Speak through camera (TTS / audio upload PPT)

[fabiosoft]

I repost here... avoid writing in other repo.

I updated to 0.1.9 and browsing to speak.html page... the UI is really nice. But i click on TTS or upload a wav file nothing happens, and the camera is quite.

I installed the engine for tts and nothing changes... Wav file format seems correct... i converted the file using the specs you provided. (S16LE / 8 KHz, 16 bit, mono)

Here is a VLC file information:

How can i use this feature? I tried on a y21ga and r30gb models

[roleoroleo]

Sorry, there is an error in the label. The format is S16LE / 16 KHz, 16 bit, mono, not 8 KHz. 8 KHz is the format for the output.

Try again please.

[thedevstone]

I've followed all instructions but when I try to send the wav to the cam it responds but I can hear nothing

[roleoroleo]

Error is false, so the web service works. Are you able to hear the sound using the official app? Did you disassemble the cam? The speaker is on the back of the box, when you reassemble it you have to tighten the screws to ensure the contact.

[fabiosoft]

From the app everything works like a charm. @Giulianini can you upload your demo file so i can test the same?

[thedevstone]

@fabiosoft Uploaded here in sendspace

[roleoroleo]

Check this commit: https://github.com/roleoroleo/yi-hack-Allwinner-v2/commit/5ed736f2b04b1e503bcaf111f3ade4fcf9301766 I pushed it after 0.1.9 release.

[thedevstone]

@roleoroleo i'm using yi-hack-allwinner-v1 (I have a 9FUS), speaker.sh differs from yi-hack-allwinner-v2, they don't have to be the same?

[roleoroleo]

Sorry, what's your model?

[thedevstone]

Y20ga

[thedevstone]

news?

[roleoroleo]

Not yet.

[thedevstone]

@roleoroleo news for speak through camera?

[roleoroleo]

I tested again the web page with my y21ga and it seems working properly. Both text and audio file. Did you solve your problem?

Are you sure you are not confusing the 2 models? allwinner and allwinner-v2? Check your model.

[thedevstone]

I'm using y20ga and all winner V1. The speaker.sh file of allwinner V1 is different from V2. They don't have to be the same?

[roleoroleo]

They are different, yes.

[d0m1n1qu3]

i have also the problem to hear something and found aout that in v 0.19 th cgi-bin/speaker.sh use the command "speaker on" but this it not in path .. if i do it over ssh with

root@yi-cam-1:~# /tmp/sd/yi-hack/bin/speaker on
Running ioctl: num 16
root@yi-cam-1:~# cat /tmp/sd/yi-hack/samples/Front_Center.wav > /tmp/audio_in_fifo
root@yi-cam-1:~# /tmp/sd/yi-hack/bin/speaker off
Running ioctl: num 17
root@yi-cam-1:~# 

it works fine so i think i could be also you problem? can we fix it in 0.20?

speaker.sh.txt

[roleoroleo]

speak.cgi should inherit PATH variable from httpd daemon, but I can add it to the next release to avoid any problem. But @Giulianini is running on a different cam, where speaker command doesn't exist (no need to move the gpio to activate the speaker).

[thedevstone]

Ok thanks @roleoroleo

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.