Mar 27

RED Scarlet-W Dragon vs. Panasonic VariCam LT – FULL Dynamic Range Comparison

 

Here’s a big one for you guys. This will be a series of posts regarding this over-under exposure test with the RED Scarlet-W and Panasonic Varicam LT. I not only compared the Varicam LT to the Scarlet-W, but also compared individual settings within both cameras, such as the RED’s Standard OLPF vs. Skin Tone-Highlight OLPF and the Varicam’s Native 800 ISO mode vs. Native 5000 ISO mode. Many more posts spawning from this test to come, so stick around ūüôā

 

RED Scarlet-W (Standard OLPF) vs. Panasonic Varicam LT Dynamic Range

I was recently invited to the Panasonic Headquarters to take part in a complete over-under exposure test with their new, not-yet-released Panasonic Varicam LT. The tests were conducted and produced by the incredible cinematographer and AFI alumni¬†Toshi Kizu¬†with Panasonic’s camera engineer Takahiro Mitsui present to take us through the camera and running the tests. ¬†The Varicam LT uses the same exact sensor as the full-sized Varicam 35; it’s just a new, smaller and compact body — the Varicam “Light” (hence LT). ¬†As far as I know, the Varicam LT has not been officially released and this is still considered to be pre-release footage and more tweaks to the camera are actively being made.

I brought my Scarlet-W along to partake in the festivities and spent a solid 12 hours in their studio shooting Xyla Dynamic Range charts, over-under dynamic range tests, and projecting them on the large theater screen to see how the two looked in its large format. We were extremely meticulous in controlling all the variables and adjusted exposure accordingly.

With the Red Scarlet-W coming in at about $15,000 and the Panasonic Varicam LT at about $25,000, both are in contention for a similar customer-base.¬†I’m not going to go into detail about the features for each camera, but you can read my RED Scarlet-W impressions in my previous post, and¬†the Varicam LT post will come in the future.

 

Xyla-21 Dynamic Range Test Chart

First off, using the Xyla Dynamic Range chart, both the Scarlet-W and Varicam LT have very similar dynamic ranges; approximately 14-15. 14 clear stops with the 15th being a bit amorphous, though I can make out slight borders for both the Scarlet-W and Varicam. Very close. Both these charts were shot at native ISOs of 800.

 

 

Testing Methodology and Parameters

This test was comparing the Red’s Standard OLPF (native 800 ISO) to the Panasonic Varicam LT (native 800 ISO). We conducted an additional test using Red’s Skin Tone-Highlight OLPF which will¬†be posted¬†in the future.

Each¬†camera’s dynamic range was tested up to 8 ¬Ĺ stops over and 8 ¬Ĺ stops under in increments of half stops. All the test footage was shot with the Leica Summicron-C 29mm at a common stop of T/4.0 and stopped up or down a half stop to adjust for exposure changes. In order to accommodate¬†the common stop of a T8.0 on the gray card, we had to adjust the camera’s settings accordingly, as the camera’s had different frame rate shooting capabilities. All the footage was shot side-by-side.

Below is the chart made by Toshi Kizu that was used to determine the settings for each individual exposure value and their equivalence to adjust for the different cameras. Please refer to this if you have questions regarding how the scene was adjusted for each exposure.

It’s worth noting that the Scarlet-W¬†was shot in 5K HD to match the Varicam’s Super35 sensor. 5K Dragon sensor is just slightly wider than Super35 (but same height) so we chose to crop to 5K HD (4800×2700) to match the Varicam’s 4K UHD aspect ratio (3840×2160). That means that all the Red 5K footage is downscaled to 4K UHD, though resolution was¬†not what we were testing.

 

camera-log

Scene readings as made by the spot meter

It’s best to think of the values in relative terms to N (T8.0). For example, her face reads +1, so N+1 is one stop above N, therefore, her face is one stop¬†above N.

 

THOUGHTS

I think it was a surprise to many to see how¬†well the Varicam LT performed compared to the Scarlet-W. The Varicam LT was able to hold greater highlight detail when overexposed with a softer roll-off. It was able to retain more color and detail information in the shadows when underexposed. The Red Dragon sensor is known to have a hard clip in their highlights and flat-line compared to say, the Arri Alexa’s ALEV III sensor, and it was clear that Panasonic took a note from Arri in replicating a softer roll-off. The Varicam’s ability to hold color information in the highlights was very impressive. Even at N+5 ¬Ĺ, the Red began to lose it’s ability to distinguish individual colors but the Varicam was able to hold its colors easily.

Panasonic’s V-Log is very flat compared to Red’s RedLogFilm gamma curve. Again, I think Panasonic took notes from Arri’s Log C and attempted to create a log curve as flat as possible as opposed to Red’s more contrasted curve. Even in D-Max and D-Min, there’s a noticeable softening of both the hard clip white and solid black; both are not quite pure white nor a solid back compared to Red’s RedLogFilm gamma.

