›
This is the forum archive of Homey. For more information about Homey, visit the Official Homey website.

The Homey Community has been moved to https://community.athom.com.

This forum is now read-only for archive purposes.
Closed

Version 0.8.24

13567

Comments

  • ReflowReflow Member
    edited March 2016
    Same here, speech-input logs nothing and doesn't work at all. Ok Homey works fine. Speech-output is full with "Oops, couldn't reach the server" sometime 2-5 times a minute. Happens both with Dutch and English language


    oeps.PNG 525.6K
  • ArjanArjan Member
    I've received oops can't reach that server just 2 times, been giving Homey lots of commands.

    Luckily i'm not the only one.
  • PhuturistPhuturist Member
    edited March 2016
    Just chippin' in here:
    • upgrading took some reboots and a power cycle even though automatic upgrades was ticked
    • Homey does not crash anymore on first voice trigger
    • speech recognition has decreased for me though, from a 70% hit ratio to a 15% hit ratio (I'm unable to finish the OK Homey learner) -> update: with adjusting my speech to a more phonetic kinda trigger sounding like okiyomi I'm back on track again with a succes rate of about 80%.
    • z-wave sockets still time out
    • refreshing z-wave nodes works again without throwing an error
    • sometimes the list of z-wave devices on the z-wave settings page stays empty but it now shows loading ids (a reboot is needed to bring the list back)
    • z-wave in general still works for me but sockets time out from time to time (see issue here)
    • KAKU in general still works for me
    • there are some 404 error on the settings pages here and there not loading an icon font
    • issue with button for device settings not solved yet (see issue here and here)
    • some small UI improvements
    All in all, it doesnt break anything compared to 0.8.22 but it doesnt fix much either.
  • bvdbosbvdbos Member
    edited March 2016
    I didn't have the "Oops..." but speech-input didn't log anything, not after reboot and not after powercycle. Then I did a system-restore while keeping settings, no luck. Finally I did a system restore with deleting all settings. Now inputting through http://ip/manager/speech-input works...
  • ArjanArjan Member
    BasVanDenBosch said:
    I didn't have the "Oops..." but speech-input didn't log anything, not after reboot and not after powercycle. Then I did a system-restore while keeping settings, no luck. Finally I did a system restore with deleting all settings. Now inputting through http://ip/manager/speech-input works...
    Good to hear that solved the problem, let's see if the team at Athom finds out what the cause is of this problem. Been trying and testing some things and added those to #296
  • PhuturistPhuturist Member
    edited March 2016
    I noticed I have more succes with the voice trigger if I articulate less. If I speak okiyomi in a single and fast spoken word I have much higher results. Perhaps someone else can try and verify this. This might be useful information for Athom related to underfitting or overfitting the machine learning of the voice trigger.
  • HarmStellingaHarmStellinga Member
    the update was good here.
    The loud noise sound is gone, and can time, date etc. call via app on phone.
    via speech input works okay.
    The ok HOMEY trainer have to wait till tomorrow .

  • Fire69Fire69 Member
    Phuturist said:
    I noticed I have more succes with the voice trigger if I articulate less. If I speak okiyomi in a single and fast spoken word I have much higher results. Perhaps someone else can try and verify this. This might be useful information for Athom related to underfitting or overfitting the machine learning of the voice trigger.
    Nope, makes no difference here.
    I have 0% success rate with the training.
    I'm in a completely silent room, about 10m away (tried closer also) but I always get 'there is too much background noise', 'try talking less loudly' or 'try to articulate more clearly'

    After discovering the ledring bug in .20, trigger was working fine.  It broke in .22 and it appears worse now.
  • PandaPanda Member
    Speech input here doesn't work either. Going to do a full reset to see if this fixes something.
  • spamenigmaspamenigma Member
    Mine seems OK for speech input, other than the OK homey not being picked up compared to hey homey. Anything I say shows in the /speech-input/ page (well what it thinks I said, often wrong). The last time I saw this it was one or more flows that was the cause, I deleted them all and suddenly it worked. Just suggesting to check flows and if you have a few try clearing one/some/all :)
  • ArjanArjan Member
    I did remove some flows, got back to 19 and now able to enter text in the /manager/speech-input/ screen. Will have to test tonight if it "really" works with voice input.
  • mruitermruiter Member

    Was expecting some news about the "Oeps ikke ben te dom om de server nog te vinden meldingen"

    @Emile al een idee wat het probleem is, want verder lijkt mijn Gody goed op spraak te reageren.

