outerbeat

I did the test shooting with and without bitrate hack to see if it helps to eliminate macroblocking. Yes, it goes in right direction, but not there yet, and not because of hack itself it's only about sd-cards speed limits. We can use 150-190 mbps at the time, there is no SD card which can provide more. Yet. But still, it is a big progress. Even at 160 mbps there is a detectable visual upgrade compares to 80 mbps, but you need to know where to search. I think it's about CTU in HEVC, at larger bitrates codec is not eliminate macroblocks in picture per se, but replaces them with Coding Tree Units more efficiently than with smaller bitrates. At some point in post processing you can see when there's large blocks occur in 80 mbps file, but in 160 mbps file - only a "three" somewhat like "fractal blocks" of much smaller blocks. H265 use CTU for replacing macroblocks. So, overall quality is improve with bitrate hack, but macroblocking is not eliminated completely. Here is the test results, two heavily destructed sequences for more visual reference, left side - 150 mbps source, right side - 80 mbps source. Both sources recorded in UHD@30 with identical setting and environment, only difference is the bitrate. Scale around 300% for better proof and to eliminate influence of export bitrate. Forgot to mention that I did the test with NX1 It depends of scene that you shooting. Some scenes can't be record with small bitrate without macroblocking at all, they even close to crash with high bitrate option, but it works in far more cases than I did imagine before I start to use bitrate hack. Now I shoot only with it, but of course, if there new options appears, I'm ready to consider them in work. Also you can shoot with pretty much close to LOG settings and even create special NX LUTs for standart situations for exact yours setup. Basically you just need a few hours and some color checker for starters And yes, I'm also wants the real log profile and 10 or 12 bit color, for shure! But really I think that bitrate hack is just a first step in better overall quality path, bc without it I highly doubt that only 10bit color itself did all the work, like, how about beautiful 4K picture with 10bit color but H-U-G-E one helluva macroblock in the center of frame, like 480p-block? It might be the case very easily

So, according to two documents of this product released on BM site, here the deal. It can capture 38 minutes of 4K ProRes 10bit 2160p24 to 256 Gb sd-card (main page of product). In additional spec pdf "Compatible SD Cards" there is only two sd-cards, which can perform such bandwidth for 4K, and both of them only 64 Gb. So, basically you get like 7.5 min 2160p30 files on each card, and they are need to be replaced by anothers. Of course there is growing market of cards but you can't get good option yet, and to be clear - sd cards vendors never release crucial and particular info about this matter, so in most cases we spend our money playing charades with exact performance of SD cards, I have particular problem with this on my NX1 with bitrate hack (up to 320Mbps), it's very detectable and recurrent. Long story short that Write Speed, UHS-II, Speed Class 10 and so on means nothing in terms of capabilities in real applications, you just need to test this yourself. It's also indirectly confirmed by the "Product Manual" (page 20) on BM site, where you can see that top-speed sd-cards can't perform 4K, but only 1080p or even only 1080i but that cards have all the logos like UHS-II and so on. So for me it might be good but right now it's very dependable of unknown performance and capabilities of sd-cards. I think that it's easier to rely on SSD-based recorders, which didn't so limited on space, file lenght, bandwidth and also fine ssds cheaper of superfast sd-cards

Transcend SDXC 64Gb Class 10 UHS-I U3, 95/60 MB/s - write 35106 kb/s SANDISK Extreme Pro SDXC 64GB Class 10 UHS-I U1 (95/90 MB/s) - write 53710 kb/s Kingston SDXC 128Gb Class 10 UHS-I U3 (90/80 MB/s) - write 32615 kb/s