Jump to content

Can't make decent HDR grade from Panasonic S1 V-Log. What am I doing wrong?


Owlgreen
 Share

Recommended Posts

I'm attempting to make a rec.2020 PQ grade from footage shot on the S1 in 6K open gate 200Mb/s long-GOP 420 V-log mode. I'm cutting in Resolve Studio and no matter what color management methodology I try, I'm consistently getting big macroblock cyan and magenta splotches in certain parts of my image. It's usually on gray or off-white objects in the picture. These splotches are only clearly visible when viewing in HDR. When the same clips are converted to rec.709 and viewed in that colorspace, the splotches are reduced to just looking like noise. I've attached a few photos I took of my HDR monitor itself while displaying the problematic images in rec.2020 PQ. Photographing the monitor showing was the simplest way to capture the splotches, as they disappear when the clips are viewed in rec.709/sRGB.

Generally speaking, I'm seeing more banding-type artifacts in clips shot with this camera than I had expected to see. I suppose I expected too much at these bitrates and resolutions, but I am left with the feeling the files are quite "thin". I would be interested to try one of the raw recorders with this camera, but I believe none of the recorders can record the open gate image, unfortunately.

For strictly rec.709 SDR output the camera seems to do quite well. For HDR, I am less sure. Does anyone have experience with these or related issues, and do you have any tips for me?  

IMG_20221217_172947_1.jpg

IMG_20221217_171126_1.jpg

IMG_20221217_171219_2.jpg

IMG_20221217_172902_1.jpg

Link to comment
Share on other sites

EOSHD Pro Color 5 for Sony cameras EOSHD Z LOG for Nikon CamerasEOSHD C-LOG and Film Profiles for All Canon DSLRs
20 minutes ago, PPNS said:

1. Why would you ever need to grade for hdr?

2. You are shooting in an interframe codec with 4:2:0 chroma subsampling, youre supposed to expect macroblocking and banding

Thanks for the incredibly helpful response, buddy! You probably get invited to a lot of parties!

Link to comment
Share on other sites

35 minutes ago, PPNS said:

2. You are shooting in an interframe codec with 4:2:0 chroma subsampling, youre supposed to expect macroblocking and banding

At the risk of feeding the troll, the S1 and S1H share the same 200mb/s 6k 420 10-bit long-GOP V-Log mode. The only difference is the AA filter on the S1H. S1H is "Netflix approved" for anamorphic shooting using the aforementioned open gate mode. Netflix requires all "Netflix Originals" content to be delivered as a Dolby Vision IMF packages. So theoretically the S1H is being used in this interframe 420 mode to produce HDR content for Netflix, and the S1 has effectively the same imaging capability. This tells me they must have figured out a color correction workflow which yields decent looking results. I want to know what part I'm missing. 

Link to comment
Share on other sites

Is anyone aware of any V-gamut/V-Log  to rec.2020 PQ LUTs? I've searched but haven't found any.

I suspect white balance has a a role, possibly a very big one, in these matters. There's an interesting thread here,

https://forum.blackmagicdesign.com/viewtopic.php?f=2&t=114403

about the "native" color temperature of the BMMCC sensor. It's not the most well defined concept, but I wonder what temperature this would be for the S1 in V-Log. In that thread someone writes, "As an Arri engineer told me, it's a property of silicon sensors that they all have their native optimum around 5,000 K". I will try 5000k and 5500k rather than balancing off a card or using one of the presets and see if that makes a difference.

Otherwise I guess adding film grain for the dithering effect could be helpful. 

Link to comment
Share on other sites

Hey Mr. Owl

It would be helpful if you could send some framegrabs in addition to the photographs of you monitor/tv. I wouldn’t recommend grading for hdr from any 4:2:0 codec, since the larger colorspace and luminosity range will only show the limits of said codec even more pronounced.

In general - any large uniform area with subtle gradients will show banding and macroblocking when you’re recording in a highly compressed codec. My advice would be to record in at least ProRes 422hq - maybe you need an external recorder?

 

Link to comment
Share on other sites

15 hours ago, Michael S said:

Looks to me as if a conversion to 8bit is happening before a conversion to a new colorspace is done. I'm assuming you are recording in 10 bit? Recording vlog in 8bit is no good. Maybe this is useful: https://business.panasonic.co.uk/professional-camera/varicam-eva1-color-grading-in-aces-davinci-resolve-tutorial

Thanks, Michael.

I'm using ACES with the V35 IDT, as per the recommendations. The amount of banding that results makes me wonder if there's a bug in the system. Maybe it's just down to not having enough encoding bandwidth. I think there's a degree of noise reduction baked into the files even at NR:0 which makes the problem more visible. Adding some film grain in Resolve helps to hide the issue. 

Link to comment
Share on other sites

2 hours ago, Owlgreen said:

I think there's a degree of noise reduction baked into the files even at NR:0 which makes the problem more visible. Adding some film grain in Resolve helps to hide the issue. 

