Understanding percent fill in Engrave Mode

I’m curious if anyone could explain this behavior, because I thought I understood what the percent values meant, but it’s not acting like I would have expected.

I’ve set various vertical and note spacing parameters, so that most of the time, the scores and parts just look done without touching hardly anything (which is, frankly, utterly astonishing to me. Holy moly!). But sometimes, I’ll see it have a situation like I just had, where it had a page at 85% full. I’d have a rest for a page turn if I just got the next system included on the page, so I “made into frame” to push that system in, and now it’s 99.1% full. If these are percents based on the “ideal” spacing choices, why didn’t Dorico automatically place the stave on the page for 99%? It’s virtually perfectly the ideal, yet it spaced it for 85% instead. Which looks fine too, but if 100% is exactly following the ideal spacings, shouldn’t it be aiming for as close to that as possible?

There are multiple things going on.
First, Dorico takes a rough first guess about how much space things might need. This does not take into account each and every thing around each staff, but is just an educated guess to make the “real” spacing faster and easier. This guessing can lead to results that are a bit off, since things might overlap in dense situations and therefore will take more vertical space.
Second, there could be things going on the next page that lead Dorico into not filling your page completely.

As Daniel has explained multiple times here, casting-off is a circular process: You know how much space a single staff will take when all casting-off is done, and you only know how all the casting-off can be achieved when you know how much space every staff needs…

1 Like

Well, I would think it would just space things based on the “ideal” settings, which would be an aim for 100% fill across a stave and vertically across a page. That would result in a predictable look that seems like it would be pretty straight forward math.

Except it occurs to me that 100% would probably make more sense as being something defined by the “minimum” required values, rather than the ideal ones. So things then have the allowance to squeeze tighter horizontally and vertically up to the minimum tightness - which would be 100% by that definition.

So I guess, I’m realizing that 100% must be defined by the “minimum” space settings, not the ideal ones? But then, what constitutes the ideal ones? It still seem strange to me that it would leave an entire stave on the following page when it could fit it in and still be in the green.

Does this answer that question?

The ideal gaps are the initial spacing of (empty) staves. The minimum gaps are for collision avoidance – to accommodate anything outside the staff, notes, markings, text, etc. Then vertical justification adjusts the gaps so the page is filled. If Dorico overestimates how many systems will fit, this can (but does not always) result in collisions. But it can also happen that Dorico underestimates, thinking the next system won’t fit when it would perfectly well. I wish this did not happen, but I don’t know what to suggest.

1 Like

Some images of the pages in question would help the discussion.

1 Like

Hmmm so it seems like what I imagined was it would run minimum gap requirements first (and thus have the maximum allowed on the page) and then expand from there to get as close to ideal gaps as possible. @DanielMuzMurray the quote from Lillie seems to describe this, which would explain why this was the process I thought it was.

That would, in theory, result in a tendency to over-estimate though as it would start with the maximum possible and then spread out from there.

But @Mark_Johnson is saying it goes in reverse, applying ideal gaps initially (but ONLY for staves, not for all objects?), which means that the number of staves on a page are determined by the wider, ideal spacing rather than the tighter minimum spacing. Then when it moves objects around to avoid collisions by applying minimum spacing, it pushes some things out and thus spreads out the staves beyond their ideal gaps, which would thus be eating into the “extra” space that was left over by the stave initially pushed off due to the ideal gaps function run initially. This sounds like a process which would tend to cause underestimation, as the staves are already beginning at 100%, and then as minimum gaps are applied to the various objects, the percent fill drops. Which does seem to fit what I see, where most staves and pages are well WELL below 100% initially.