4 ms·
Why is voice audio always so hard to get right? Seems like it should be easier than video because the data rate is lower than video but computer generated audio
by davidhyde 6y ago
Why is voice audio always so hard to get right? Seems like it should be easier than video because the data rate is lower than video but computer generated audio has always been terrible.
- 082349872349872 6y agoEither a possibility or a tangent: video is less time sensitive (think 24 fps film) than audio, which in turn is less time sensitive than haptic feedback. (Sentiment du fer may seem like a strange idea to us, but in disciplines where 40 ms is important, it may help to base reactions on senses which are more time sensitive than the eyes.)
- LudwigNagasena 6y agoMaybe because video is 2D, so it doesn’t seem real anyway, but audio is very close to real-life experience, so it is easier to notice the difference.
- air7 6y agoMy thought is that human speech contains a huge amount of delicate, not conciously noticable quailites that we are innately attuned to pick up on but find hard to explain explicitly. This is also true of visual facial expressions, but deep fakes know to avoid close ups on the face otherwise the deceit would be apparent, and we are accustomed to seeing faces from afar. Audio on the other hand, we just dont accept perceiving it badly: We'll ask the person to speak louder and clearer, turn up the volume, come closer etc.