I seem to be starting to figure this out. We've known for a long time that my hearing loss makes "normal" (i.e. sit in a classroom with 25 other people, teacher lectures) language classes mostly useless as a primary learning strategy, but I'm beginning to figure out what things work for me self-study wise. Rambly braindump notes follow; I may someday write this up in a better form.
The short version: since I understand spoken English via a Markov model, the trick to being able to understand spoken something-else is priming my Markov model - acquiring sentence structures and vocabulary words in a medium that works for me (text) and a "translation key" of how that looks and sounds like when spoken, because otherwise I end up with great reading/writing abilities and an utter inability to recognize spoken words, even if they're exactly the same as the ones I've written down.
I think of language-learning in three parts: written language, spoken language, and the ability to go bidirectionally between the two. I want spoken language - that's the final goal, conversation - but I suck at it, but I'm also really good at written language - so how do I capitalize on that and transfer some of that strength into my weak area?
Here are the resources I've found useful so far. This is not a final set of "thou shalt buy these!" recommendations - some can probably be made on your own for cheaper! However, this stuff has caused me to learn something about my language-learning strategies and I've found them useful.
Basic phrasebook: Survival memorization of a couple vital things like where's the bathroom? Which is completely not to be discounted; when I need to ask where the toilets are, I don't care whether I understand the grammar of the phrase or not, I just want to pee.
Pimsleur, Earworms: Helps with memorizing the contents of a basic phrasebook. I must know the words before I hear them on the tape; the tape is simply for drill - more on this later.
Basic grammar book: Also duh. Used as a reference rather than a study text - I generally touch the grammar book when I'm at the point of pain - what the hell is up with such-and-such a conjugation? - so the rules have context to grab onto when I learn them, and the book stays a happy "this causes moments of clarity and revelation!" item instead of "stupid thing I need to slog through."
Michel Thomas: Gets me through the basics of a grammar book with much less pain than simply reading the grammar book. I need to pre-prime my Markov model by reading the transcript before or during the first round of listening. Again, more on this later.
German reading book: This is a slim little book for scholars who want to be able to read in German (Goethe and Faust and opera lyrics and so forth) but don't give a fig about speaking it - this may be available in other common "scholarly" languages as well, but I haven't checked. (Spanish? French? Search for books or classes labeled "for reading knowledge" or knock on doors at your local university and ask.) It's like a grammar book, except written in a way you actually want to read - and it includes reading exercises and that sort of thing. It is much less comprehensive than an actual grammar book, and incorporates high-frequency vocabulary lists once in a while.
Books translated into German that I've read in English: Exercising my ability to guess in German. This is important - it's how I learned English as a wee tot. If you know the story, then you know the meaning of the words in the book, and what you're trying to figure out is how the meaning that you know already was expressed in German. I pick up an eclectic set of vocabulary this way; for instance, Der Kleine Prinz taught me the words for drunkard and thorns. I may soon start looking at translated versions of technical essays and books I've already read in English, or on topics I already know (thanks to Catherine Devlin for the suggestion). Note I said translated into. Especially if they're nonfiction, books translated into your target language tend to have simple sentence structures and no unexpected cultural idiosyncrasies and references to trip you up; for the importance of this, imagine what it would be like for some poor unwitting English learner trying to read T. S. Eliot's The Wasteland. Or anything by James Joyce. It's hard enough to figure out the grammar and vocabulary, so put as few barriers between you and the meaning of the text as possible.
Books in German that I haven't read in English: A test of how well I've learned to guess - basically an advanced version of the previous resource. I'm currently limiting myself to books that also have English versions so that I can check my comprehension afterwards, because my brain needs feelings of closure ("yes, I'm sure I understood that!") to feel satiated and reading stuff in English and going HAH! YES! gives me that closure and a big boost of encouragement. Without it, I get a growing pile of uncertainty that slowly eats my confidence away. Therefore: English versions too, please. For now.
Vis-Ed Flashcards: Drilling my ability to guess the meanings of more offbeat/complex vocabulary and sentence structures with minimal context. They're like mini-books; you can pick them up and put them down and do them in bite-sized pieces. Sure, books give you the whole story structure in your head to help you figure things out... but you also have to load all that into RAM before it's useful, which (at least for me) takes up the entirety of the few seconds I had to practice a flashcard anyway. I'll note that I skipped straight to level II for Vis-Ed flashcards (discovered a lonely box, heavily discounted, in the back corner of my campus bookstore) so I'm not sure what the first set's like.
Rosetta Stone: A bit of basic vocabulary memorization, but this was (surprisingly!) most useful at the start of my attempts to learn German, because I wasn't speaking - I was too shy. I mean, hell, I was dating a native speaker and I couldn't bring myself to try speaking German to him - it was just this mental block of oh noes I shall get it wrong frozen-in-terror-ness. But a computer? I can talk to a computer! I can repeat words into the speech recognition portion of a computer. I can do that while in the same room as my (very patient) boyfriend (who would barricade his ears with a WWII TV series so the gunfire and explosions drowned out my attempts to pronounce the number 12). And then I could ask him for help when I got stuck on a sound, because it was the damn computer's fault. And that meant I was saying German words in front of him (not really to him, per se - but in front of him) and gradually I realized that wasn't so different from trying to speak German to people and why didn't I just do that too... so in effect, Rosetta Stone acted as a (very expensive) mediating artifact that tricked me into talking with people.
Kaffeestunde: The weekly German coffee hour at Purdue. Most universities with foreign language programs have some sort of language lunch table or club or something of the sort; there are also meetup groups or language partners or just plain talking with a friend who speaks the language. This is conversation practice, which is the point of everything else - after all, you learn a language (usually) so you can communicate with people. So... communicate! (I was so proud the first day I managed to tell some other grad students about my trip to a bookstore in Portland - I didn't know a lot of the words in their responses and questions, but I could guess most of them, ask about the others, and understand the definitions they gave... without once resorting to English. There is hope for me yet!)
You'll notice some of these are audiovisual or audio resources (Pimsleur, Earworms, Michel Thomas, Rosetta Stone). That's where the "bidirectional transfer" part comes in. Basically, I know I want to understand audio, so I use these resources as goals - these are all-audio courses that you "don't need a book" to use (if you're hearing), so can I get to that point? What do I need to do to be able to use these as audio drills as intended?
The simplest way to do this is to either read the transcript before listening to the tape - or (more time-consuming but also more effective) listen to the tape while reading the transcript the first time. I suspect the reason the latter works better is that I'm naturally a speed-reader; I don't often subvocalize while reading texts (which is one of the first things a good speedreading program will try to train you out of - you're limited by how fast the invisible voice in your head can talk) so I get the words when I read transcripts, but I don't link them to audio; the word-as-written and word-as-spoken remain separated by a large gap in my head. And given that I don't want to lose my speed-reading ability or to inadvertently train myself to subvocalize all foreign-language words, I'll take the extra time to listen through the first round - after all, I'd like to speedread in German someday.
The awesome thing is that once I've gone through an audio resource, I have a good enough memory and enough residual hearing to be able to understand it without transcripts on subsequent uses. It's probably more memory than hearing; after listening to the first Michel Thomas CD twice, I could write out a reasonably accurate transcript from memory. But in any case, it becomes sort of like reading a book in German that you've already read (however long ago) in English; you know the plot, you're just waiting to find out how they express the meaning you already know in the language you're trying to learn.
Learning how to learn a new type of thing is awesome. I'm already thinking "whoa, how much easier is French going to be after this?" (Or Spanish, or Italian, or any other European language... I think Asian ones are in a slightly different class, but I have some strategies for those already.) Go go gadget metacognition!