Jump to content

Calling all colourists - Grade Panasonic GH4 4K ProRes next to Arri Alexa 2K ProRes


Andrew Reid
 Share

Recommended Posts

  • Administrators
PSR20007b

gh4-3


The GH4's internally recorded 4K files can be converted to a chunky 2K 10bit ProRes 4444 image which grades extremely well.

To test this theory I call on colourists to have a go at it.

Arri Alexa ProRes files are available from their FTP server. Details below.

Read the full article here
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

Is converting ProRes LT to ProRes 444 going to actually give you more data to work with or "room to breathe" as you put it? Isn't it going to be the same data in a larger file size container?

You are correct there will be no advantage to converting to pro res 4:4:4. Any high end grading system will covert the footage to 4:4:4:4 dpx 10bit files to work with internally anyway. Going to test this footage later, shame its not full res as I would love to see how much secondary isolation I can achieve.

Link to comment
Share on other sites

You know what kind of footage I'd like to be able to play around with? Underexposed, poorly shot, or otherwise difficult footage that needs lots of spot treatment and reconstruction such as keying, masking, etc. When I'm color correcting well-shot Canon footage, I don't find myself running into too many walls. Of course I have a longing for better quality, and the GH4 clearly offers that.

 

I so often find myself shooting in poorly lit rooms, or rooms with straight-up gross lighting. There are also times where I just do a plain old bad job of exposing correctly. It's when I'm grading this kind of footage that I start to pull my hair out because there's not nearly enough latitude to make the corrections I need to.

 

I'm not explicitly asking you to go out and shoot poorly on purpose, buuuut if there happen to be any shots that you aren't technically happy with, it might be useful to throw those online as well.

Link to comment
Share on other sites

Thanks for sharing Andrew. Just loaded the footage and converted to ProRes 4444, it really shines. I really like how exited you are about this camera. Here we all thought the DSLR (small still camera that also shoots video) was dead. This camera seems like a perfect documenty camera. Looking forward the seeing the rest of your thoughts.

Link to comment
Share on other sites

I'm a bit worried sharing these, mainly because I'm paranoid I did something wrong on my end, regardless, I think it's good to share. First, I think the shot I used is a "worst case scenario" type of shot - because the shot itself is of a bland, bright sky that contains strong highlights and deep shadows (a real killer for 8-bit codecs as many of you know). Second, I think it reminds us that there is nothing like a true 10-bit codec, as opposed to 8-bit 4K being down-sampled to 10-bit 1080P (which in theory, works, but I think in reality it seems to fall apart somewhat - again, this may be a fault on my end). All in all, I think people shouldn't even begin to pass judgement on the GH4 until we see how it record in true 10-bit. I really hope that someone can show us true 10-bit out of the GH4 soon! I'd love to use the 10-bit 1080P at 60FPS that the GH4 is capable of to complement my Blackmagic Pocket Cinema Camera for slow-motion shots  :) 

Anyway, here are the shots I spoke about above. I simply imported Andrew's 4K ProRes file into Premiere Pro CS6, and did a light "Shadow/Highlight" recovery to see what kind of data I could get out of such an extreme shot - and guys, please remember that it is an extreme shot as it contains both strong highlights and deep shadows - I think any camera would struggle in this situation to be honest. However, the increase in colours that 10-bit brings would have definitely helped in the gradation I think. Also, I should point out - these are 8-bit JPEGS, however, I did render the shots as "Uncompressed 10-bit  YUV" (by way of Blackmagic's Premiere Pro rendering preset for Premiere CS6), and the results (in regards to the gradation) were the same. Anyway, enough of my ranting and raving, here are the shots:

ORIGINAL
33jljps.jpg

ATTEMPTED SHADOW/HIGHLIGHT RECOVERY
kb60eo.jpg
 

Link to comment
Share on other sites

The 4K to 1080 is (8+10+8)/3 = 8.67 bits; what you're seeing in practice makes sense.

I too saw limited DR when playing with these clips. Masking for the sky and applying noise can help reduce the banding.

