Jump to content

Automated generation of Voice ATIS files


Recommended Posts

Hello all!

 

Some weeks ago, I decided to create the necessary files so vZID Euroscope controllers could use the automated ATIS functionality. At the time, though, I was at a loss about how to create things quickly and easily, especially having to record the names of all towers airports in our airspace. This is no longer a problem!

 

I now have a set of programs, where if you give it a list of things to say, you will get the .wav files and the AtisFiles.txt file, automatically!

 

The requirements are:

 

  • A Macintosh computer (required by the next item)
  • Mac OS X (for the built-in speech generation and scripting functionality)
  • QuickTime Pro (for access to the file export functionality)

 

At http://dl.dropbox.com/u/18635216/vZID%20ATIS.zip you can download a .zip archive containing the current vZID files, the two scripts used to create them, and a README file. (If WinZip also extracted a "__MACOSX" directory, you can delete it.) Good luck!

 

The drop was last updated on 6 March 2010.

 

P.S. I have also created, using the magic of Perl and recursive-descent parsing! You can get to it using the URL http://karl.kornel.name/metar2es.cgi?metar=$metar($atisairport)&atiscode=$atiscode&arrrwy=$arrrwy($atisairport)&deprwy=$deprwy($atisairport). It should work with all U.S. METARs. If you ever get an error that a METAR can't be p[Mod - Happy Thoughts]d, please send me the METAR and the Extracted URL via private message!

Edited by Guest

A. Karl Kornel - vZID C1, FE, and Mentor

Smoke Bomb! POOF

Link to post
Share on other sites
  • 1 month later...
  • 2 weeks later...
any way this can be ported to Windows?

 

Jonas' answer is correct. The best I can do is to give you some info on the technologies that are used.

 

The process of creating the .wav files is in two phases. Phase I is to create the sound files in AIFF format. A Perl script is used to drive Phase I. Perl already exists on Windows, so that is good, although the script might need to be adjusted to deal with Windows path naming conventions (for example, using \ instead of / for directory separation).

 

The actual program that generates the AIFF sound files, `say`, is part of the Mac OS X text-to-speech subsystem, which has been around since the original Macintosh. You would need to find a substitute program that would work on Windows. Also I should note, the Mac OS team have had over 20 years to work on and improve the voices available; your substitute should be of similar quality. The conversion to a 7.35 kHz Wave file really does reduce the quality of the recording.

 

As I said above, the `say` command outputs in AIFF format, so Phase II is to convert the recording files into their final format. I actually tried to go open-source first on this, by looking into the mencoder/mplayer set of programs. Unfortunately, not only could they not recognize the AIFF format, they also could not deal with sample rates below 8 kHz. In fact, I found that alot of programs would fail when asked to go below 8 kHz. In the end, QuickTime Pro was the only program that would open AIFF, work with sample rates under 8 kHz, and do all of this in an automated fashion. QuickTime Pro is available for Windows.

 

QuickTime Pro is a Mac OS X application, not a command-line tool like `say`, so it is going to use AppleScript as an automation mechanism. The script that implements Phase II, therefore, is written in AppleScript. I do not believe that AppleScript exists in Windows, but I do know that QuickTime in Windows has a COM/ActiveX interface available. It may be possible to use this.

 

So, there you go! The major barrier to a Windows implementation of this would be an exemplary text-to-speech program that can be used in an automated fashion. After that, it would be a matter of figuring out the QuickTime COM/ActiveX interface. Will I work on this? No; I already have a working solution, and I don't plan on switching from a Mac anytime soon!

 

There is one other thing I can do: I can add additional words and airport names to the download. I would just need to know what they should be (best to send that info by private message).

 

Later!

A. Karl Kornel - vZID C1, FE, and Mentor

Smoke Bomb! POOF

Link to post
Share on other sites
  • 2 months later...
The drop was last updated on 26 September 2009.

 

Since my voice ATIS package was released, I've made a number of updates: I have updated both the CGI script behind the ATIS maker URL, and the voice files that are included in the .zip archive. I am now keeping a log of changes made on the drop.io site. I do not want to constantly post updates here (or else the topic will get too big), so I will simply be updating my original post to show when the most recent change took place.

 

If you are using the voice ATIS files that I am creating, then you can check the date on my first post. If the date means that you downloaded an older .zip file, you should go to the drop.io site to see if the changes made are important to you.

 

Later!

A. Karl Kornel - vZID C1, FE, and Mentor

Smoke Bomb! POOF

Link to post
Share on other sites
  • 3 months later...
  • 2 months later...

In under an hour, I'm going to be pushing out a new version of the code and audio files. This update will include a 5-second period of "silence" at the end of the ATIS. This was a request submitted to me a while ago, and today it's here!

 

Oh, BTW, it won't actually be silence. If I just put in five seconds of silence, I expecting people to say "did the ATIS just disappear?" For those who are wondering, it will be 5 seconds of pink noise, mono, amplitude 0.01, from Audacity. It was recorded in at 7.35 kHz 32-bit float, then converted to 16-bit PCM and exported as a .wav. If you're going to try this at home, don't forget to set the project rate to 7350 Hz before starting!

A. Karl Kornel - vZID C1, FE, and Mentor

Smoke Bomb! POOF

Link to post
Share on other sites
  • 9 months later...
May I have an updated link, i am getting an error message saying that the service is discontinued?

 

Thanks!

 

It happens from time to time. We're also using Karl's ATIS template, just wait for it to come alive and try again.

Romano Lara
vACC Philippines, Manager - Training & Standards
04819c_4181f294a6c34b5aa4d8a82c0fb448c5~mv2.webp

Link to post
Share on other sites
  • 2 weeks later...
  • 10 months later...

Wow, people were/are actually using my stuff! SHOCK!

 

I didn't know if people still wanted the audio files, but I've moved them over to my Dropbox account. I've already modified the main post to point to the new URL.

 

Let's see how long Dropbox lasts!

 

Anybody else looking for airport name generation?

A. Karl Kornel - vZID C1, FE, and Mentor

Smoke Bomb! POOF

Link to post
Share on other sites

Hi all,

 

I've been thinking: It really annoys me when somebody makes something available, and then it disappears forever. I really dislike having that feeling of helplessness, how something you use regularly is suddenly gone.

 

I have decided to publicly release the server-side program that converts METARs into EuroScope script. The file is available here:

 

http://dl.dropbox.com/u/18635216/metar2es.cgi

 

This is probably not going to be a simple drop-in for your FE or Webmaster. It's written in Perl (but don't let that scare you ), and it requires one Perl module (P[Mod - Happy Thoughts]::RecDescent) that isn't installed by default in most hosting environments. It may take a little bit of work to get going on your own server, but I at least want to open the door for you.

 

If you decide to take it, and use it for yourself, would you mind letting me know? Thanks!

A. Karl Kornel - vZID C1, FE, and Mentor

Smoke Bomb! POOF

Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...