Could we actually talk to NPCs in future TES games?

Post » Thu Dec 13, 2012 12:46 pm

I've been reading various tech blogs lately and I've noticed a popular trend. Smartphone devices, most notably iPhone, have been utilizing speech recognition software which is incorporated into a knowledge navigator program. If you've had the opportunity to use Siri, Vlingo, or Dragon Go!, you probably have noticed that this technology is advancing rapidly.

In addition to speech recognition and knowledge navigation, speech synthesis has advanced greatly as well, and will soon (5-10 years) be nearly indistinguishable from genuine human speech.

What if Bethesda incorporated a system like this into TES:VI or Fallout 4, where a player could simply approach an NPC and speak to them? The NPC could then draw upon a database (with lore, items, characters, enemies, locations, quest info, player achievements, etc.), and the game's AI could construct legible responses, that the speech synthesis could then deliver to you.

You could literally enter a shop and say "Do you have any swords for sale?" into your microphone, and the NPC would run a check on its inventory, identify "swords", check player achievements, construct a legible sentence, synthesize the verbal reply (all in about 2 seconds) and respond "Sure, I have several! I'm told you're very skilled with the longsword. I have a nice ebony one you might be interested in.", or perhaps the merchant could say "Sorry, I don't carry swords." You get the idea.

Please note, this system would not necessarily replace traditional voice acting. Main characters would still be voiced by professional actors, especially for important quest content. However, for non-scripted speech an actor's vocal likeness could be emulated.


PROS

- A fun way to interact with the game.

- Every NPC interaction would be unique.

- Every NPC would have their own distinct voice and accent. Next-gen speech synthesis software will be able to create very subtle differences in the tonal ranges, level of vocal pitch, and the variation of syllable stressing for unique accents and speaking styles. Filters can even be added to make a voice raspy, nasal, guttural, or a range of other distinct sounds.

- You wouldn't have the same ten actors used over hundreds of characters.

- You wouldn't have to hear the same lines repeated over and over again.

- You wouldn't be limited to what you can say to an NPC. You could ask them anything, and there would be thousands of possible responses.

- You could find information about a merchant's inventory without opening a menu.

- You could give followers detailed verbal commands.

- ZeniMax could package this system as middleware and license it out to other developers.


CONS

- You will need a microphone.

- You will feel very silly at first... and look even sillier to others.

- Increased potential for glitches and bugs... many of which might be hilarious though.

- Speech may sometimes sound unconvincing, which might be immersion breaking.

- NPCs may have difficulty understanding you.
User avatar
Claire Lynham
 
Posts: 3432
Joined: Mon Feb 12, 2007 9:42 am

Return to The Elder Scrolls Series Discussion