- Feb 3, 2024
- 454
- 3,228
Renders are static renders plus animation run time * fps.What is your "Renders (including animations)" number indicative of? Going by the mpa71 Big Table if you're just adding the number of animations on to the number of renders, Episode 10 would only have ~6200, but the graph looks like over 6500, so I'm having a hard time understanding what that means? Obviously animations and renders aren't 1:1 so I'd understand if you didn't just add it on, but I don't really get where the number comes from otherwise.
Regardless, I don't think that the graph really shows that the newer episodes are "very proportional". They fit the trend if you include them in the trend (thereby weighing the trend more towards themselves), but if you created a trend line from Episode 2 to Episode 7 or even Episode 8, that line would be quite a bit lower than the plotpoints for Episodes 9 and 10. Even moreso if you just did the trend line for Season 2; look how flat that trend is between Episodes 4, 5, 6, and 7. This shows that DPC was on a faster pace in the episodes before Season 3.
I don't like including animations or giving DPC credit for having 25 minutes of animations vs past episodes where he has less than 10 anyway because oftentimes all that contributes to the project is a 30 second Halloween animation that could have been 10 renders. Or a Zoey running animation that could have been 3 renders. If I am gonna include animations, I think doing it the mpa71 way of counting seconds of animations is the best way, and if you look at that, once again, Episode 7 and 8 have some of the highest ratios of seconds Animated in the end product vs development time. Including animations can explain why its taking longer, but it doesn't explain why there's such a hyperfocus on animations if they're taking so long for so little payoff.
And keep in mind that Episode 11 (and part of Episode 10) is using tech that is supposed to be MULTITUDES better than what he used to have.
I think Episode 7 is the Gold Standard for what DPC can do if he has his priorities straight, and even Episode 8 starts to show a trend of favoring animations and unsurprisingly having a worse ratio of development time vs material gains than the Episode before it.
I actually just averaged out all the fps to 30 even though episode 10 lewd scenes were 60 fps, because the new hardware and the bloated framerate kinda cancel out.
That said, no matter what factor you multiple the animation runtime by, you still end up with roughly the same trend (unless you put a massive number in there).
Not sure what point there is in removing episodes to determine different trends. There are many factors influencing delivery times (render factors, game factors, hardware upgrades, and even external factors) that are not accurately factored in at all, so the point of collating all the episodes render data and trending it makes more sense the more episodes you include.
Of course you could say that if you just trend episode 2-7, then using that data, episode 10 should have been delivered 2 months faster. But on the flip side, if you just trend episode 7-10, then using that data, episode 2 should have been done in -22 days... There just isn't enough data in those subsets, so it doesn't make sense using it like that. The more data you include, the more useful it is.
But, if you scrap animation data altogether and just go on static renders, you surprisingly get an even better trend:

So, the renders really are dictating the delivery times. Which makes a lot of sense, since while he has renders churning away, he's working on the other stuff (like posing, coding, writing, music, etc.) in parallel.
I probably shouldn't have used the term "directly proportional", because there's no way they'd be, but the correlation is so strong, like, the coefficient is 0.97!
And episode 11 is running right in line with that. Using the formula for the trend (days = (renders + animation time * 30) * 0.0484 - 5.2547, (estimating 20mins of animation time since we don't really have those details yet, but we have animation count, and the big ones are done I believe), we end up with a dev time of only 9 days less than the time recorded.
TLDR; The time DPC is taking to deliver the episodes is directly linked to the number of renders. Number of renders is the strongest factor among all factors for determining delivery time within +/- a month's accuracy. And he's probably gonna just keep increasing the render count, so episodes are most likely gonna take longer and longer to come out.