I think it is a design question.
When I layout with facing pages I try to set the images so the upper edge is aligned with Cap Height of text on the baseline grid of the same spread. That means for me that I try to get (depending on the font) the upper border aligned with text and image as well. So I choose the baseline grid that the Cap Height is aligning with the border. Because I work with mm and InDesign with Inch and DTP-Points I have to take care on internal rounding problems (e.g. sometimes InDesign tells me that a 5mm distance is to small and should be at least 5mm, that is caused by rounding errors to the metric system). So I add always small values to get rid of those messages.
In my opinion (and there's 2 long shelves of books that I've designed
over the past few years sitting in the bookcase behind me), the best
choice is "fixed."
The main reason is this: any of the other choices can lead to subtle,
but significant misalignments -- and this can cause the page to be 1
line short, depending on the text that appears on the top line of the page.
Why? Because everything except "fixed" is dependent on the font that
happens to appear on the top line. So clearly, if you happen to have a
larger font in the middle of the book text (not just subheads -- perhaps
for some reason one word, or even one letter, is enlarged somewhere in
the book; or, if you've chosen "leading", then the leading might be
different for one character), and that letter happens to fall at the top
of the text frame, all the text on that page will not be aligning to the
So let's say you fix that for the one page. But then there's a late edit
-- and that top line gets bumped to the second line on the page. Now
everything on that page is badly adjusted again, and the page has to be
You might say, How often, in a regular book, is there a larger word in
the middle of a line? So here we come to the most insidious problem with
the other settings: italics.
Believe it or not, the x-height of an italic font is not necessarily
identical to the x-height of the companion roman font. It can be
slightly taller than the roman, even on well designed pro fonts. This is
a design decision of the font maker.
So, if you've aligned everything perfectly to the baseline grid, and set
your text frame first baseline option to "x-height", and you've made
sure that the depth of your text frame is a perfect multiple of your
baseline grid, what can happen is this: The first word on the page is
italics. So all the lines are imperceptibly shifted down a fraction
below the baseline grid. Therefore, the LAST line on the page will be
bumped off the page because there won't be quite enough room there, and
the result is that your page is one line short! And good luck finding
the cause of that (it took me ages first time because I never imagined
that the italics x-height was taller than the roman)! (Of course, now
I've told you the secret!)
And once again, if you have a late edit to make, you'll have to check
the entire book to see that no pages are inadvertently short.
And if you decide to be clever, and make the text frame (or margins)
just that little bit deeper to accommodate the necessary extra space for
italics at the top of the page, it will mean that your footnotes (if
there are any) won't be aligning perfectly to a regular line of text on
the facing page (admittedly by an almost negligible amount).
So what are the cons? Why not use "fixed"?
The biggest reason against that I know of is that for some reason, using
"fixed" disables the option of keeping any paragraph rules within the
text frame. Keeping a rule (or line) within the text frame can be a
useful way of making some text start lower down on the page (if you use
an invisible white rule). With "fixed", for some reason, this option is
Another reason against is that with "fixed", if an inline graphic
happens to fall at the top of the page, it will often end up floating
above the text frame.
But for books without graphics (and even with), I would recommend
"fixed" as the best, most predictable option.
I haven't done as many books as Ariel, but I think it bears mentioning that running a line short, or having a misalinment at the bottom, in my experience, is usually unrelated to the first baseline position.
Misalignment is caused by shifts in leading (as mentioned) which might be caused by adding a block quote or some other differently formatted paragraph or line, or by adding paragraph spacing that is not equal to a full line space.
Whitespace at the bottom of a page is usally the result of keep options kicking in. INn those cases I usually "cheat" by making the frames on both pages in the spread one line longer or one line shorter to try to move the paragraph withthe problem keep option off it's current position. I find this to be less obvious to the reader than a blank line, but it's a matter of personal taste, I think.
It depends on what comes at the top of each page.
If you are designing plain-text-only books, you wouldn't notice the difference between the options! For ease of calculating how much text lines you can fit on each page, I'd recommend Leading -- then the total text frame height is simply number of lines * leading.
However, if you have freely-positioned material such as tables or figures, that may appear at the top of a text frame, and where there may be text on the verso, you need to take different font sizes into account. If you use Leading and your caption text at the top is smaller (and thus should have a smaller leading), your table text would appear higher than the text on the reverse page. So in that case you could use Ascent or Cap Height -- but then you'd have the *baseline* visibly higher.
I prefer using an Object style for floating matter, so I can adjust the top position of floating materials. Depending on the font, font size, and sometimes other graphical elements (top lines etc.), I adjust the top distance so I get either the top of capitals *or* the baseline equal to "regular" text -- whatever looks best.
Thanks to all for this great advice.
From what I think I am gathering, the top line on the page can be taken into account to determine a suitable baseline option. Since the top line of my chapters have a large font, it may be appropriate to use the leading option.
For example, here is a page using the fixed option:
Here is the page using the leading option:
Most pages line up well automatically now, but a few need to be fixed. Using @Peter's advice, I can fix most of them by adding or subtracting a line from the facing text frames. But what is the correct way to handle it, when even that doesn't work? Here's a 30-second video with an example:
Thanks very much in advance for any info.
I realize this doesn't apply to all cases, but if you are in command of the text (meaning you are allow to split paragraphs), I would usually "cheat", splitting a paragraph in two where is makes sense. That way a line will be added at the bottom and I get my two pages even. In your video, I would split a paragraph in your left page or if not possible there then in previous pages.