markr041

It makes a big difference, because you are shooting at much higher bitrates and sometimes with more bit depth and color sampling - less compression, more bits. All intra 400 Mbps is not close to ProRes 422 HQ, which is 1900 Mbps. Compression artifacts include banding and macroblocking that all the pixel peepers here obsess on. Also, you are bypassing a lot of other internal transformations of the data, perhaps the aggressive noise reduction. Also the GH5 cameras shoot 4K 60P at 8bit 420, but externally 10bit 422. With less compression, more color sampling, and more bit depth, the external recording will be much superior. Now, sure, when viewed via YouTube, which is 8bit and highly compressed, the difference may indeed be marginal. But still, claims about camera performance should be based on the videos it records alone. By this argument, there is nothing the GH5 has over Sony cameras that shoot 422 8bit other than the highly subjective "color science." And with a tinier sensor and sub par AF.

I have no quarrel with the quality of the GH5 or GH5s. But if those videos were shot using the Shogun Inferno, then all of the internal compression (and maybe the noise reduction algorithms) of the camera are bypassed. Compression is the enemy of good quality video, so shooting in ProRes at its extraordinary high bitrate (and 10bit 422, of course) really makes a difference. Many cameras would look as good if one is using the Inferno. The Shogun is the video recorder, not the camera. This does not make the videos worthless for evaluating the GH5, but it should be considered in drawing any conclusions about the camera (of course, one of the good features of the GH5's is that they work well with the Infernos).

Anyway, here's a video shot in-camera using HLG 8bit (!): It's not that bad...

We need a non-screen, small and light recorder that has the same recording capability as the Shogun Inferno from Atomos. Now, I power the Inferno and the FS5 from one battery, using D-tap. That lightens the load considerably. But the size of the Shogun still makes it an awkward combo. My other technique is to to just put the Inferno in a shoulder bag, attached to the FS5 by the thin SDI cable. I really do not need the screen.

This is a good discussion. While you can shoot HLG internally in the FS5, you cannot externally shoot HLG directly using the Atomos Inferno from the FS5. Internally, the FS5 is set to Slog2, even for SDI RAW output. One has to convert Slog2 to something; could have been HDR10 or REC709. And, the differences one sees are due both to the bitrate difference and the large difference in the amount of compression. All Sony cameras below $8,000 shoot in Slog2, 8bit. Sony has been touting this ability for many years. It's a feature allegedly designed for pros. So, if you want you can think of this test as a test of Sony's Slog2. But, again, Slog2 needs to be converted to a delivery format for viewing. And HLG is one of those. One can record RAW video externally from the FS5 and then convert in software to REC709, HLG, or HDR10 using 8bit 420 and 10bit 422 and different amounts of compression for comparison. That would reveal what matters and by how much. But then such a comparison would have little do to do with what Sony cameras actually deliver internally, which is the point. One could also record HLG internally and compare that with HLG converted from the RAW output in post, choosing different bitrates and bit depth.

Is 8bit HDR or HLG a "joke"? Here is an HDR HLG video shot from the same camera (Sony FS5) using its internal 8bit 420 Slog2 XAVC and Slog2 recorded externally from 14bit RAW to 10bit ProRes HQ (Atomos Shogun). So the 8bit is heavily compressed (100 Mbps), the 10bit is not (950 Mbps!). Totally unfair to the 8bit. But, let's see... Same scenes - for each the first version is from Slog2 ProRes 10bit, the second from XAVC 8bit SLog2 placed on the same timeline. The video was rendered in HLG 10bit DHxNR HQ in Resolve Studio and sent to YouTube. No color correction other than the Resolve presets converting Slog2 to HLG and level adjustments. It is a real HLG video. You can easily tell which is which, since the 10bit source from RAW is 4K DCI and the 8bit source is 4K UHD. The video has plenty of blue sky, ripe for banding. And lots of detail, and high contrast with those pesky white clouds ripe for blowouts. There is a difference. But does it matter? Is it fixable? A laughable joke? If you do not like the test ("the test is a joke," "the test tells me nothing," "YouTube converts to 8bit anyway, so what?"), suggest what you would want to see instead. I can do anything with the original Slog2 clips, including making them available for download. Finally, why do we care about this? ALL Sony mirrorless cameras and camcorders less than $8,000 that have log gammas (Slog2/3) and/or HLG only shoot 4K in 8bit 420 - the RXxxx series, the A7xxx and the A6xxx series, and the FS5. If shooting in log or HLG at 8bit to make 4K HDR videos is a joke, that would call into question what Sony is trying to sell, and what anyone should consider buying (note, the GH5 can only do 10bit 422 in 4K30p, not 4K60p). No Canon cameras/camcorders do 4K 10bit below $8,000 either I think. And they have log gammas. Is this a big deal?

