wswartz Posted December 16, 2016 Posted December 16, 2016 Great thread! Wanted to jump in and share my TTS solution. Probably not easy to replicate, but works really well... Using Cisco Call Manger Express with 13 IP phones throughout the house. Phones are also touch screen capable and are my control interface to a number of ISY programs and scenes. For TTS, I use software by Cepstral: http://www.cepstral.com/ I tested a number of TTS solutions (both free and paid) and this solution seemed to be the most clear and understandable. I still need to convert the TTS audio into a format that the phones can consume. I do this with some scripts that run the audio through FFMPEG. Then, using the same process outlined by Thanatar previously, the ISY can provide triggers to the audio alert system. The kids have affectionately named the system, "Jarvis" (not original, but they think it's cool), and "he" has almost become another member of the family. Here's a sample of what Jarvis does for us: - Wakes the kids up (and if they don't get up, he wakes up mom/dad and tells specifically which kid isn't up) - Tells kids when to go to bed and if they are goofing around instead (via motion sensors) - When the bus comes in the morning (motion sensor outside) - Alerts on water in the basement, under sinks, and water in basement windows wells - Alerts on intrusion - If our chickens are free-ranging, alerts that it's time to put them away (sensor on chicken door) That's just the short list, there are a bunch more. Finally, Cepstral does offer a voice that works with the Raspberry Pi. Go to their demo page (http://www.cepstral.com/en/demos) under the "Voice" drop down and listed to the US English voice called "Callie". Last I checked, they offer this voice for only $35. The voice I'm using is a UK English voice called "Lawrence" and he has that English butler kind of sound. Thanks, Bill 1
jasont Posted December 16, 2016 Posted December 16, 2016 I don't believe it was a very hard sell to make me consider that HiFi Berry Amp - LOL, SOLD! Sold here, too! And it (along with everything else I needed) was Amazon Prime, so I'll be geeking out Sunday when it arrives. I wish If found this thread a month ago, it'd save me the hassle of having to return an Aeotec ZW056 doorbell that I'm pretty sure I'm getting as a gift for Christmas. lol
Thanatar Posted December 16, 2016 Posted December 16, 2016 Teken, jasont Be sure to visit this site to add the amp to your system. https://support.hifiberry.com/hc/en-us/articles/205311302-Configuring-OpenElec-with-device-tree-overlays It'll walk you through everything you need to do. 1
Teken Posted December 16, 2016 Posted December 16, 2016 (edited) OK just got home and couldn't stand to wait for the week end to come so decided to move forward with this. I downloaded the image provided by Thanatar and unzipped the file and mounted it with Win32DiskImager. Once complete I moved the Micro SD card to the RPi and loaded it up. When the RPi booted up I see the system indicating various messages but nothing is happening? Did I miss a step here or is there something else that needs to be done to make the program continue? EDIT: Never mind the system just stated rebooting! I guess I'm too used to seeing a scrolling dot or status bar. Let me continue on and go from here - Thanatar, you're indeed Santa Clause in the making! Thank You Sir! Edited December 16, 2016 by Teken
Thanatar Posted December 16, 2016 Posted December 16, 2016 OK just got home and couldn't stand to wait for the week end to come so decided to move forward with this. I downloaded the image provided by Thanatar and unzipped the file and mounted it with Win32DiskImager. Once complete I moved the Micro SD card to the RPi and loaded it up. When the RPi booted up I see the system indicating various messages but nothing is happening? Did I miss a step here or is there something else that needs to be done to make the program continue? What messages are you seeing? I haven't set one up in a while, but I can load an image in a bit. There should be an option for Samba and SSH. Make sure both of those are on. From there set up the network address to your network. Once that is done you should be able to see the device under Network on your PC.
Teken Posted December 16, 2016 Posted December 16, 2016 OK on to the next steps how do I configure Audacity to record the sound from the computers sound card and not from the built in microphone?
Thanatar Posted December 16, 2016 Posted December 16, 2016 At the top right of Audacity should be a picture of a mic. On mine I can drop it down and pick my sound card as the source. Doing that allows me to record whatever comes across the speakers.
