Your examples aren't exactly the same. In your first example, you're actually missing a comma:
I heard his deep, warm voice filling the room.
You need the comma here because both deep and warm describe voice. What kind of voice is it? A voice that is deep and warm.
She was wearing a deep crimson shirt.
No comma here, and also no hyphen; deep describes crimson, and crimson describes shirt. What kind of shirt is it? A crimson shirt. What kind of crimson is it? A deep crimson.
I think '32-bit' being hyphenated is just because it's become a convention to write it that way. If you wrote 32 out as a word, you wouldn't hyphenate:
The thirty-two bit Pentium had a sixty-four bit data bus.
ETA: Upon further research, it seems that in the first example, you can either use the comma or not use the comma; either would be correct. (Source)
I found this example online that illustrates when to use the hyphens. Basically they are used to avoid syntactic ambiguity; when not using the hyphen makes the sentence mean something else, you use the hypen.
Jacob took the well-fatted calf to the riverside.
('well-fatted calf' as in a very plump calf)
Jacob took the well fatted calf to the riverside.
('well fatted calf' could be construed as a 'well' (i.e., healthy) and 'fatted' calf.
In the first example, the 'well-fatted calf' could be ill.)
This is similar to your added example about the well-known actress:
If you do not use a hyphen, you have:
The well known actress accepted her award.
In this case both well and known apply to the actress. This means she is well (not ill) and known. But that's not what you want to say; you want to describe to what degree she is known, and you don't mean to say anything about the state of her health. So you use the hyphen to make well-known into a compound adjective everyone will understand:
The well-known actress accepted her award.
In your deep crimson shirt example you don't have any ambiguity. A shirt cannot be deep, so no one is going to be confused if you don't use the hyphen. We understand inherently that deep applies to crimson and crimson to shirt. So it seems we use the hyphen to make compound adjectives when otherwise the sentence's meaning would change.
Sometimes when an adjective + noun combination is used very frequently and becomes very common, we make the transition from writing the two as separate words and make them one combined word.
So initially I'm willing to bet that "story book" was correct, just as "picture book" is two words, and "recipe book" is. It is a book that contains recipes, pictures, or stories. This makes perfect sense. But over time when the combination of words becomes very common, we can see evolution such as "story book" to "storybook". "Story book" is no longer correct; we always use "storybook". This hasn't happened to "recipe books" or "picture books", though, for which we always write the words separately. But check back in a couple hundred years and that might have changed. English is always evolving! There's no real way to know which words are written separately and which as a single word, except to learn and memorize them.
Regarding the hyphen, it would never be used in a case like this. We do not use a hyphen to connect the adjective to the noun in the simple case. A story-book doesn't make any sense. However we do use hyphens to connect compound adjectives which then refer to a noun. So if you are talking about a single book which contains two stories, you would call it a "two-story book". That is, a book which contains two stories. If you are referring to two items which are called storybooks, you would simply say "two storybooks". That is, these are storybooks, and I have two of them.
Best Answer
As per StoneyB's comment, the hyphen is effectively optional.
In OP's specific example, most people probably would include it, but grammatically the usage is no different in, say,...
...which as you'll see if you follow the link, most writers don't hyphenate.
Since language is primarily spoken, one must be careful when suggesting the presence or absence of the hyphen has any semantic significance. But sometimes there really is a difference - for example,...
...where it's possible the right (correct) mouse-click is in fact a left (or middle) click. This is one of those rare cases where (in principle) the written form can be less ambiguous than the spoken form...
Some native speakers will think they can unambiguously indicate whether they mean right=correct or right=[opposite of left] depending on stress patterns and pauses, but in practice many native listeners still won't be 100% sure what was meant.