Shorthand Suite

I read about when you guys were talking about the Shorthand Suite that Tyler was talking about (“SH List Demo.zip”), and I said “Hey, maybe I could do that.” Thus the 2013 Shorthand Suite.

I haven’t really started anything yet, aside from the already-ongoing Greggory text-to-shorthand converter, which I might release around March/April. On deck is a shorthand-to-text recognizer, which might require a lot more work than the converter. I can get these down in a few months, but if anyone knows to program in C# out there and could help me, it would make things very much easier.

What other programs do you want to include in the 2013 Shorthand Suite?

14 comments Add yours
  1. I'm not really sure what the Greggory-to-Text converter does or is used for, but since I'm self-teaching and thoroughly having a blast with learning shorthand that whatever is/will be included would be fantastic.

    I don't know C# but if there is anything that I can help with just let me know. 🙂

    1. Greggory (or Gregg.Convert) is a shorthand -> text converter, but right now I'm stuck at generating AGS. I don't know of any process to generate outlines.

      The other one (still unnamed) will let you write in shorthand then convert it to English text.

    1. My program does TTS. You can have a copy. It's in Python 2.7.

      It takes a text file, adds extra codes control the speed and extra time between sentences and paragraphs, then calls Cepstral to convert to speech.

      Calibration was tricky.

      The quality is adequate for speed-building, but not for cold dictation. Once the material is typed, it takes 30 minutes to convert thousands of words to 5 different speeds (plus supper while the computer works). Most of the time is converting WAV to MP3 and editing tags. I got bored of programming before I found a way to automate that.

    2. I'm writing my code in C#, because the GUI is much better than on Python. I've tried TTS in C#, but I don't know how to control the speed.

      What I'm trying to do is to implement the dictation software without having to bundle some extra software packages (in this case, Cepstral).

  2. You're right about Python's GUI. I used the MS IDE for a year, and still miss it.

    For the TTS part, look up SSML and SAPI. They're mark-up languages. In SSML, the speed is controlled by and you can add a delay with , or something like that. It's been a while since I worked with that part of the code.

  3. It all sounds neat. After all is said and done it would be so wonderful for there to be a program that would have a nice user-friendly interface, take a text file, and just simply type in (say 50 WPM) and the software takes care of the rest. I'm sure it is all very complicated. I'm sure you can do it with Cricket's program but it sounds like a lot of time & work to get a dictation file just right.

    1. Time? Oh yeah. The best way to learn to program is to have real projects, so it counted as learning to program rather than shorthand time. Packaging it nicely for non-programmers is, wildly-guessing here, another 10 hours, mostly learning. And they you'd still have to pay for Cepstral, and I get nightmares thinking of all the install cases. Which version of Windows? Which version of Cepstral? Where does Cepstral put its own files on your machine? What if Cepstral stops supporting the product? Way too many different cases to worry about.

      It's faster for you to email me text and I convert it to several speeds than to package the program properly.

    2. I wish I'd started that way. I started by emailing a few TTS manufacturers about slow speeds, hoping it would be easy (or they'd think it was a large enough market to go for) and Cepstral was the only one that replied. "I talked to the engineer. He said you'd have to program it. Here's a link to our page on SSML." It looked easy enough that I didn't look into the MS voices, other than to see they didn't use SSML.

      Now you've got me wondering how hard it would be to change the program. Change the tags from SSML to SAPI is easy. Calling the program automatically (so it generates several speeds for several files at once) would be a bit of research and, if it's possible, just edit a line or two.

      Our first version just took in text and spit out SSML, which I pasted into Cepstral's window. Once per file per speed. Then copy all the files to my mp3 player so I could work in another room.

  4. Sounds very technical. I really wished I knew programming. It sounds like it would be a lot of fun to work on a project like this. I'll sign up to beta test if it ever gets to that point (hahaha).

    1. It will get to that point… eventually. I'm stuck at outline generation. I haven't figured out a way to piece together strokes to form a textbook outline. Right now it looks like something from the Handywrite page. If I do figure it out, hopefully within the next few weeks, I'll let you know.

Leave a Reply