I'll start a bit from afar.
Each segment ("sound", allophone) covers a lot of possible actual articulations. And, while it does seem like for [v] the spectrum is somewhat wider than for [f], their articulations are most similar and voiced/voiceless is the only difference which is stable across different allowed articulations. (As an aside, I would be interested in how you believe mouth to be positioned differently for [v] and [f]. If @Sergey Slepov's characteristization of lower lip covering upper teeth - presumably instead of merely touching them as in every labiodental - is right, this is certainly the first time I see it.)
Now, why is it relevant? Because literally the same holds for Russian: [в] and [в'] similarly do have more possible articulations than [ф] and [ф'] but the only stable difference is voice.