Just getting back to the voice issue: for non-passerines and as far as is known the vast majority of sub-oscine passerines, song is genetically coded, not learned. So if two birds do not recognize each other, it means they're genetically programmed to a degree that they do not have a conspecific response. Of course, conducting playback experiments has to be done rigorously and one has to have contextual understanding of what family/genus is in question. IE, a lot of tapaculos will approach tape of any other tapaculo if they happen to be close. For oscine passerines, song is learned, so it gets a lot more confusing. You have cases like Golden-crowned Warbler where you have distinct populations with some morphological differences and wildly different songs where common sense says they are almost certainly congeneric but specifically distinct. But a good example of messy birds are House Wren, or Brushfinches (Atlapetes) where a lot of divergence times are relatively recent, the morphology / coloration is not a good indicator of how different they are, the songs are complex, varied, and all learned, and the genetics sometimes don't match up to what one observes in the field. It again reinforces that "species" is a human concept and evolution leads to things that don't neatly fit. But many/most of these cases are not "House Wren" level of complexity, they are dead obvious genetically coded songs that vary across geographic boundaries. The burden of proof is more on "maintaining clearly different birds as one species" rather than splitting in a lot of these cases, to be fair.
americanornithology.org
According to the plot there, the peak is 370, that may be achievable with normal cameras just with filters - on a bright day.