Link to comment
Share on other sites

Not posting any images yet but in my testing so far I am seeing very little difference when I grade the 8bit 4k as apposed to the quasi 10 bit 2k footage. They both have banding when pushed too far at about the same point. Honestly the resolution loss is the most noticeable difference I see. I even programmed my own method for doing the down sampling to insure that the pixels were being combined in floating point format to get 10 bits of data.

 

What I am seeing is that having 10bit doesn't have near as much of an impact if there is no increase in dynamic range. Down sampling may give something similar to 10bit color but it doesn't increase the dynamic range captured in the scene in any way so it is not a noticeable improvement in what I have seen so far. 

 

Also it is not anywhere near the Alexa with 10bit log encoding, the amount you can push the footage is insane because the 10bit log encoding is being efficiently filled with more dynamic range rather than just higher precision colors. 

 

The 4K footage looks great but that is about it.

Link to comment
Share on other sites

Hi,

 

 

A quick grade test of the EOSHD shot GH4 rushes using DaVinci Resolve. First up a quick disclaimer, I'm not a professional colourist, though I'm often called on to do colour correction as part of an edit.

 

Workflow - Down sampled the 4K footage to 1080p QT ProRes 4444 in Resolve then re-imported and graded. I went for a warm look then applied the Osiris KDX LUT dialled back to around 50-60%. On the first shot I pulled a secondary on the skin tones to add some colour and tracked a window to subtly pull up the face. Couldn't do anything about the moire on the shirt though - just one of those shirts I think and it was only noticeable when the camera pulled focus.

 

Similar process on the other shots with some stabilisation where needed. On the last shot - the backlit smokey street scene I softened the look by using a layer mixer node set to Overlay and added a small amount of blur to one of the layers. I tried to recover highlights where possible using a soft clip curves adjustment. Overall the footage graded really well. No it's not Alexa and log or raw but there's a lot in there to work with if it's shot well and a huge amount of detail to play around with.

 

Looking forward to seeing some nighttime footage with a fast lens and some close up faces too.

Link to comment
Share on other sites

Here's a quick grading test mainly using my beloved Osiris LUT (Vision Color)

https://vimeo.com/90805624

 

First feelings :

GH4 footage grades REALLY well ! When downscaling 4K to 2K or HD, the image retains insane details.

Still a bit uncomfortable with GH' color rendition on natural green hues (foliage, grass etc.) 

Yet a BIG fan of lumix GH since GH1.

Being recently using 5DMK3 Raw I notice Panasonic color accuracy issues even more. I wish we could import picture style on GH4... (I mean Vision Color)

Link to comment
Share on other sites

Sounds about right dhessel- what we'd expect from 8.67 bits and no sensor-provided increase in DR.

HurtinMinorKey- is there an error in the math? Only green gets 2x unique samples in the Bayer array, red and blue are 1/2 resolution; Y is R*.3 + G*.6 + B*.1; only Y gets 4 samples to combine vs 1 for U and V in 420. If on-camera debayer is very good, the low res R and B will be better than 1/2 resolution, however at best we are still looking at 8.67-bit YUV after downscaling. After conversion to RGB, it will be somewhat less than 8.67 based on how good the debayer method performs. Thus not being able to see a difference matches what the math predicts.

Link to comment
Share on other sites

Not posting any images yet but in my testing so far I am seeing very little difference when I grade the 8bit 4k as apposed to the quasi 10 bit 2k footage. They both have banding when pushed too far at about the same point. Honestly the resolution loss is the most noticeable difference I see. I even programmed my own method for doing the down sampling to insure that the pixels were being combined in floating point format to get 10 bits of data.

 

What I am seeing is that having 10bit doesn't have near as much of an impact if there is no increase in dynamic range. Down sampling may give something similar to 10bit color but it doesn't increase the dynamic range captured in the scene in any way so it is not a noticeable improvement in what I have seen so far. 

 

Also it is not anywhere near the Alexa with 10bit log encoding, the amount you can push the footage is insane because the 10bit log encoding is being efficiently filled with more dynamic range rather than just higher precision colors. 

 

The 4K footage looks great but that is about it.

 

I can say that banding is very rare when grading 10-bit ProRes from my BMPCC, and I have so push it very, very far to bring out any sort of banding.

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...