).</off-topic>

But anyway, since this is (more or less) open source project, this could be implemented by Jolla or a community member(s). Got to stay tuned on this one! :) And thanks again for the link.

raketti ( 2015-03-24 08:34:00 +0200 )edit

have just to find now someone who can port it into sfos...

cemoi71 ( 2015-07-22 01:13:30 +0200 )edit

7

answered 2019-01-29 13:01:47 +0200

leon.anavi
1421 ●5 ●8 ●10

It is funny that I started this discussion 5 years ago and it is still a hot topic :) In the mean I had a technical talk at Embedded Linux Conference Europe 2018 about smart speakers: https://www.linux.com/blog/2018/11/closer-look-voice-assisted-speakers

Although in Jolla and SailfishOS the environment is a bit different, some of the open source technologies used by MyCroft could be relevant for this discussion too.

Best regards, Leon

edit flag offensive delete publish link

Comments

Thanks, looks interesting! :)

MartinK ( 2019-01-30 12:20:10 +0200 )edit

2

answered 2016-01-07 17:16:45 +0200

xllndr
1164 ●15 ●24 ●29

Like the "answer" before, I also want to mention a library that recently got opensourced. Homepage GitHub

edit flag offensive delete publish link

2

answered 2019-01-29 12:41:24 +0200

This post is a wiki. Anyone with karma >75 is welcome to improve it.

updated 2019-01-30 07:25:01 +0200

raketti

2837 ●44 ●58 ●65

Mozilla's Common Voice project can also be used for this task.

Similar requests on together.jolla.com:

Voice control projects in other operating systems:

Edit 2019-01-30: There's also Snipek, based on Snips

edit flag offensive delete publish link

Comments

quite interesting info. thx
hmmm all those actor don't mention the localisation of the server of this application...
for google cortana bixby and siri it's quite clear. but snipek or mozilla....
Is just because of private sphere usage....

cemoi71 ( 2019-02-11 23:11:00 +0200 )edit

You self host the Snipek (Snips), so you are in control of the server/data.

raketti ( 2019-02-12 10:58:17 +0200 )edit

:-)
did you tried some of them?

cemoi71 ( 2019-02-12 16:28:28 +0200 )edit

2

answered 2019-12-05 07:19:26 +0200

_David_
21 ●1 ●2

There is a lot of choice now.

Mozilla's had an open system out for a few years. Facebook has released an engine-only internal project.

However, Kaldi doesn't need to deal with GPUs/DSPs and is already trained. Does Sailfish support the needed libraries?

https://mc.ai/how-fast-is-facebooks-end-to-end-speech-recognition-toolkit/

I found Google Keep with its voice support amazing. But a second best might be to record the audio and let my home PC do the transcription.

edit flag offensive delete publish link

1

answered 2015-04-17 17:14:58 +0200

ApB
27604 ●302 ●421 ●409

updated 2015-04-17 17:16:27 +0200

https://www.youtube.com/watch?v=XDlmgEa1aJY

alive

Would be nice to have it in the partner screen (a swipe away)

edit flag offensive delete publish link

1

answered 2015-06-26 16:16:01 +0200

arjwright
3038 ●20 ●30 ●36

http://antoinerjwright.co...

Given the following information about Amazon's Alexa, this feature request might need to move into at least a prototype state: http://techcrunch.com/2015/06/25/amazon-unbundles-alexa-virtual-assistant-from-echo-with-new-dev-tools/

edit flag offensive delete publish link

Comments

1

Do you trust amazon?

ApB ( 2015-06-26 17:10:53 +0200 )edit

1

answered 2015-07-23 15:56:07 +0200

Tucky
1081 ●14 ●28 ●27

There is link text that runs on a Raspberry Pi. It is not an answer, but I thought it is worth to mention it.

edit flag offensive delete publish link

Voice Control

Comments

9 Answers

Comments

Comments

Comments

Comments

Comments

Question tools

Stats

Related questions

Voice Control

Comments

9 Answers

Comments

Comments

Comments

Comments

Comments

Question tools

Public thread

Stats

Related questions