S-Cinetone - Codec/Framerate/Drive Space Sweetspot?

rdbot wrote:

I'm on a Sony A7RV. I've read recently that if one is not using log then there may not be a lot of gain to shoot in 10bit or 4:2:2 and one can save drive space. Others say shoot as high quality as you can.

I'm a bit new to this and am looking to shoot with the best quality drive space sweet spot.

I would like to shoot in S-Cinetone and there are several variables.

• Should I shoot in XAVC-S or XAVC-HS?

I read that XAVC-HS would save me some space, but might be harder to edit on some systems. I'm on a mac mini M2 Pro, macbook pro M1, or Macbook Air M2.

• If shooting XAVC-HS at 24p, I have the option of 100M 4:2:0 & 100M 4:2:2
• If shooting XAVC-S at 24p, I have the option of 100M 4:2:0 8bit & 100M 4:2:2 10bit
• If shooting XAVC-S at 30p, I have the option of 140M 4:2:2 10 bit & 100M 4:2:0 8bit

I have a lot of older DV footage shot at 29.97 that I'll have to mix with in editing, so not sure if to shoot new material in 24 or 30, but want to take the above considerations and pick the best option going forward.

I don't know if it matters but most of my SD cards are only v30 so don't know if that matters. I still would like to have the best ability to grade later, but if it won't make much of a difference, then I'd like to save on drive space.

What are the pros and cons of shooting 100M vs 140M or 4:2:0 vs 4:2:2 or 30p vs 24p?

Thanks for the advice!

R

1. If you are going to grade, as you say, then you should not shoot in S-Cinetone. It is not meant for grading.
2. If you have no interest in grading, then 8bit 420 is fine.
3. If you think you will grade then shoot in Slog3 422 10bit. It is easy in post to transform Slog3 to REC709 and then grade to get the look you want. Or simply transform, no need to grade. It is just a setti g in the editor. Slog3 gives you the maximum dynamic range of the camera.

4.Your M2 Apple computers can easily handle XAVC-HS. They are designed to do so.

Markr041 wrote:

rdbot wrote:

I'm on a Sony A7RV. I've read recently that if one is not using log then there may not be a lot of gain to shoot in 10bit or 4:2:2 and one can save drive space. Others say shoot as high quality as you can.

I'm a bit new to this and am looking to shoot with the best quality drive space sweet spot.

I would like to shoot in S-Cinetone and there are several variables.

• Should I shoot in XAVC-S or XAVC-HS?

I read that XAVC-HS would save me some space, but might be harder to edit on some systems. I'm on a mac mini M2 Pro, macbook pro M1, or Macbook Air M2.

• If shooting XAVC-HS at 24p, I have the option of 100M 4:2:0 & 100M 4:2:2
• If shooting XAVC-S at 24p, I have the option of 100M 4:2:0 8bit & 100M 4:2:2 10bit
• If shooting XAVC-S at 30p, I have the option of 140M 4:2:2 10 bit & 100M 4:2:0 8bit

I have a lot of older DV footage shot at 29.97 that I'll have to mix with in editing, so not sure if to shoot new material in 24 or 30, but want to take the above considerations and pick the best option going forward.

I don't know if it matters but most of my SD cards are only v30 so don't know if that matters. I still would like to have the best ability to grade later, but if it won't make much of a difference, then I'd like to save on drive space.

What are the pros and cons of shooting 100M vs 140M or 4:2:0 vs 4:2:2 or 30p vs 24p?

Thanks for the advice!

R

1. If you are going to grade, as you say, then you should not shoot in S-Cinetone. It is not meant for grading.
2. If you have no interest in grading, then 8bit 420 is fine.
3. If you think you will grade then shoot in Slog3 422 10bit. It is easy in post to transform Slog3 to REC709 and then grade to get the look you want. Or simply transform, no need to grade. It is just a setti g in the editor. Slog3 gives you the maximum dynamic range of the camera.

4.Your M2 Apple computers can easily handle XAVC-HS. They are designed to do so.

Also note that going from 10 bits down to 8 does not necessarily save that much drive space.  After all, the camera is recording in CBR mode.

While in theory 10 bits is an extra 25%, the nature of lossy video compression is such that the actual increase in utilization will be less than that.

Plus, 100 megabits/second is 100 megabits/second no matter which codec or bit depth you use, and 100 megabit H.265 is more than enough at any supported bit depth.

rdbot wrote:

Thanks so much for the reply.