  • ArjanArjan Member
    edited March 2016
    Arjan said:
    I didn't have the "Oops..." but speech-input didn't log anything, not after reboot and not after powercycle. Then I did a system-restore while keeping settings, no luck. Finally I did a system restore with deleting all settings. Now inputting through http://ip/manager/speech-input works...
    Good to hear that solved the problem, let's see if the team at Athom finds out what the cause is of this problem. Been trying and testing some things and added those to #296
    Got it! If you have more than 1 flow which contain a microphone in when, the  http://ip/manager/speech-input stops working!

    For all other people having issue's with speech commands, please check that you do not have more than 1 flow with a microphone in the "when" and than check if speech issues are resolved.
  • Fire69Fire69 Member
    Seems like the same problem as before. 
    Is this a general problem, and if so, how did it get through testing? 

    @arjan, did you remove the flows? Or is just unchecking them enough? 
  • ArjanArjan Member
    edited March 2016
    Fire69 said:
    Seems like the same problem as before. 
    Is this a general problem, and if so, how did it get through testing? 

    @arjan, did you remove the flows? Or is just unchecking them enough? 
    You need to remove the microphone from the flow to get it to work.

    Edit 1: In my case i replaced them with input of random IR Remote signals (IR is broken anyway) to  test.

    Edit 2: I'm currently still at work so i can't check voice input but i'm confident that this should fix the issue.
  • EmileEmile Administrator, Athom
    edited March 2016
    Arjan said:
    Got it! If you have more than 1 flow which contain a microphone in when, the  http://ip/manager/speech-input stops working!

    For all other people having issue's with speech commands, please check that you do not have more than 1 flow with a microphone in the "when" and than check if speech issues are resolved.



    Damn, that was one tricky bug! Fixed it :-) Thanks for finding out when it was caused!

  • brot84brot84 Member
    I have only one micro flow in when but i doesnt  work! 
  • mruitermruiter Member
    edited March 2016

    Have no flow with Mic because it didn't work before this version.. Well.. before...... Still haha

    Issues with the ooooeps still in progress

  • ArjanArjan Member
    brot84 said:
    I have only one micro flow in when but i doesnt  work! 
    mruiter said:

    Have no flow with Mic because it didn't work before this version.. Well.. before...... Still haha

    Issues with the ooooeps still in progress


    What happens when you go to http://Homey's-ip/manager/speech-input and then type in a sentence like "What is the time" ? The sentence you entered should appear in the http://Homey's-ip/manager/speech-input screen and in the screen http://Homey's-ip/manager/speech-output the answer should be visible.


  • MartijnDeRhoterMartijnDeRhoter Member
    Not sure if it's thanks to 0.8.24 or KaKu 1.0.15, but I can no longer use my doorbell (ACDB-7000AC).

    It pairs, but it doesn't register - all was fine on 0.8.22 with KaKu 1.0.15 even, but I had to do a full reset and now it's broken :)

    Speaking of KaKu, the new pairing flows are a little broken graphically. When pairing a dimmer socket, it won't allow you to pick the device (so it stays a socket, instead of a lamp for example) and when you want to copy a remote signal, the wizard says "press a random key". That should be "press the key you use to control the socket".

    Small regression perhaps?
  • mruitermruiter Member
    edited March 2016
     


    What happens when you go to http://Homey's-ip/manager/speech-input and then type in a sentence like "What is the time" ? The sentence you entered should appear in the http://Homey's-ip/manager/speech-input screen and in the screen http://Homey's-ip/manager/speech-output the answer should be visible.