If I remember correctly, the S1H has a -1 setting for noise reduction in vlog which disables all noise reduction, so yes, zero should mean some noise reduction is still active. And all these codecs are optimized to discard visually imperceptible detail. I'm sure Panasonic tuned these codecs to the best of their ability but when you are shooting in a very flat profile, some detail might be classified as "visually imperceptible" and get discarded even though after adding contrast, the detail wouldn't be so imperceptible anymore, and now you've got these flat, featureless (chroma) surfaces in your video.

I remember reading in some forum years ago that at some point a new camera from Canon was giving such clean images that banding acros an even gradient like a sky became vary apparent in the 8bit rec709 recording. The solution was to add a bit of gain while recording to hide it. So this became common practice with that camera.

Link to comment
Share on other sites

If 

On 12/21/2022 at 4:08 PM, Alexis Fontana said:

Why do you need to shoot Open Gate? 

It's a matter of preference. If the sensor is there, I want to have the option to use it. For some things I like the 4:3 aspect ratio and don't want to be losing sensor and lens coverage on both axes. If I wanted to shoot something with the Sirui 50mm 1.6x anamorphic lens, the S1 open gate would yield a 2.4:1 aspect ratio instead of a 2.84:1 aspect ratio were I shooting 16:9. I prefer the former. Obviously cropping gives a endless number of alternative aspect ratios for any and every camera, but as a matter of principle I like to crop as little as necessary. My ideal sensor would probably be square.

Link to comment
Share on other sites

On 12/24/2022 at 1:53 PM, Owlgreen said:

If 

It's a matter of preference. If the sensor is there, I want to have the option to use it. For some things I like the 4:3 aspect ratio and don't want to be losing sensor and lens coverage on both axes. If I wanted to shoot something with the Sirui 50mm 1.6x anamorphic lens, the S1 open gate would yield a 2.4:1 aspect ratio instead of a 2.84:1 aspect ratio were I shooting 16:9. I prefer the former. Obviously cropping gives an endless number of alternative aspect ratios for any and every camera, but as a matter of principle I like to crop as little as necessary. My ideal sensor would probably be square.

No offence but that just sounds uninformed. You have an idea that one arbitrary format is better than another because of principles you haven’t explained and sounds to me like you’ve bought into the false idea of more K’s = better. What good is a perfect, no crop desqueezed anamorphic image when you can’t grade it without banding and cheap artefacts. I really don’t get where you’re coming from.

Link to comment
Share on other sites

3 hours ago, Alexis Fontana said:

No offence but that just sounds uninformed. You have an idea that one arbitrary format is better than another because of principles you haven’t explained and sounds to me like you’ve bought into the false idea of more K’s = better. What good is a perfect, no crop desqueezed anamorphic image when you can’t grade it without banding and cheap artefacts. I really don’t get where you’re coming from.

👍 Cool story, bro. 

When did this place get so full of unhelpful trolls?

WTF?

Link to comment
Share on other sites

These spots sometimes appeared on skin filming in VLog with the GH5 in 4K 10bit 422 Long Gop in harsh daylight

I didnt reshoot, testing All Intra, though, because I saw the problem only later in post.

You could do a test open gate 10 bit 420 vs 16 to 9 10bit 422 and see if the later shows the same artefacts. If so, 6K+4K yikes!

Link to comment
Share on other sites

Thanks, Panny.

I will do this, but the open gate mode is the only mode in the camera which really interests me. Also I think the issue is more to do with compression and noise reduction than 420 vs 422 chroma. And all the modes max out at 150-200mbps.

What was your fix? I've been able to mostly hide the issues by adding film grain.

I do have a deeper appreciation for the fp's image now. 12-bit at 2400mbps is pretty uncompromised. Not always the most practical in the field, but I think it's worth the trouble if you don't need to shoot a ton of coverage. I look at it a bit like a 16mm film camera. I wish it too had an open gate mode, but you can't have everything. 

Link to comment
Share on other sites

No true fix but rather workaround by colour selection and changing the pink into the surrounding colour. I had the blotches one someone with reddish skin. With my own S1 I recognized some atomic reds and pinks under certain circumstances. I like the Lumix cams the best with tungsten light or light at that colour temperature. I don´t like them at daylight that much.

Link to comment
Share on other sites

4 hours ago, mercer said:

If it's a 4:20 vs. 4:22 issue, then you could desaturate the footage you already shot and the artifacts should disappear.

That’s kind of a big compromise - don’t you think? Mitigating a colour fidelity issue by removing colours is hardly optimal. 

 

23 hours ago, Owlgreen said:

👍 Cool story, bro. 

When did this place get so full of unhelpful trolls?

WTF?

Not trying to troll at all. I’m coming from a background of being a working cinematographer represented by a very decent agency, shooting commercials around the world and just had my first feature premiere at TriBeCa last year. I’m trying my best to understand why you feel it is a necessity and also hint in a non-demeaning way that you’re putting your focus in all the wrong places. But hey man - you do you. Looking forward to seeing your kitchen film👍

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

  • EOSHD Pro Color 5 for All Sony cameras
    EOSHD C-LOG and Film Profiles for All Canon DSLRs
    EOSHD Dynamic Range Enhancer for H.264/H.265
×
×
  • Create New...