Audio IF

So Final Destination 3 has a DVD extra that allows you to choose plot direction/variation in a couple of scenes using alternate takes, reshoots, and specifically filmed extra footage. It doesn’t change the movie much, but you can end the movie in 5 minutes and save all the characters if you want with a slide show of “where they are now”.

There are also a couple of movies with existing formats that could be potentially viewed in different orders, like Pulp Fiction and Go which consist of different viewpoints of the same events. Also, Clue had three filmed endings and the DVD lets you pick one (but it’s not as good as the combined ending with all 3.)

This is probably one advantage in IF: text is cheap and is only limited by author’s imagination and stamina.

6 Likes

I’m coming up on 100 published IF games for my newsletter Adventure Snack. Maybe I’ll adapt Facts About the Robot Uprising to Ink and make that the 100th game. :thinking:

4 Likes

Well, I hoped to play the audio version ! :slight_smile:

I have the means to experience it in its original form through an old phone I hacked with a RPi inside, and the prospect of adding a new original game to what it’s capable of playing fills me with anticipated joy.

old phone, open, with apparent wiring

10 Likes

Children on journeys might enjoy audio IF, and this would remove the distraction element. This also has the benefit that stories as a whole (and individual segments) are short enough that the children are likely to remember any formatting rules for selecting the next action/choice (e.g. “[nickname], [choice]”, “[shape], [letter]” or “[name], [action]”) the particular IF might impose.

The other obvious audience is people who have visual impairments or are blind, though a lot already play IF using screen readers.

3 Likes

An obvious parody game that comes to mind is a Phone Menu Text Adventure, where the player has to navigate an exceedingly poorly designed phone menu to save the day.

6 Likes

Wow, that’s super cool! Is there a 4-track cassette player hooked up to your rotary phone? (That’s what you’d need to play a 2-XL game in its original format.)

4 Likes

I’ve used the .wav tracks found on archive.org so far ! It’s digital only (but with the phone acting as the interface)

5 Likes

Hi there, I am the developer of the app mentioned above:

The app still in active development, i have some 200+ active users it seems and on last update i added some android voice configuration settings button for people who don’t know how to find those settings on the phone.

Just wanted to say that some of the voices on newer phones are pretty good, has anyone tried it?
You just need to download .z game files, after that the app will scan your phone for them (it needs permission to be able to find the games), then you can play it in a conversational way.

Can you give me some feedback about the app if you try it?
Thanks in advance

8 Likes

There were also many IF systems that attempted a primarily text-messaging format - either focused in conversation with characters, or giving them instructions remotely, like Lifeline.

Those might actually make good candidates for audio IF with regular text-to-speech and speech-recognition tools.

5 Likes

I thought that the gameplay of Infocom’s Suspended was something like interacting using text messages – created way before text messaging was a thing.

7 Likes

A little late to the party here, but this is along the lines of my rationale for creating a choice-based game system with emphasis on interaction.

Limiting puzzles to the kind that can be solved with mouse clicks seems to go hand-in-hand with the idea of the simplified user interface necessary for the games you describe. Additionally, since most of the “text” in these games are static outputs to the user, they lend themselves to be more readily converted into audio clips.

This solution wouldn’t address the matter of audio input, but since the requirement is to avoid too much typing, the point-and-click interface should cover that.

3 Likes

Let’s see some real examples, so we can learn from others:

Escape the Gloomer is voiceover by professionals. And it shows. It has a parser oblivious to input errors. Seriously, it is a great effort, not perfect (as UI) but it deserves to be studied.

https://oreolek.me/skill-stamina-luck/

The Interactive History of Interactive Fiction documentary in Twine, contains the very first instance of a choice based IF by audio, playing by phone: “Fighting Interactive Scenarios by Telephone”, or “F.I.S.T.”.

The format was invented by Steve Jackson in 1980 (I think) and it still holds up. It is compelling to play (although it has a very silly story).

You can test it playing that twine, press the * red link.

This is the Fighting Fantasy classic Deathtrap Dungeon voice over by Eddie Marsan.

It has a strong storyteller at the fireplace mood. I still need to play it, so I can’t tell you if it works for me or not.

Of course, some voice assistants has already several games like this, but I feel this genre has not taken off yet.

6 Likes

In 2017 I made an Android app (GIKI, ГИКИ, in Russian) for playing interactive audiobooks. The site as well as the app in Russian, but if you interested you can Google translate it or just ask me.

The interpreter took the plain text story file. The structure of the file was very simple:

1
Paragraph
2
Paragraph
...
etc.

You could easily navigate between paragraphs by calling it number. In fact you could try to save your gamebook pdf to txt and play. It had dozen of built-in features like dice roller, progress status, undo, inventory, paragraph notes, etc. The app supported both voice and keyboard input. It’s free, of course.

5 Likes

In terms of voice recognition accuracy, which is better? A, B, C, D, E, F, etc., or 1, 2, 3, 4, 5, 6, etc.?

I was trying to find information on this, but searching for this is cursed.

If one is more distinct to voice recognition and results in statistically better results, it seems like knowing to build a multiple choice game around that option would be helpful.

1 Like

It depends on the recognition algorithm which is on the user’s side, right? What can you do about it?

1 Like

Anecdotally, radio communication had enough trouble with A, B, C… that people switched to Alpha, Bravo, Charlie, but didn’t have that problem with 1, 2, 3.

Which makes sense, since more than half of our letter names are just one consonant plus an “ee” vowel, while the numbers don’t even share a vowel until you get to 5 and 9. (ə, u, i, ow, aj, ɪ, ɛ, ej, aj…that’s more vowels than most languages even have before hitting a repeat!)

8 Likes

Though now that I’m thinking about it, a game that presented the options as Alpha, Bravo, Charlie, Delta, Echo, and so on would be very easy to recognize and would also add some great flavor…provided it’s a story about a World War II fighter pilot or something like that.

4 Likes

Well, I was thinking about that. I’d love to choose the set of words with the lowest failure rate for speech recognition. I couldn’t find exactly what I was looking for, but I did find that speech recognition has the hardest time with monosyllabic words, which includes both numbers and letters. Something like the NATO phonetic alphabet should be easier for speech recognition to accurately recognize. They even have a standardized pronunciation guide to eliminate confusion with accents.

The thing is, this phonetic alphabet was made for optimal human recognition. I can’t help but wonder if this could be further optimized for digital speech recognition. For example, Bravo, Echo, Kilo, Romeo, Tango all ending in -o. Could alternate words be chosen to increase how unique the combination of syllables is to the program?

Furthermore, a speech recognition program that was tailored to only recognize items from this idealized alphabet would have fewer variables to consider and could be potentially more accurate within its limited word set than a more general purpose speech recognition program might be.

I’m just imagining someone driving, and choosing options verbally for an Audio IF game. This might reduce problems caused by synonyms, accents, homonyms, and rhymes. Saying one thing and having the game do something else is definitely the worse case scenario and would likely sour someone’s experience.

1 Like

Hi there!

As for where THIS idea went (Augmenting 100% of the text with recorded audio - #14 by CaptainEdgecase):

Tons of local code and various learnings of lessons, mostly focusing on leveraging Vorple.

I think audio-IF has a lot of potential, just the level of effort to do it well (as hinted at in the discussion above, from various perspectives) is exponentially higher than pure text.

But best case, some of us will make it happen, at least for some small, rich games.

4 Likes

It might be. I like podcasts and audiobooks, consuming these with much more enthusiasm than plain text OR video with subtitles.

On the other hand, I’m not convinced that text to audio is quite ready to replace humans in roles where nuanced dramatic voice acting is important.

3 Likes