I understand that in a strict exercise all tendency tones should resolve properly. For example, 7ths resolve down by step in the same voice, leading tones usually resolve and always do so if in the top voice. However once you venture outside strict Harmony exercises do these rules still apply? For example if I'm writing an accompaniment or an instrumental piece on the piano then each voice is no longer sung by a singer and is played by the same instrument in which case does it matter that the 7th does not resolve in the same "voice"? Does it matter if there is an augmented second in my guitars chord progresssion which I have written for a singer? I am basically trying to understand which rules will still apply outside the realm of writing a chorale or 4 part harmony exercise or should I understand that these rules taught in harmony and voice leading from the common practice period really only apply to 4 voices singing independent lines in a choir? In a pop song for example the only two "lines" would be the bass and singer right? so if my singer sings a 7th then perhaps it would mean that the 7th should probably resolve down as taught in strict voice leading but if in my accompaniment there is a 7th chord does it matter that the 7th resolves in another voice?
The sort of 4-part vocal writing we do in 'harmony exercises' are based on the idea that the voices actually ARE independent singers, each one of which deserves if not an interesting melodic line, at least something that makes melodic sense.
When writing block chords we can certainly be less strict about a tendency tone resolving in the same voice. Probably still a good idea to have it resolve somewhere though!
Sometimes an accompaniment includes the entire melody, doubling and supporting the solo singer or instrument. That's fine. But it's also fine to have it specifically avoid the melody notes, and in particular avoid doubling tendencies and resolutions.
And although voice-leading may not be too important in a chordal accompaniment, texture is still an issue. Random shifting between 3-note and 4-note density can be noticeable and is probably best avoided.