In reply to jackie.

I don't fully remember the 90's and 00's -- I swear I was there -- but aren't there only 30 Seconds from Mars?

K3n.

I wonder if there's audio versions of adversarial images , like subsonics that a clanker will speech-to-text into something that would be unprofitable to add to their training data.

In reply to DiscoStu.

Sounds epic!

In reply to acb.

Luckily, SAM was ported to more easily available environments: https://discordier.github.io/sam/

Noticed that the poster says "9:30PM - 2:30PM". It's going to be a 17 hour dance party?

https://hawkingsvoice.com/

Just repeat: “Head on, apply directly to the forehead” for the length of the ad.

In reply to GnP.

I dunno. It's a video ad, which is a thing that they sell. Not an audio interstitial which is a different thing that they also sell.

Oy... I'm guessing this is part of the whole industry-wide "second screen" shit (ie. Netflix demands its writers have characters describe every goddamn thing 'cause it figures "viewers" are scrolling)?

Spotify's doin' this wrong. Well, they do everything wrong, but if this so-called "radio" company is trying to get spots that work for both audio and video, a certain Bay Area business cracked that code decades ago:

As dumb as it sounds, maybe they have a point? I would assume most people on Spotify are just listening to music and not looking at the screen. If your ad is only some Bruno Mars music that will play between 2 playlist songs, they may not even realize there was an ad  or what it was about.

Spotify probably has some hard data backing this up (as in, "90% of users are playing on their phone with the screen turned off" or something).

Or maybe they're just dumb, no way of knowing for sure I guess.

In reply to Lady MacBeth.

I am told that the overlap between people who listen to Spotify and people who might want to go to a Bruno Mars tribute dance party is quite high, but I don't know what the kids are into these days.

In reply to acb.

DECtalk.  To my eternal shame I failed to liberate one of these which was sitting in a cupboard where I then worked.

In reply to acb.

I was going to suggest MacOS say -v Zarvox 'whatever' but this is even better

Maybe it's for the best. What do you think the overlap is between Joe Rogan listeners and people who would go to the DNA Lounge?

I’d be tempted to get a primitive speech synthesiser to read out the event name (maybe SAM running in a C64 emulator?)

you need to get like a popular sports commentator to do a voiceover track of what's happening