    That works like a charm. Oepsie only comes when i command my slave true the finally working mic.
    Speech recon looks pretty good on my homey. He's just to stuppid to find the servers
  • wvanderspekwvanderspek Member
    edited March 2016
    still have the problem that after a couple of hours my homey becomes slow, the pages are loading very slow and the reaction to pressing a button on my kaku remote is very delayed.

    I this something that is understandable for this update? or is it maybe something in a workflow.

    Also triggers on sunset are not working.

    to be clear microphone is off and there are no flows or any kind of speech related items.
  • mruitermruiter Member
    wvanderspek said:
    still have the problem that after a couple of hours my homey becomes slow, the pages are loading very slow and the reaction to pressing a button on my kaku remote is very delayed.

    I this something that is understandable for this update? or is it maybe something in a workflow.

    Also triggers on sunset are not working.

    to be clear microphone is off and there are no flows or any kind of speech related items.

    Do you have Z-wave device's ? I Noticed this also when some of my defuct z-wave devices where configured
  • HarmStellingaHarmStellinga Member
    HarmStellinga said:
    the update was good here.
    The loud noise sound is gone, and can time, date etc. call via app on phone.
    via speech input works okay.
    The ok HOMEY trainer have to wait till tomorrow .
    I tried the ok HOMEY learn proces, no results never get green ring when i say oke HOMEY , only a red ring.
    you get advice on the screen, but why is it hiding behind the 5 buttons ?


  • brot84brot84 Member
    Arjan said:
    brot84 said:
    I have only one micro flow in when but i doesnt  work! 
    mruiter said:

    Have no flow with Mic because it didn't work before this version.. Well.. before...... Still haha

    Issues with the ooooeps still in progress


    What happens when you go to http://Homey's-ip/manager/speech-input and then type in a sentence like "What is the time" ? The sentence you entered should appear in the http://Homey's-ip/manager/speech-input screen and in the screen http://Homey's-ip/manager/speech-output the answer should be visible.



    That Works for me, Input Bad Output visible and Homey unterstand correct what im writing,i dont want Do an hard reset because the Z-Wave Devices
  • AndreOerlemansAndreOerlemans Member
    edited March 2016
    After update to 0.8.24 i experience an extremly loud and distorted voice trough the internal speaker, tried to set the volume in flows but it doesn't work anymore. Any ideas?

    Problem solved, made a new flow and add a volume control regulator, speech is normal again
  • wvanderspekwvanderspek Member
    mruiter said:
    wvanderspek said:
    still have the problem that after a couple of hours my homey becomes slow, the pages are loading very slow and the reaction to pressing a button on my kaku remote is very delayed.

    I this something that is understandable for this update? or is it maybe something in a workflow.

    Also triggers on sunset are not working.

    to be clear microphone is off and there are no flows or any kind of speech related items.

    Do you have Z-wave device's ? I Noticed this also when some of my defuct z-wave devices where configured
    Yeah i have 2 greenwave sockets, so do i have to remove those sockets or is there something else i can do?
  • mruitermruiter Member

    Reading up on the last days, did not have much time.

    My Homey looks like most of it is working after .24 update. Ok Homey works like a charm, even from the kitchen on the other side of the room.

    However every voice command results in a Oeps could not find servers.
    Punching in command as text in speech input page does not have this problem and results in time etc etc.

    Am i not looking correct or did nobody at Homey respond yet with a sollution for this problem. Lot of people seem to have this problem.

  • Fire69Fire69 Member
    @emile looked at the problem after @arjan reported it on Slack.
    It was a bug and is fixed now, so the next release should solve your problem.
    In the meantime you can try removing (or just disabling, don't know if that works also) flow that use the microphone.  That seems to get rid of the problem...
  • mruitermruiter Member
    Fire69 said:
    @emile looked at the problem after @arjan reported it on Slack.
    It was a bug and is fixed now, so the next release should solve your problem.
    In the meantime you can try removing (or just disabling, don't know if that works also) flow that use the microphone.  That seems to get rid of the problem...

    Thanks , did not see it was on slack.
    Have no flows with mic / speech because it did not work before .24 on my homey. Double checked it.
    So no resolution for me then (Maybee a restore to default but that will kill my flows again)
This discussion has been closed.