ByteEfficiency reports no savings in animated GIF->video cases

[paulirish]

repro:

1. use new_audit: efficient-animated-content - use videos instead of gifs #4885
2. lighthouse --perf https://gif.ski/ -GA

the report says its 100 even tho there was almost 5MB of savings.

why?

The computeWasteWithTTIGraph reports 0 savings because there's only 1 network request affected and no long tasks touched. However taking the end of each graph, the difference is 23 seconds.

Patrick said one option is to calculate impact on end of graph, rather than impact on estimated TTI. Onload would also be indicative especially when the focus is on 1 or 2 problematic requests. Both seem good to me.

[patrickhulce]

Yeah I think we decided after the original change that we're just trying to report "load"ish savings rathe than on specific metrics anyhow, it seems like savings on load instead of TTI is what we want in most cases (except offscreen images, render-blocking, etc where we're telling folks to "defer" rather than "eliminate")

[kaycebasques]

I was going to suggest a GIF->video audit in response to Jeremy's new article, but it looks like this issue covers it

https://developers.google.com/web/fundamentals/performance/optimizing-content-efficiency/replace-animated-gifs-with-video/