Not trying to bust your chops, but these types of photos depend upon being able to literally SEE, and read, and guess at, the expressions of the people in the shot. Without the ability to CLEALRY see the faces of the peple, we have to look at body language and big,broad gestures...we cannot literally "see" enought to "get" the feeling, the vibe, the details. For example, the woman on the far left of the frame in the top photo appears to be looking out a window, but in the latest crop,we see that she has a man to her left..he appears to be standing...she might have a name tag on...soooo...I guess she might be a waitress...what APPEARED to be a window in the top photo NOW appears to be a big,open doorway leading to another room, or more likely to an outside patio seating area. Again, not trying to bust ye olde chops Scuba...it's hard to say which shot I prefer since I cannot really SEE this picture. some shots simply do not translate well to being down-rezzed for the web...
This type of street photography image, with around 30 people in the frame (yeah, I counted 30 people before I gave up) just can not be fully appreciated at web size. For me, that sad fact is one of the single most-disappointing things about displaying photos on the web...we make these neat photos that look GREAT at 2,506 pixels wide, but which when down-rezzed just...sort of disintegrate into tiny pin-sized faces and stuff...
So, summing up--I do not know which of the two I prefer. The later 2x3 ratio crop adds a whole 'nother interaction, and what appears to be a second room, there on the left...in some ways I think it might be a fascinating interaction if the camera can SHOW us that woman's expression. But here, I dunno...both seem equally good. Now, if this were to be printed poster sized...