[Ruud-v-A]

That is what I expected as well, but it turns out not to be the case. See also https://www.ece.cmu.edu/~ece548/hw/lab3/faq-do~1.htm and https://encode.ru/threads/2922-Knusperli-%E2%80%94-a-better-....

[vanderZwan]

So could one define an "inverse" JPG standard that works with the current version but has the advantages that tialaramex spoke about?