Panasonic GH4 and BMPCC Tests (short version)
All tests with Canon 50 mm f/1.8 set at f/5.6 on two different EF to MFT adapters. One is a dummy adapter, the other is so-called RJ Speed Booster. All footage at 25p except for GH4 4k at 24p.
In all the adjustments, no color correction and grading was made. All the adjustments were on the exposure (i.e. luma) levels to see the exposure and detail capture capabilities of both cameras at different settings.
The first reference clip: GH4, Cine D profile, ISO 800, 1/50, dummy adapter, 4K was conformed to FHD frame in Adobe After Effects.
The following clips are from Blackmagic Pocket Cinema Camera. The first set of clips were shot at the following settings on a dummy adapter WITH a slim fader ND 2-400 filter.
ISO 800 at 1/180 degrees
ISO 800 at 1/360 degrees
ISO 1600 at 1/180 degrees
ISO 1600 at 1/360 degrees
The second set (starting from 1:16) was shot at the same settings but WITHOUT the ND filter. To see the difference, a simple corrective adjustment layer was slid over the clips. The adjustment was made with Color Finesse on After Effects. It is a simple push&pull of pedestal and RGB Gain values followed by a subjective gamma adjustment. The same layer is copied on all clips and that's why you don't see an improvement over the first clip as the whole adjustment was made on the first clip (ISO 800 at 1/180 degrees). I later made custom adjustments on ISO 1600 at 1/360 deg. clip to save more highlights and am happy with the result that it can be corrected safely with no problems.
There was a serious problem we realized much later: we made the test on a snowy day in Ankara, Turkey but the cameras were positioned on the edge of a window and there was a heater just below the cameras, sending heatwaves upwards. That's why the footage was slightly distorted by the passing jelly-like heatwaves. This reduced the perceived sharpness of the sensors but had no effect on the exposure, which we were testing. We are happy with the result.
Further tests were also made to see the performance of FHD at 200 Mb/s All-I on GH4 and ProRes 222 on BMPCC but these are not going to be published. We may repeat the test under better conditions.
The final stage of the video (1:52) shows a clip to test the ETTR approach on the BMPCC. A rooftop with a bird perching on the edge was shot from a distance of 50 metres. All the histogram was pushed to the right, to the point where the first 100% zebras appeared. This was the highest point at ISO 800, 1/180 where you do not blow the highlights. It was impossible to see the screen at this setting and as can be seen in the video, there was very little detail to make out. In the post-production stage, a simple adjustment pass was made to see how much detail could be restored. As you see, BMPCC has captured a great amount of detail in this already flat scenario. The sky itself did not contain anything more in real life and that's why it looks flat (it was snowing). The building has survived rather well and as a test the video was a success.
