[arm-allstar] modifying write-node-callsigns script

[Patrick Perdue]

This is true. What I am proposing is to basically make a new version of 
that script which doesn't use any of the pre-existing sound files, but 
rather, creates a new one by passing the callsign string through some 
formatting directly to a TTS engine, putting a space after every 
character in the callsign field before it is sent to the TTS.

What I eventually would love to do after building a few sound packs 
using different TTS voices is offer a public directory through rsync for 
each of several voices, keeping things up-to-date, with a file in the 
main /var/lib/asterisk/sounds/ or /var/lib/asterisk/sounds/rpt directory 
that can be used as a reference, in combination with a script that pulls 
new nodenames with rsync, to get files matching your currently installed 
sound pack so that voices are consistent.

My ultimate goal is to make everything more efficient by replacing 
Allison and cutting down on TX time per utterance, while maintaining 
intelligibility.


On 1/27/2020 10:40 PM, "Doug Crompton via ARM-allstar" wrote:
> Patrick,
>
>     You don't pass the entire text to be interpreted. There is a
> speaktext.sh file in /usr/local/sbin that does this as an exampple. It
> takes each letter and processes it individually like  W A 3 D S P not
> WA3DSP.  Each letter has a gsm file in /lib/asterisk/sound/letters  or
> digits, etc.
>
>
> *73 Doug*
>
> *WA3DSP*
>
> *http://www.crompton.com/hamradio <http://www.crompton.com/hamradio>*
>
> On Mon, Jan 27, 2020 at 9:23 PM "Patrick Perdue via ARM-allstar" <
> arm-allstar at hamvoip.org> wrote:
>
>> Hi:
>>
>> I've been toying with the idea of making a modified write-node-callsigns
>> script, using an up-to-date astdb.txt as the source, but on a machine
>> not running HamVoIP, using a Mac to generate the TTS clips. However,
>> instead of concatenating existing GSM or ulaw files, I want to have it
>> say the entire text string of the callsign field with a space between
>> each character, so that TTS weirdness doesn't happen. I'm already using
>> a bash script on a Mac to generate replacement sound packs for HamVoIP
>> (more on that in the future.) In the name of efficiency, I would like to
>> have callsigns spoken at a single utterance. I'm not the most well
>> versed shell scripter around, so just looking for ideas of how to go
>> about formatting a text string in this way, which can then be fed to the
>> say command, then SoX to convert to ulaw as <node_number>.ul. The rest I
>> think I can handle.
>>
>> This may be a little overkill, but as a daily TTS user, I thought it
>> would be a nice touch, especially considering the availability of TTS
>> voices of Mac OS is greater than what is generally available in Linux.
>>
>> Also, the sound packs I've generated so far use ulaw, as it sounds
>> considerably better than GSM. Thus, I have to do a batch find/replace on
>> the HamVoIP scripts that specifically look for .gsm files. I'll probably
>> make both GSM and ULAW versions of these sound packs available when I
>> smooth some things out, and perhaps even a test in the scripts to use
>> either GSM or ULAW, depending on what is available, rather than only
>> assuming GSM.
>>
>> _______________________________________________
>>
>> ARM-allstar mailing list
>> ARM-allstar at hamvoip.org
>> http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
>>
>> Visit the BBB and RPi2/3/4 web page - http://hamvoip.org
>>
> _______________________________________________
>
> ARM-allstar mailing list
> ARM-allstar at hamvoip.org
> http://lists.hamvoip.org/cgi-bin/mailman/listinfo/arm-allstar
>
> Visit the BBB and RPi2/3/4 web page - http://hamvoip.org