Hi! I am a big fan of this site and of this forum, but I have never felt the need to post myself. There is however one thing about these new speedboosters for the BM cameras that makes me curious. I hope I can write this in a way that makes it understandable, as I hardly understand it myself. It is regarding the question of f-stops, t-stops and pixel vignetting.
So, I have a Metabones Speedbooster for Sony Nex and I bought it hoping that it would give me an extra stop of light (besides increase in field of view of course), and usually it does, but not at all f-stops.
Pixel vignetting (as explained in this dxomark article http://www.dxomark.com/en/Reviews/F-stop-blues) seems to be causing light loss at digital sensors at wide apertures. This is something I have experienced myself when using the speedbooster (but it is not the speedboosters fault, it is the digital sensor).
All of my lenses gain one full stop of light with the speedbooster up to and including f2.8. At f2 the lenses I have tested gain 1/2 to 1 stop (usually as much as that specific lens gains on that specific camera when going from f2-f1.4 without the booster). At f1.4 the speedbooster "only" gives me an increase of 1/3 of a stop compared to when I am using the lens at f1.4 without the booster. I have not tested any faster lenses than f1.4, but dpreview found something similar in their first impressions report (http://***URL removed***/articles/2667195592/first-impressions-metabones-speed-booster), and there should be even less of a gain when using lenses with maximum apertures of f1.2 or more (with or without speedbooster).
Anyhow, this might not be true for the BMCC and the BMPCC as I guess they have a lower pixel density than digital photography cameras, and, on top of that, the speedbooster might not make the angle of incoming light more oblique when exiting the pupil (my speculation) than when entering (so at f2 the t-stop could still be about 1.2 on the bmpcc, as intended).
I think it would be interesting to know how these new speedboosters behave in regards to brightness increase at wide apertures, if t0.74 really is possible when using a digital camera. Especially since low light performance could be a decisive factor when choosing between, for example, the 5d mark III, the BMCC and the BMPCC.
And as a side note I am very satisfied with my speedbooster, despite the minor issues at f1.4.