I purchased this camera recently. Here are a few impressions.

• It's tiny. Your GH2 or 3 is a tank comparatively.
• Image is really good. Straight out of the camera you get the equivalent of a film negative. An ugly, flat, bland image. But you can do anything with it in your NLE.
• Moire and aliasing, though present in some situations, is generally not a big problem.
• RAW files are too gigantic, ProRes however is what I use and totally okay.
• I hate cages, but I'm not sure you can do without one. The camera body is so thin, I could not get a decent mount, you need two connecting points. BH was selling the Wooden cage for $50 a few weeks ago. Highly recommended, it's dinky thing and really locks down your camera.
• batteries don't last, but they're cheap. I have 5 batters and they cost me $50.
• the LCD is pure crap. But is usable in conjunction with the $155 Kinotehnik loupe.
• Cage, Kinotehnic, and extra batteries are the basic kit IMO. Very affordable.
• 32 gig cards (20 minutes of footage approx) and the stock batteries pair up nicely to give a film like workflow. Change batteries and cards at more or less the same time typically.
• Watson dual battery charge is excellent if not overpriced. It can test batteries even when uplugged and read out how much charge is in your battery. Because of the short battery life, I consider a dual charger of some sort part of the essential package.
• This camera can do things your GH can't. High dynamic range and great codec. But it's not the right camera for all jobs. GH2 seems to compliment it well because of the GH's effecient file size. I haven't tried matching the images.
• Da Vinci Resolve is most likely not going to be turnkey for you. It's very picky about what systems it likes, especially with regards to video cards. I conceded defeat quickly. There are too many other options, although it probably is the Rolls Royce of color correcting software. But I don't need a Rolls on my 999 camera. YMMV.
• You can still do nice shallow focus photography with the s16 sensor. Easy. NOTE: Kinotehnik's loupe seems to be the only reasonable choice right now for non electronic vf. However, though it's made for the pocketcam, the thing itself isn't going to win any engineering awards. The mold they're using isn't 100% straight perhaps, it has a slight but problematic bow where the unit mates the bmpcc. The pocketcams data read out, showing fstop, iso, etc is tiny and the loupe is likely to obscure part of the readout. Very annoying. If you're on the fence about this unit, waiting for Zacuto or another manufacturer should be a consideration. The camera data is easily visible behind the mounting frame with the loupe removed, but attach the loupe and the loupe's slightly irregular shape will obstruct part of the data and make unreadable.

@Sarah839 If you select the Nikon mount version of the Sigma lens, then you will have full manual control over the aperture and focus of the lens on the BM Pocket.

I am working on one right now ... send me a PM

@Vitaliy Thanks so much for the link. A very enlightening read, indeed.

@CaptainHook So the results of Photo Numerique are all wrong?

@jonpais - i don't know, i don't speak french sorry. Just responding to the comment that the Pocket raw is not linear.. We've known since the BMCC first arrived it's stored as LOG. But John also said back then it unpacks in Resolve back into 16bit linear.

@CaptainHook I do read French, and from what I read in the article, it looks as though Black Magic has made grossly exaggerated claims about the dynamic range of the BMPCC. To begin with, exposing correctly proved extremely difficult for the reviewers (En pratique, l'exposition de cette petite caméra nous paraît laborieuse et peu précise). It also seems that the camera clips very easily in highlights (La première sensation, c'est que la BMPCC clippe rapidement les hautes lumières, même en RAW), which apparently is not recoverable in post. And just to clarify, the reviewers at Focus Numerique did use DaVinci Resolve Lite for editing. At the very most, according to the testers, the camera can do some 10 stops of dynamic range, and even that is stretching things. On the other hand, the camera did quite well in low light, and the reviewers recommend using RAW under these conditions for the least noise possible (Sur les très basses lumières, le capteur de la BMPCC s'en sort très bien). No wonder all the videos I've seen from this camera look so horribly flat and lifeless. But I'm approaching flame territory...

@jrd I'm not 'conflating' anything. I've never read anyone else remark about the flat and lifeless images out of the BMPCC before, only how wonderful the images look (and that goes for Black Magic's most recently released camera as well). And for that matter, I've never heard anyone else comment that most of the videos from this camera that have been posted online are poorly color corrected. Excuse me if I haven't seen those posts. Edit: I just Googled BMPCC and "flat, lifeless images", and it would indeed appear that if anyone makes an innocuous remark about 'flat' or 'grey' images concerning the camera, they are instantly burned at the stake. I should also add that I've yet to see a crisp image made with the pocket camera.

I admit that I'm a novice when it comes to video and that I don't understand all the terminology. I am eager to learn, and that's why I come to this and other sites. I also understand the principle behind shooting 'flat' in order to be able to manipulate the image in post. I would like to provide examples, but nearly every BMPCC video I've seen has been 'fl... and lifel...' I can however share an example of what I do consider to be a 'correct' video, with beautiful highlights, mids and shadows, shot with the GH1 and GH2 cameras.

@brunoalmeida Are you saying that this contrasty video is 'flat'? In that case, I need to either exchange my MacBook Pro Retina or rush to an ophthalmologist!

@jonpais and boring too! Try this one, I really like the grading

@brunoalmeida Give me at least ten minutes to download the Vimeo video. Impossible for me to wait ten hours for Vimeo to buffer...