Does Zee S2 have enough power for the skill?

[jhearty]

I played with voice commands via HSTouch android app to my Zee S2 which works, but it was explained to me the Google VR in the phone is leveraged there and simple text is passed to the Zee. This is consistent with an old post by macromark, but seems to conflict with posts by Wade on this forum. As it has also been explained to me, the Echo API would send the audio to the cloud for VR and text passed back to HS, so Zee controller power should not be an issue with the API. Finally as I understand it the SKILL will pass audio from the Echo directly to HS for the controller to do the VR processing.

So my main question is if I use the SKILL, will the Zee S2 have enough processing power to do the VR processing? What kind of delay might I expect for said VR processing on a Zee S2? Another question is for the API, what kind of delay are others seeing for the voice to go to the cloud for VR and then text sent back to your HS controller?

John

[rmasonjr]

I can only comment on the API, since the skill requires you to say "Alexa, tell homeseer to..." which I dont like...

The API is very fast - maybe 1-2 seconds after speaking does the device get controlled.

I'm running HS3Pro on a Raspberry Pi2 (same as a Zee S2).

[Sgt. Shultz]

This information is in regards to the regular Amazon skill, rather than the API one (they're both called skills now :/), as well as commands through the phone (which is not associated with the Echo or Amazon)

All commands, both from the phone and the Echo, are text which is converted straight from audio.

All processing is done in HS3 regardless of the source..., the only difference is the path being taken to get to HS3.

If you use your phone on your WiFi, then the command is sent directly to your HS3 system via your router and processed.

If you use your phone outside your WiFi, it is routed through the Homeseer tunnel which has a private connection to your HS3 system (through your router)

If you use an Echo or Dot, the command is sent to Amazon, then to our pass through site (myHS), then to the tunnel, then to your HS3 system (again, through your router).

If you use an Echo or a Dot, you need to preface your command with a phrase like "Tell HomeSeer to".

With your phone, you need to press a button to talk, but you don't need to preface your command with anything.

[jhearty]

This information is in regards to the regular Amazon skill, rather than the API one (they're both called skills now :/), as well as commands through the phone (which is not associated with the Echo or Amazon)

All commands, both from the phone and the Echo, are text which is converted straight from audio.

All processing is done in HS3 regardless of the source..., the only difference is the path being taken to get to HS3.

Hmm. I'm still a bit confused by this. Are the commands you mention above that are in text from the Echo the "Tell Homeseer to..." part? Is the rest that follows still audio, processed by HS3? If doing this from the phone, the below post from Mark led me to believe the phone was processing all speech and passing text of what it thought it heard to HS3. Re-reading it, perhaps there is some intermediate step between the "VR heavy-lifting" done by the phone and perhaps the Echo, and the HS3 processing to turn it into text?

http://board.homeseer.com/showthread...d+voice&page=2

"When you use HSTouch-Android voice recognition, we're piggy-backing on the Google speech engine. Google is doing ALL the VR heavy-lifting. We just take what they give us, convert that to a text file and parse it internally."

[Sgt. Shultz]

The command from the phone and the command from the Echo are both text that were converted from audio.

The HS3 receives that text, parses it out, and processes it.

The text from the phone is converted using the Google speech engine prior to it being sent to HS3.

[jhearty]

OK, I'm almost with you. So HS3 always receives text, whether it is from the android HSTouch app, the skill, or the skill formerly known as the API.

Just want to be clear now on what is doing the VR and converting speech audio to text in the case of Echo/Dot for both the Skill and the API. You said "the command from ... the Echo" is text. Is that converted to text by the Echo itself, or by Amazon cloud on its way to HS? And again question is for both skill and former API. Sorry if you've answered this 5x before. I think I read all those but I'm not yet crystal clear. I was under the impression the only VR the Echo did itself was the leading commands like "Alexa" and "Tell Homeseer to..." and the rest was done outside the Echo.

John

[Sgt. Shultz]

For the Amazon devices, audio is sent from the Echo and Dot to Amazon where it is converted to text.

From there it is sent to HomeSeer's tunnel, then sent to your HS3 system where it is parsed and processed.

Response are sent back to Amazon as text where they are converted to audio and sent back to the Echo or Dot.

[jhearty]

Thanks! Now I'm crystal clear. I asked for a Dot for Father's Day.

John

[joey52685]

I have the Echo, it's awesome. All voice recognition is done by Amazon's servers. It's very fast.

My only complaint is that the Dot is currently back-ordered