The different pitch accent patterns is easily the most noticeable phonetic difference when you look at the Kansaiben dialects (and it's important to mention that this is a group of dialects rather than a single dialect with no internal regional variation), so it's easy to conclude that this is the only real different in pronunciation between standard Japanese and Kansaiben.
If I understand correctly you are not interested in the many differences in vocabulary and grammar between the two dialects, so I'll stick to phonetics and phonology.
In the realm of basic phonology, modern Kansaiben seems to maintain more or less the same phoneme distinctions as standard Japanese. It doesn't have (or at least I'm not aware of) distinctions that don't exist in standard Japanese such as distinguishing an extra consonant or vowel sound.
When Kansaiben does seem to differ phonologically, it's usually deeper into the territory where these phonological differences may have already morphed into vocabulary differences. Kansaiben has a noted tendency to extend short, one-mora words (a lot of which are body part names), into two-morae words by lengthening the vowels. Oddly enough, these words also may have different pitch accents depending on the word, so めぇ (目, eye) has a LH pitch, while けぇ (毛, hair) has an HL pitch and ちぃ (血) blood bears a flat HH pitch. In standard Japanese there are only two possible pitch patterns for these words and they can both be properly distinguished only if some particle follows that word (because the pitch accent is relative and can only be distinguished when hearing the pitch rising or falling).
While short-vowels in one-mora words are lengthened, long vowels are sometimes shortened (as in the long O-form of verbs) and conversely again, a っ (small-tsu) sound is often replaced by lengthening of the previous vowels. But both of these changes are not as regular or sweeping as the previous one, and they seem to fall more properly into the realms of vocabulary and grammar.
Sometimes the boundary between phonology and vocabulary becomes quite murky, and this seems to be the case here. Even the lengthening of one-mora words, which does seem to be sweeping, is probably not an enforced phonological constraint anymore, since practically any young (as in younger than 50 years old :)) Kansaiben speaker can easily say these words without lengthening the vowel, and in fact most do just that when they speak standard Japanese. This is actually a tendency you everywhere in the modernized world, where mass media and higher mobilization has caused people to be bilingual in both their dialect and the standard language. This means that even if your dialect lacks some phoneme that is present in the standard language, when you borrow that word into your dialect you're much more likely to keep this phoneme as-is, since you can pronounce it anyway.
But there is one phonological feature of Kansaiben (besides the accent) which is strikingly different from standard Japanese and cannot be attributed to vocabulary: devoicing of vowels.
I think the video jkerian linked makes this quite apparent: when Tokyo Chicken and Osaka Dog pronounce 箸 (chopsticks), they don't only differ by the pitch accent they use, but also by how much they elide the last vowel of the word. Tokyo Chicken sometimes discards the vowel completely (saying [haʃ] instead of [haʃi]), but Osaka Dog will always pronounce it.
In general, devoicing of the vowels /i/ and /u/, which is so common in standard Japanese, is much more rare in Kansaiben. What's more interesting, is that even if it does occur, it may not necessarily occur in the same places as in standard Japanese, since the pitch accent does have some influence on whether a vowel gets elided or not. In this case, if Tokyo Chicken would have said 箸 with the same pitch accent as Osaka Dog (or if it just said 橋 or 端 in Tokyo dialect, like it did later), the vowel /i/ would never get devoiced.
