Seriously! I don't understand why people waste so much effort presenting information in such a high-overhead, low-bandwidth format. Wouldn't it be easier to just post the notes and skip all the AV folderol?

An echo of the Youtube effect; it's less work for the producer. Just turn on the cameras and start babbling, no editing, annotating, summarizing, or other further work required.

(tbc not faulting the speaker here).

I think these are relevant:

http://web.eecs.umich.edu/~gurevich/Opera/209a.pdf

http://web.eecs.umich.edu/~gurevich/Opera/210.pdf

Can we replace the article link with these? :)

Would be cool if a web app existed that could take a video and produce not only a transcript but also a series of intelligently selected images from the video, basically producing a full document that stands alone.

I feel like transcribe + pull out slides would be a generally useful thing for conferences. Even a specialized piece of software that has the original slides and just records when slide changes happen and inserts the appropriate slide into the transcription.

I’m convinced this would help conference videos that frequently do just have points where just the slide is shown (WWDC videos have this a lot - where frequently the speaker isn’t being shown). There should clearly be a standardized video codec for exactly this ;)

Pay me and I'll do it for you. If there is enough demand I'll look for ways to automate it.

My coworkers are currently on this same bent and it drives me mad.

They’d rather make a video than write an architecture document. And what happens when someone has a question or complaint about the middle of the video? How do you come back in a week to review the second to last part?

Learn to take screenshots and do some light typesetting. Spell it out for people (in my case, they are literally getting paid to do the work).

Listen at 2x speed and skip ahead?

Doesn’t work for people who can’t listen.

Doesn’t work for people who can’t see.

Doesn’t work for people in environments where the above two are essentially true.

25 minutes is still longer than 10 minutes (transcript, slides would be even shorter although less informative), and I was being conservative about reading speed.

People who can't see whould be unable to read as well

however, if it were written in text, they can read by using a screen reader, which is very established techology by now.

This is not (as) available for a video if at least some component of the lecture requires being able to see

>listen at 2x speed

I'm trying 1.5x, but then when he is speaking it is too fast

>skip ahead

I skipped the introduction, but the standstill periods are too short, and you have no way of knowing when something interesting will happen, so skipping ahead entails missing things. This is where a copy of the slides is particularly useful: you can identify those slides you don't comprehend, or of particular interest, and sit through the speaking just for these portions.

Listening to things fast is a skill - you can do it but it requires practice. If you know any blind/vision impaired people who use screen readers you should ask them to use speaker for a bit while they’re browsing/navigating etc.

It is absolutely incredible how fast some of them have the audio going at. (It also makes any pausing or lag in your application excrutiating. It turns out 10ms being “noticeable” is more a factor of how you interact with an application.

They've practiced with their specific text-to-speech program. The more familiar you are with a voice the faster you can understand it. The skill won't completely transfer to different speakers.

For their standard reading speed that true (but it’s still an amazing thing to hear, and I think more devs should communicate with blind people just so they can grasp the sort of latencies involved).

I would expect that the skill does translate to them being able to understand arbitrary speakers sped up more than a regular person could manage. I’d be curious if any HN readers using screen readers could confirm/tell me I’m talking nonsense :)

>I would expect that the skill does translate to them being able to understand arbitrary speakers sped up more than a regular person could manage. I’d be curious if any HN readers using screen readers could confirm/tell me I’m talking nonsense :)

I use a text to speech program at ~850wpm and I also watch lectures sped up to x4 regular speed. So yes, it does generalize.

I think it may also be the case that (old-school, non-concatenated but generated) synthesized speech, with practice, is more intelligible at higher speeds than natural speech. The sounds are more distinct and consistent.

The speech synthesis for screen readers is heavily tailored for that job (I remember the emphasis put on the new voiceover voices on Mac many years ago, and they apparently were noticeably superior to people who used them as screen readers at the high speed settings)

How do you scan forward for the interesting stuff? How do you search for a quote a friend sent you? What if you’re deaf? Or blind and can’t read the slides?

Just transcribe it and include the slides! If you can afford the speaker you can afford to make it accessible.

here are the captions printed out: https://txt.fyi/+/933087e8/