Teken Posted December 16, 2016 Posted December 16, 2016 (edited) At the top right of Audacity should be a picture of a mic. On mine I can drop it down and pick my sound card as the source. Doing that allows me to record whatever comes across the speakers. I just don't know why this is so hard I've selected every option in the program and yet the software insists upon using the laptops built in speaker? The good news is my voice is clearly record via the built in microphone but sadly selecting the sound card for the computer still makes the robot ladies voice to be recorded from the speaker opposed to the sound card? I must have selected every option in dozens of combinations and it still won't record her from the sound card. EDIT: I believe this has been sorted out - Not sure why the previous selection didn't work as I selected the same bloody thing five other times! The program must have been in some kind of unknown state. Cripes . . . Edited December 16, 2016 by Teken
Teken Posted December 16, 2016 Posted December 16, 2016 I have to tell you this website that does the TTS offers some of the most useful (basic) and expected features. Most of the websites I have seen and played with there was no method to make the robot stop, pause, or emphasize a phrase / sentence. This one however allows this and when you record the voice it offers that *Almost* real human voice. Once again much thanks to all of you who have taken the time to share your invaluable projects, insight, and knowledge. Each and every one of you will be 2016 super hero's in my books because this is really starting to round out my HA. Right now I am making all of the low battery, leak sensor voice alerts. Next are some security related TTS I have always wanted to have and deploy in my home. Julie (US) is my preferred girl and thus far meets the balance of soft, distinct, and focused voice. I don't know how many weather and energy voice alerts I will make but thus far I have come up with at least 89 conditions I want to be vocalized! Guess I know what this week end has in store for yours truly after hitting the gym!
Thanatar Posted December 16, 2016 Posted December 16, 2016 I have to tell you this website that does the TTS offers some of the most useful (basic) and expected features. Most of the websites I have seen and played with there was no method to make the robot stop, pause, or emphasize a phrase / sentence. This one however allows this and when you record the voice it offers that *Almost* real human voice. Once again much thanks to all of you who have taken the time to share your invaluable projects, insight, and knowledge. Each and every one of you will be 2016 super hero's in my books because this is really starting to round out my HA. Right now I am making all of the low battery, leak sensor voice alerts. Next are some security related TTS I have always wanted to have and deploy in my home. Julie (US) is my preferred girl and thus far meets the balance of soft, distinct, and focused voice. I don't know how many weather and energy voice alerts I will make but thus far I have come up with at least 89 conditions I want to be vocalized! Guess I know what this week end has in store for yours truly after hitting the gym! That's great! I picked Julie when I did mine for the same reasons. So were you able to make the pi play your files correctly? Are you just using amplified speakers for now? If you are I'd be curious how the sound was for you. Mine always had a faint hiss. With 89 alerts you far surpass my list. Glad someone else was able to put this setup to use. Keep us posted on how your experience is with everything.
Teken Posted December 16, 2016 Posted December 16, 2016 That's great! I picked Julie when I did mine for the same reasons. So were you able to make the pi play your files correctly? Are you just using amplified speakers for now? If you are I'd be curious how the sound was for you. Mine always had a faint hiss. With 89 alerts you far surpass my list. Glad someone else was able to put this setup to use. Keep us posted on how your experience is with everything. I'm just in the beginning stages of this awesome project and haven't even loaded anything to the RPi. Right now I am just focused on making all of the voice announcements. Once that is done I will import them to the RPi and hook up some cheap powered computer speakers and give it a listen. What program are people using to access the RPi with a GUI so I can see the directory and dump the files to the Micro SD card?
Thanatar Posted December 16, 2016 Posted December 16, 2016 I'm just in the beginning stages of this awesome project and haven't even loaded anything to the RPi. Right now I am just focused on making all of the voice announcements. Once that is done I will import them to the RPi and hook up some cheap powered computer speakers and give it a listen. What program are people using to access the RPi with a GUI so I can see the directory and dump the files to the Micro SD card? As long as you have Samba enabled on Openelec you will see it listed under My Network Places on your PC. It should be labeled openelec, but you can change the name under the openelec system menu. Another option is WinSCP. It is a free program that offers a tree style interface.
Teken Posted December 16, 2016 Posted December 16, 2016 As long as you have Samba enabled on Openelec you will see it listed under My Network Places on your PC. It should be labeled openelec, but you can change the name under the openelec system menu. Another option is WinSCP. It is a free program that offers a tree style interface. Thanatar, I must have fell off the couch and hit my head because (Samba) was just enabled and that was so obvious . . . Thank You!
Thanatar Posted December 16, 2016 Posted December 16, 2016 Thanatar, I must have fell off the couch and hit my head because (Samba) was just enabled and that was so obvious . . . Thank You! LOL
Teken Posted December 16, 2016 Posted December 16, 2016 I just transferred the first 20 or so files and totally forgot the RPi was attached to my HDTV. I was running all over the house trying to find that silly USB speaker. When I selected the music folder I almost got blasted on to the master bedroom floor as I had the volume cranked up watching *Saving Private Ryan* last night. Anyways the entire folder started playing consecutively and just WOW! Crystal clear audio being pushed out via the HDMI ports and through the sound system. No hiss, no hum, no buzz, nothing - just a sweet voice which I can tell you would date anytime! Once I finish most of the voice announcements I will push forward with making the NR's to connect to the RPi. Once that is done I will sit down and craft some programs and see how it operates in a real world condition. I know I've said this a few times but thank you so very much Thanatar! This has been one of the easiest voice deployments I have seen yet and will really be helping me complete my project. I plan to document all the steps and the resources that were needed to complete this task once I am done. Full credit will be called out in Project Titan to you and many others. Thank You!