That's right, and the M50 also cannot shoot 25p or 30p in 4K (only 24p). Anyone vlog in 24p? Cinematic vlogs? And I am sure its IQ will be inferior to that of the NX500 in 4K.

Do you have any comment on the video clip I bothered to render from your clip above (without a de-squeeze)? It is not "too blue". Is this what you envisioned it should look like, or not? Your snapshot is colorless and too contrasty compared to what I get. I like what I got much better.

Here is the REC709 version of one of the clips based on the assumption that the clip was shot in Slog: As rendered in Resolve Studio using Resolve Color Management.

1. The metadata from your file says REC709. 2. It looks like Slog, however. 3. I used your clip in Resolve, set to Slog input color space and REC709 time line and output color space. The rendered video looks beautiful. So, it is actually Slog. I will upload the rendered clip to Vimeo, but I am uploading another video now, so this will have to wait. I attached a tiff file from the rendered video. Looks good. Test_1.1.1.tif

Another snowstorm, more intense! 120 FPS 4K DCI.

The snowstorm aftermath: snowbells + tree destruction

Yes, umbrella when outside, also I look for overhangs..

It's not that bad a workflow but internal is my wish too (second-best: a tiny external recorder that has the same capacities as the Shogun Inferno, in 5" or without a screen at all). Right now I save a lot of weight by jointly powering the FS5 and the Shogun with one battery that fits in the FS5 and has a D-tap connector (Swit). Shogun without a battery is quite lightweight, just too big mounted on a handheld.

"120p in a 30p timeline? That must’ve been some fast snow." Yup.

Sure. We are getting off topic here, but when you connect up the FS5 to an external recorder (eg, the Atomos Inferno) you get some major specs: 12bit RAW for DCI 4K at 30P, or ProRes HQ 10bit at DCI 4K 60P or even 120P, all from 12bit RAW. You are out of 8bit 420 land (however important or unimportant that may be). These specs are just as good as the bigger, more expensive FS7, but less good than the upper end if you think that 16bit RAW importantly beats 12bit RAW! And there is that electronic, continuously variable automatic ND filter..

Exactly. But most of the people here are discussing and shooting with mirrorless cameras or DSLRs. The FS5 is not "lower end" in that category, it is in a whole different category - not cinema or broadcast but not consumer either.

Sony FS5 Slog2 RAW 12bit 120 FPS->Shogun Inferno Pro Res HQ->DNxHR HQ 422 10bit REC2020 30P HLG. Yes, you only can get 4-second clips.

In what sense is the FS5 "lower end"? Relative to ARRI cameras? The FS7 also shots HLG, speaking of the lower end...

HDR HLG 4K DCI 60P: Same as above but with Sony 10-18mm F4 lens.

Sony FS5 + Shogun Inferno=4K DCI @ 60P. Slog2->HLG in Resolve Studio with Luminance mapping.

HLG allows more room for highlights. Night video has bright lights surrounded by dark. With little DR (like for SDR), either all the lights are blown out (and thus lack color) or if shot to avoid blow outs, there is zero information in the dark part. So here is an HLG video of a street scene at night:

You cannot replace videos on YouTube, so here is the corrected replacement, with the other deleted, still 60P, 4K, HDR (HLG).:

Here is why I do not think there will be a new FS5 soon: there was an FS5 firmware upgrade in October that added, among other things, the HLG Picture Profile. This gave the FS5 a feature that only the newest generation of Sony cameras (2018 models) have - the new camcorders like the AX700; the A7 iii, the A7r iii, etc. The FS7 ii has this too. However, the A6500, the A7s ii, the A7r ii, the A7 ii did not get that upgrade, despite the fact that they are certainly capable of shooting in HLG. And, voila, we got the A7r ii and the A7 iii. So, which cameras did not get that HLG upgrade are the candidates for a new model, not the FS5. Note also that so far, no new Sony model does 4K 60P or 422 10bit internally. So, other than in autofocus (which, of course REAL videographers never use), the FS5 is up to date. Not to mention that externally it can do tricks no other camera can do except the FS7. And the price gap between the flagship mirrorless and the FS5 is shrinking. If Atomos finally offers a compact external recorder with the capabilities of the Shogun Inferno - no need for even a screen - the FS5 would be by far the best alternative - 4K DCI 60P and even 120P in ProRes from 12bit RAW in a relatively small and light package. Unless you really need AF, of course.