I guess I'm trying to simplify the whole process (shooting and viewing), and shooting in log seems complicated in terms of exposure and being able to see the footage w/o having to use an editor to see the footage properly or use luts which I'm not versed at yet. New to this and everything already feels a bit overwhelming.

I do have a ninja V but no idea how to set it up to view w/ a lut (or which lut I'll finally be using) and won't always be using it the ninja.

As far as grading, I imagine I might have to color balance or match and adjust things a little, but not trying to create any specific look. For that purpose, is your answer the same?

And if the computer is able to easily handle XAVC-HS, then why do so many mac youtubers XAVC-HS will give me about 20 min more per 64GB card and over the long run would need less backup storage. Can I view XAVC-HS just in quicktime and if I want to share or post to social media, does it present any extra hurdles or slow the computer down if I don't have a high-end system?

Best,

R

1. The Sony has a built in way of viewing Slog3 in standard REC709 while you shoot. It is just a menu setting. No problem viewing Slog3 shooting and seeing what you will get after transformation. It is a great feature, built-in. You do not need to do anything special with exposure. I shoot with Slog3 all the time. You can see my videos with various Sony cameras posted in this forum and in the Sony forums. Not complicated shooting in Slog3.

2. "Color grading" sounds complicated. It is not. In Davinci Resolve color management, for example, you just tell it in project settings that the clips are Slog3 and you want REC709. All your clips will then appear as if you shot in Standard mode. And it will convert nicely and automatically. And the clips will look exactly as you saw them in the viewfinder, given 1 above. That is all the "grading' you need; you can muck around more if you want with the colors, but it is an option. Not complicated making videos shot in Slog3.

3. Forget using the Ninja V. That does make things complicated with zero increase in quality.

4. You said you have an M2 Mac. That is a high-end system. On that computer you can play XAVC-HS clips in QuickTime smooth as butter. I post videos to YouTube in H265 (which is what XAVC-HS is) with no problem - you can see one shot in Slog3 right here in this forum posted yesterday. There are no extra hurdles with M1 or M2 Macs (any other Macs or laptops without M1 or M2 chips will struggle; most Intel PC's struggle). I edit on Mac minis (M1 and M2) and MacBook Pro's (M1).

Btw, the GoPro Hero 11 also shoots with the same codec as XAVC-HS, with a 180 Mbps bitrate and 10bit color. Works smoothly with an M2 chip (you can see my posted Hero 11 video here too). This is just to show you I actually use the equipment and features you are asking about. Not just an armchair poster.

rdbot wrote:

That is all very helpful, but doesn't one have to expose in a certain way for S-Log? Also isn't there a min. ISO or other things I need to set up properly? Never shot in S-Log and haven't yet begun to even think about which program I want to edit in as I haven't edited anything yet. (Next step)

And which S-Log do you recommend? I see in the Gamma Disp. Assist Typ. that I have S-Log2 & S-Log3 as option.

I guess I'm still trying to figure out whether I need the Ninja, but it is nice to have a large screen, also because I can get 5 hours or recording in ProRes LT on a 1TB SSD which costs me about $60 on Amazon. And also I'm shooting a backup to the camera when I have enough SSDs for short shoots. I also like the waveforms and false color and being able to record more than two audio tracks. But there are downsides too, like weight and batteries, and larger file sizes. And for run and gun, I wont want to lug it around. I got it on a trade and feel I should use it, but if the answer is to ditch it, then so be it.

So, to summarize, you're suggesting to shoot in S-Log, XAVC-HS, 100M 4:2:2 10bit, 24p (30p isn't available for some reason, and don't know if that would be better for me bec. I have a lot of 29,97 DV footage that I'll be mixing with).

But if I were to stick to S-Cinetone, what would you suggest then and would I be able to do some basic color correction with that setup?

I'll check out the vids you recommended now.

Cheers, and thanks again.

R

1. No special exposure for Slog3. You, again, will see in the viewfinder what the final video will look like. You also have a lot of latitude with 10bit color and the large dynamic range. People used to recommend special exposure (two-stops over!), but those were for cameras that do not have the capability of the latest Sony cameras. Just shoot normally being guided for exposure as you would normally.

2. The minimum ISO will be set automatically when you shoot Slog3. You do not have to worry about going below it.

3. Slog3 has the highest dynamic range, and you need to shoot 10bit. Those other Slog's are legacies that older cameras had that could not really use Slog3 because they were 8bit.