Thanatar Posted December 16, 2016 Posted December 16, 2016 You may want to make a setting change for the consecutive playback. System>Settings>Music>Playback>Play next song automatically > OFF I can't remember if that is on or not on mine, but it' s best to be off. If you liked the HDMI audio you'll be amazed at the sound a little add on board can produce. 1
Teken Posted December 16, 2016 Posted December 16, 2016 You may want to make a setting change for the consecutive playback. System>Settings>Music>Playback>Play next song automatically > OFF I can't remember if that is on or not on mine, but it' s best to be off. If you liked the HDMI audio you'll be amazed at the sound a little add on board can produce. I will check for that setting for sure - but when I call the NR the system is simply going to play only that WAV file, correct?
Thanatar Posted December 16, 2016 Posted December 16, 2016 (edited) I will check for that setting for sure - but when I call the NR the system is simply going to play only that WAV file, correct? I'm not entirely sure........ I set mine up way back and was trying so many things when I was figuring out the NR calls. It shouldn't keep playing is my guess. Now I'm tempted to hook a monitor to mine and check. EDIT: OK. I checked, mine is turned off. Edited December 16, 2016 by Thanatar
mmoore99 Posted December 17, 2016 Posted December 17, 2016 Most of the websites I have seen and played with there was no method to make the robot stop, pause, or emphasize a phrase / sentence. This one however allows this and when you record the voice it offers that *Almost* real human voice. Teken, how do you make the robot stop, pause, or emphasize a phrase / sentence?
Teken Posted December 17, 2016 Posted December 17, 2016 (edited) Thanatar, I have almost completed all of the predefined voice messages. Network Resource: When you select add another pop up window appears to select *Header* details. Is anything being select and if so what option is chosen here? As I noticed your last field indicates *Content Length = 0* Edited December 17, 2016 by Teken
Teken Posted December 17, 2016 Posted December 17, 2016 Teken, how do you make the robot stop, pause, or emphasize a phrase / sentence? Hello Mmoore99, Surprisingly, the TTS web application simply accepts common spelling punctuation's and spaces to invoke a stop, pause, and emphasis. I had recalled this feature in a similar TTS service I trialed more than eight years ago which unfortunately was abandoned. If you enter say: Warning!!! Julie will actually emphasize that phrase. If you place a comma like: Hello, how are you <- This will create a short pause. I also noticed if you simply added space between the words this too would cause a pause in her vocal speech. So if you wrote out the following: Hello how are you <-- The space between the hello and how would cause her to pause then speak. A simple period would also cause her to make a line break which in laymens terms is simply a pause. But using a combination of each this makes the voice and tempo of her speech come across more natural. As some phrases give off an odd synthesized sound at least to me. In a few cases I used the reverb feature because no matter what I did the word I selected didn't sound human enough to me so adding the reverb oddly enough made her sound more human in tone and pitch. Which I used also but didn't find any of those options to have much benefit besides sounding like chip monks, or some cheap sci-fi sound effect. 1
larryllix Posted December 17, 2016 Posted December 17, 2016 I'm just in the beginning stages of this awesome project and haven't even loaded anything to the RPi. Right now I am just focused on making all of the voice announcements. Once that is done I will import them to the RPi and hook up some cheap powered computer speakers and give it a listen. What program are people using to access the RPi with a GUI so I can see the directory and dump the files to the Micro SD card? I use FileZilla to copy files back and forth between Windows and RPi very easily. No SD card removal. sent from a really small mobile keyboard, so may be a short response.
Bill Morrow Posted December 17, 2016 Author Posted December 17, 2016 I had thought about having the wife make the recordings, But she says I don't listen to her now, So I figured I would never hear the announcements. Here is the link: http://www.oddcast.com/home/demos/tts/tts_example.php?sitepal The cool thing is it doesn't have to be speech. I have the Mario fail sound play for my daughter when it is time to leave for school. I got extra WAF points when I added the AOL "You've got mail" sound for our mailbox. Also one more cool Openelec usage. I have several IP cameras around the house. I placed a few monitors in key areas with a Pi mounted to the back. A simple stream file loaded to the Pi allows me to view the camera feeds using Alexa to say which one is displayed. I could waste some serious time just playing with the voices. I'm kinda fond of Bridget.
Thanatar Posted December 17, 2016 Posted December 17, 2016 Thanatar, I have almost completed all of the predefined voice messages. Network Resource: When you select add another pop up window appears to select *Header* details. Is anything being select and if so what option is chosen here? As I noticed your last field indicates *Content Length = 0* Nothing was selected in the add window. I'm not sure why it says Content Length = 0. I just made a fresh NR and it did not add that line. It worked with no trouble. Must have been leftover from my testing. Safe to ignore.
Recommended Posts