Sony a7RII – 4K Video Test at North Wollongong

Second of a series of tests with the new Sony, looking at moire. In our previous test we showed the impact on full-frame recording in 4k (or UHD as it really is). In today’s test we looked at recording on the cropped setting, where no pixel-binning happens. This is our recommended setting for video with this camera, unless in extreme low light, as the moire is much less apparent. We also show some directly comparable full-frame footage, which again shows bad moire. These recordings were made with the out of the box settings, other than slowing the auto-focus to stop hunting. We also applied some colour correction. Our next test will be with an external recording device, but we think the noticeable moire on full-frame means the better video comes with crop. but then the field of view is reduced, suggesting that a wider lens is required for landscape and the like.

 

Sony a7SII 4K Sydney Test

A short test recorded in 4K, on the new a7SII, in and around the Central Railway Station in Sydney. We find the camera does a good job of handling difficult light situations, and the internal codec is ok for static shots, but falters with moving images, like passing trains. We found the image quality was good, with very limited moire, and so provides a better full-frame option than the a7RII. Graded with Film Convert, using default camera settings, and no picture profile. We will do some profile tests later.

 

Sony a7RII – 4K Video Test Sydney

Further test in 4K recorded in and around Sydney Central Business District. Used the crop setting, cinema picture setting 1 and auto focus. You can see the moire still in some of the buildings, and also the focus “searching” especially in lower light areas. Manual focus fixes the focus searching, but not the moire. It is a function of the sensor, not the codec, because the same artefacts also show on the Shogun. I conclude that the A7SII is a better bet for 4K Video.

 

 

Our Thoughts On Sony FS5, RAW and The Atomos Shogun

We have had the Sony FS5 for several months, and recently added the RAW upgrade, to use in conjunction with the Atomos Shogun Inferno.

Together they offer a range of additional functionality, including slo-mo, and capture to ProRes. The Inferno cannot capture RAW in native form. Of course you trade off the light-weight FS5 by adding the external monitor.

We did a few tests to examine the capability of the system and drew some interesting conclusions.

For this segment we compare the 4K RAW output from the FS5, recorded on the Atomos Shogun Inferno, at 50 fps which converts to Prores HQ, with the internally captured HD 422 footage. We use a side by side comparison (thanks to Doug Jensen for the idea), showing the 4k image alongside a scaled version of the HD version, with LUTs applied. We also include a 4 times zoom version, which shows the differences. We show both ungraded and graded footage. But we conclude that in normal usage, even when mastering to 4K, the internal 422 HD image is surprisingly good. The RAW version does capture more detail and colour (some thanks to the higher resolution) – and we found the internal 4k option on the camera performed less well. But the advantage of RAW 4K, in normal use is quite small, especially if mastering down to HD, and of course there is the extra costs of the RAW upgrade and external recorder. We also found the Shogun was a great tool to assist in the correct exposure of the SLOG2 S-Gamma image, essential to get the right exposure. Get the exposure wrong, and both images turn to custard. RAW is a good investment, but really only comes into its own for specialist purposes, or when cropping hard.

Further analysis and comparisons between the various RAW slo-mo modes, We use the Shogun Inferno to capture the footage, which is converted from RAW to Prores HQ. In this segment we compare the standard 4k full resolution standard 50 FPS, with the 2k 100 FPS and 200 FPS continuous slo-mo and also the HD burst mode (Not RAW) at 200 FPS. We have graded the results as best we can to make a series of side-by-side comparisons, and we also show the sequence at full 4k up-scaled resolution and also at 4x zoom. You should inspect the results, but our take is that the colour depth is not quite as good in the slo-mo modes, compared with the full 4k. We found there to be little difference between the 100 FPS and 200 FPS RAW modes. Whilst the resolution was better in 2K, the final results from the HD burst mode are very similar to the RAW versions. So we conclude the RAW slo-mo modes have the benefit of continuous recording, but the final real-world results are pretty much indistinguishable from the in-camera burst mode. All tests were done in S-Log2 and S-Gamut, and as we are a PAL country, all results were taken to a final 25 FPS time-line.

We think the RAW upgrade offers some extra flexibility, but at the expense of convenience and using an external recorder, but in output terms, it does not offer a big lift in final production quality. The Inferno is an excellent tool to get the exposures right. Actually the best feature is the more secure BNC connection over SDI compared with HDMI, as the latter cannot be locked in place, and is at risk of dropping out. But is the cost of the RAW upgrade and the Inferno worth getting a secure locked connection?