4. You can do anything you want to S-Cinetone clips, but they are not meant to be changed. It is a final look. It is not clear why you are fixated on that; there are lots of other profiles that do not need grading, including Standard (REC709).

5. It is too bad you cannot shoot in 30P using XAVC-HS. If you mix with other footage at 30p, maybe you do need to shot using codecs that allow 30P, XAVC (still with 10bit 422 if you shoot in Slog3). Normally, I would never shoot in 24p.

As I have said to you before, if you want more exposure aids and a bigger screen, you can get that with cheaper, lighter monitors than the Ninja V.

If you are worried about backup, if your camera has dual slots, you can shoot to two sd cards at the same time. That is a lot more efficient.

rdbot wrote:

Just wanted to add something, when I select XAVC HS 4:2:2 10 bit, I get a message that says, "A corresponding environment is required for playback of movies recorded in XAVC HS 4:2:2 10 bit". Does this mean a mac environment? Or something else?

Some computers cannot playback HEVC 10bit clips. Your Mac can.

rdbot wrote:

This is all great info, and I will do some testing now with the settings you recommended!

I think I've just been nervous to add more complication and S-log seemed more complicated than you have explained it, so thanks again for that.

Do I have to view the clips though in an NLE or can I somehow open them in Quicktime or another player just to watch what I've shot w/o having to import them into a project. I'd like to view them on the computer or show them to people. I dont even know yet which app I want to edit in and dont have time to learn how to do it.

I do have a final question though, why would you not shoot in 24 in general or did you mean just in my case w/ mixing w/ 30? In both cases? Someone on this thread mentioned pulldown, but I dont know what that entails or how complicated it will make things. Why do some people swear by 24 and others 30?

To your point, then, despite the savings in file size, you'd recommend shooting in XAVC-S, which allows 30p (not sure why it's not available in XAVC-HS).

Best,

R

If you shoot in Slog3, you must convert the clips to make them look good. In original form they look flat and colorless, not something you would want to show anyone.

24P does not capture motion well; 30p does. Cinema film is shot in 24P, so some people like the look. Computer monitor refresh rates are 30p/60p/120p. So, motion looks natural shot in any of those frame rates.

I actually now shoot only in 60p (except in very low light), which gives the best motion and does not require relatively low shutter speeds. But I do not want to start a discussion of that.

So, yes, I guess for 30p you can shoot in XAVC at the highest bitrates offered, 10bit or 8bit depending on whether you shoot in Slog3. It will look good.

Finally, you must use an editor to make a video, unless you intend to make a one-clip video. So, you will need to put your clips in an editor anyway, and then you will have nice-looking clips transformed.

If you just want to show people clips right from the camera, then shoot in S-Cinetone or Standard REC709 in 8bit 420. Your clips will look good and you never have to do anything else. But then, you might as well use a cellphone (they now take very good-looking video right away, automatically - no fuss, no complexity because they use artificial intelligence in place of your own). With your camera, and some learning and a little work, you can produce some great video, better than any cellphone.

rdbot wrote:

Hi again.

I had a question. After diving into shooting w/ S-Log3, some say that HLG3 is even better. Do you have an opinion on this?

Cheers,

R

Was that discussion based on an 8-bit camera, or a 10-bit camera?

NEVER use S-Log3 with an 8-bit camera.  It tries to cram too much dynamic range into too few code values.  I had a plot comparing S-Log2, S-Log3, and HLG in this regard, but since DPR's search is kinda garbage, I can't dig it up at this point.

S-Log3 on an 8-bit camera is universally worse than the human banding perception threshold across the entirety of its dynamic range.

S-Log2 for luminance only is in theory better than HLG (which assigns too many code values to the highlights part of the dynamic range), but you cannot mix S-Log2 and Rec. 2020 gamut.  S-Gamut/S-Gamut3 are also too wide for 8 bit recording.

S-Log2+Rec. 2020 would be the ideal combination for 8 bits but is not available.  HLG + Rec. 2020 is the closest you can get.

In theory, S-Log2 + S-Gamut with the "saturation" value cranked up to max might work (I've seen a bunch of YouTubers say they use this and it makes sense), however Sony does not characterize EXACTLY what the "saturation" adjustment does, so you would need to profile/characterize the altered gamut with a ColorChecker to properly transform it.  I was starting to go down that path, but switched to an A7M4, which supports 10-bit recording, eliminating all of the above concerns.