Live stream chit chat

I read how your script thing worked earlier but this came to mind now, how does it work when it has been on a static screen and ads kick in and then back to static screen for long time, like how do the ads impact it as wont that reset the timer?
Also, since the ad audio is identical every time for each respective ad, I want to play with it and extend the script so it can tell the difference between an ad break and an actual censorship switch. That way it can roll ad time into the total automatically, and I can still break out the ad-specific stats on their own.
 
Also, since the ad audio is identical every time for each respective ad, I want to play with it and extend the script so it can tell the difference between an ad break and an actual censorship switch. That way it can roll ad time into the total automatically, and I can still break out the ad-specific stats on their own.
Nice! Would be a pain when they change the ad for the next daily episode preview though, luckily not many dailies left as we close to end
 
Also, since the ad audio is identical every time for each respective ad, I want to play with it and extend the script so it can tell the difference between an ad break and an actual censorship switch. That way it can roll ad time into the total automatically, and I can still break out the ad-specific stats on their own.
Oh god I woke up at 5am this morning with that Optus ad tune on loop in my head for a full hour.

That's why I haven't watched the feed all day. Thank dog it's their off day anyway.
 
Nice! Would be a pain when they change the ad for the next daily episode preview though, luckily not many dailies left as we close to end
Kind of, but not as bad as it sounds. Most of those promos reuse the same audio somewhere, usually at the start or the end. I can build the script so it looks for those pieces and flags anything that matches. Then I just check it once, approve or reject it, and the script includes or excludes that time automatically.

Live-stream audio is never perfectly identical unless it’s a pre-recorded ad loop, so as soon as a new ad plays twice, the script has enough data to treat it as a match.

This is basically what you see in an audio fingerprint: two different ads with similar openings. The top section shows the shared music and voice-over, stuff like “Tonight on BB” or “This week.” The more black you see on the right side, the stronger the match. The bar on the far right shows the similarity score, and anything above about 95% gets marked in blue as a match.

ChOt1o3.png
 
Censorship Stats for Today
  • Total censorship swaps (Fish/Wall/Bathroom Mirror/Outdoor Chair Cam): 29
  • Total time with both audio and video censored: 3 Hours, 30 Minutes 04 Seconds
  • Percentage of the feed that was either muted or on a censor cam:
    ~15.91% of the total broadcast was censored
  • Average censorship duration: 08 Minutes 05 Seconds
  • Shortest censorship event: 35 seconds
  • Longest censorship event: 32 Minutes 08 Seconds
Midnight hits while the feed is still actively muted, so these stats count up until midnight only

Edit: These numbers would also be skewed because it's from Midnight to Midnight, A lot of that time the house mates are sleeping. If I adjust for only the time the house mates are awake
~28.16% of the total broadcast while housemates were awake was censored
 
Last edited:
I get the sense that the muting is only going to get worse as we get closer to the finale.
They'll get their PR team onto it and explain some BS about how the final days will be filled with twists and turns that they want to keep a surprise for the daily shows. (That doesn't explain why it's been so bad this weekend, given that none of this makes the highlights shown on the Sunday live show, but they'll still run with it because they think we're stupid and that we'll buy that non-explanation.)
 
Back
Top