Computer Program to automatically follow a script?

[Sparky2012]

I was wondering if anyone knew about a program for mac or pc, that could listen to what actors are saying on stage, and follow a script, so that if people in the green room were not following the script, then they could quickly find there place. I know the best method is to have a person following the script at all times, and as a tech, if anyone I was in charge of was not following the script, they would be gone pretty quick.

Thanks in Advance!

Gareth Ellis

 

[Footer]

Not that I know of. That is a bit of a complicated piece of software for a rather simple task. I have seen music and script systems synced to a click track, but that is for touring shows that do the same thing every day for ever.

 

[Van]

I was wondering if anyone knew about a program for mac or pc, that could listen to what actors are saying on stage, and follow a script, so that if people in the green room were not following the script, then they could quickly find there place. I know the best method is to have a person following the script at all times, and as a tech, if anyone I was in charge of was not following the script, they would be gone pretty quick.

Thanks in Advance!

Gareth Ellis

Yes, there is a thing and it is Mac, PC and Linux compatible. In addition it also takes notes and makes coffee. It's called a PA. they are relatively inexpensie and seem to be willing to do almost anything for next to nothing.

 

[Sparky2012]

Haha, I am very aware of this option, and as I said in my question, I realize that that it is probably the best option. This is more just to satisfy my curiousity then anything else.

 

[sk8rsdad]

Let's break down the requirements of this software to see why this is still in the realm of science fiction:
To be able to follow a script:

• scan a document with character recognition to convert the image to text (let's ignore copyright and digital rights issues)
• artificial intelligence (AI) to extract the lines of dialogue from the stage directions, publishers comments, etc.
• accurate text-to-speech on everything including made up words, names, etc. in order to produce a baseline for comparison
• voice recognition with some more AI to deal with actors misspeaking their lines, jumping cues, etc. to say nothing of the beam forming and digital signal processing to filter out that speech from all the other sound that might be happening.
• some truly inspiring pattern matching to compare what is being heard on stage with the baseline
• still more AI, including predictive algorithms to anticipate the lead the time required for an actor who isn't paying attention to get themselves ready to appear on stage

To do it all in real-time with sufficient accuracy, something with a little more oomph than a PC or Mac is required. Siri lives on the internet, listens to one voice, has a very limited vocabulary, and doesn't come up with the correct answer often enough.

 

[StNic54]

I would add to that accents: I went to college in Alabama, and I'm pretty sure actors in the deep south working out their best Shakespeare are going to easily foil any voice recognition software. Just hand Siri to any friends with a heavy accent - you'll see what I mean.

 

[josh88]

sk8ersdad touched on it. The other problem is line jumps, flubs and making lines up completely. Lets hope everybody has it perfect but in a lot of the work with students I've done we're happy to get something at least close to the original lines and that doesn't always happen.

 

[jbreezy]

I would add to that accents: I went to college in Alabama, and I'm pretty sure actors in the deep south working out their best Shakespeare are going to easily foil any voice recognition software. Just hand Siri to any friends with a heavy accent - you'll see what I mean.

True facts. I live in the great state of slow Southern drawl, and half the people i know that have the technology can't use the voice recognition on their phones for that very reason.

 

[FACTplayers]

This is a really cool idea. It could be used to run mics and lighting cues. While it may seem impossible, I think it could be done with some VERY talented programmers, a lot of time, and a lot of money. If we can find people with the various skills needed, I would be interested in working on this project.

 

[FACTplayers]

The basics:

1) Speech recognition software. (already exists)
2) Import the script. (already exists - I would think the President has something like this for his speeches)
3) The software then allows for light/audio cues to be inserted at certain points.


Even if the actors mess up/skip lines, the software could use confidence levels and be able to skip lines if they are missed/changed.

 

[DuckJordan]

The amount of programming required would take a company like Google or Microsoft about 10 years to develop its just that difficult to get reliable speach to text recognition.

 

[techieman33]

The amount of programming required would take a company like Google or Microsoft about 10 years to develop its just that difficult to get reliable speach to text recognition.

It'll happen sooner than that, voice recognition is improving by leaps and bounds every day. Mainly because Apple and Google are throwing billions of dollars at it to try and one up the other.

 

[cpf]

Maybe you don't need speech recognition? What if you could plug in an audio recording of a reference performance with the audio matched to the cues, then have the program "follow" that audio?

 

[DuckJordan]

How is the computer going to compare the audio to audio? there is no way to program that at the moment, you have to reduce it to data down to ones and zeroes. to techieman they may be throwing tons of money at it but it still doesn't solve the problem. They are designed for simple commands not extensive sentences that are critical to making the show happen. It comes down to figuring out how to program the computer to understand all kinds of words, spellings, pronouncing and not require to go word by word syllable by syllable. At this point in time we don't have the technology and I don't see that developing until 20-30 years down the road.

 

[Chris15]

1) Speech recognition software. (already exists)
2) Import the script. (already exists - I would think the President has something like this for his speeches)

Yes, the President and others do have a system for this.
It's a human sitting there with a jog wheel controlling the teleprompter the same way as every broadcast news room runs.
Includes the massive advantage of being able to rapidly adapt to changes in pace, skipped or substituted lines.

 

[FMEng]

Letterman and Leno still have a cue card person crouched in front of them. Some things are better done the old fashioned way.

 

[FACTplayers]

Letterman and Leno still have a cue card person crouched in front of them. Some things are better done the old fashioned way.

*cheaper, not necessarily better

 

[museav]

*cheaper, not necessarily better

Probably better as well. As FMEng noted, to be effective such a system would have to be able to accommodate missed lines, ad libs, last minute changes and so on. You're also talking about having to follow different actors at different times with the added possibility of understudies and accents. And possibly having to track audio from something other than a close-mic'ed actor. Maybe it can be done but it seems like potentially a quite different situation than typical speech recognition applications.

Uisng something like that trigger cues potentially adds an entirely different element in terms of the overall situation and especially safety. I can see it maybe being used to give the SM warning but I don't see it replacing them.