As much as I wanted to believe that my Scarlet-W was superior, it’s hard to ignore the performance¬†of the Varicam LT. I believe it can be a top tier contender when compared to the Arri Alexa or the RED Weapon at about half the price. This test didn’t even touch upon the Varicam’s ability to switch it’s native ISO to 5000¬†in-camera¬†giving you incredible low light performance with minimal noise introduction.

Going into the future, it still needs to be tested to see how well Panasonic’s AVC-I codec and V-log will hold up to¬†aggressive color grading and correction. At the time of the test, the Varicam LT’s RAW codex was¬†not yet released. Red’s R3D format has already been proven to hold up against great amounts of grading flexibility in post, and I’m curious to see if the Varicam LT will have that ability as well without too much degradation to the image.

In closing, the Panasonic Varicam LT displays great dynamic range and exceeded expectations in both over and underexposure. To be clear, the dynamic range was the¬†only thing we were truly testing with this series and it will be exciting to see how it holds up¬†in real-world situations. I’m still very happy with the performance of my¬†Scarlet-W and this doesn’t necessarily mean that one camera is better than the other; dynamic range is only one small part of the equation in the overall picture. I do believe that there is a place for both cameras in the market, and it’s also important to not overlook the $10,000 price difference between the two.

 

Special thanks to Toshi Kizu for having me and allowing me to share the results, Takahiro Mitsui from Panasonic for taking an entire Saturday out of his weekend to have us, the crew Ryo Araki and Yusuke Shiratori, and our lovely talent, Shavvon Lin, for being so patient while conducting an inordinate amount of camera settings.

 

Thanks for reading and please feel free to leave me a comment and/or suggestions. I love hearing your feedback!

More content on the way soon. Stay posted.

-Brian

 

Some fun BTS stills taken throughout the day

FULL RESULTS

Overexposure +8¬Ĺ

 

 

Underexposure -8¬Ĺ

 

 

About the Author:

Brian is a freelance cinematographer based out of Los Angeles, CA. He’s currently a Cinematography Fellow at the prestigious AFI Conservatory.



7 Comments:

  1. Gavin Greenwalt
    March 29, 2016
    Reply

    To be clear “Red’s Redlogfilm” isn’t RED’s log curve it’s “Cineon” Log Aka the official international standard log curve. Log shouldn’t have a fancy head or tail or be “as flat as possible” it should be a known mathematical international standard so that you can linearize the data and apply whatever look you want. That’s why the academy’s ACES is linear to start with, it’s a known standard. One of the most frustrating parts of working with Alexa ProRes footage is that in order to get to a linear starting point you have to use a proprietary inverse gamma curve. Almost every application on earth supports Cineon (REDLogFilm) and you can get to the “real” data. If you want to start from ArriLogC then you can go [Log (cineon) -> Lin] -> [Lin -> ArriLogC.] Or even with a LUT go from Log (Cineon) -> Arri LogC. I’m disappointed that Varicam is creating yet-another-useless-log-format that nobody can work with in their software applications for effectively no purpose.

    • Brian
      March 30, 2016
      Reply

      Gavin, thanks for the great comment, that was incredibly insightful. It looks like I need to brush on my log curve knowledge. What would be a more accurate way to describe what Panasonic is doing with their V-log? I guess the bottom line of what I was trying to say was that V-log looked similar to Arri’s Log-C and was very flat.

    • Dean
      March 30, 2016
      Reply

      The guy that created the Cineon formula is nowadays the president of Arri. If they aren’t using it, they surely have a reason…
      And the most likely reason is they don’t want to lose tonal precision in the top end, which the Cineon curve does since it is meant for coding highlights compression that occurs in film capture (but not in digital capture).

  2. Esat
    March 30, 2016
    Reply

    Can you upload stills from raw ? (In raw format)

    Cineon is a good curve for working space, Arri uses log-c because of limited data space. It’s better to linearize all media and than apply log-c or cineon function at the and. You can use Resolve to manage it, just set working space scene linear, output as any log format and debayer media in scene linear ACES or device rgb. Do not use luts for scene linear color management.

  3. Ian M
    March 31, 2016
    Reply

    Great work on the test, very well done. Really like your website as-well.

    The Log/Linear ambiguity does make it seem a bit tough to really determine how much latitude the Scarlet-W has, not that your test is off or anything, but I’d love to see a version of this same test where you grade all the exposures back to middle grey. That way we can more clearly see what starts to be lost when the images is pushed and pulled, and each frame can be brought back into a Rec 709 color space. Supposedly Panasonic engineers unofficially have been able to get 16 stops out of the Varicam 35 which is why it’s designated with the 14+, though that’s probably pushing it to the absolute limit. That is awful close to the Alexa though, very exciting camera.

  4. Derek Doublin
    April 04, 2016
    Reply

    Nice test.

    If Panasonic had only released this same camera with 120fps slow motion with the sensor SCALED instead of cropped, I would sell my Epic Dragon and my old C300 and buy the Varicam immediately. It‚Äôs a great camera, but the cropping of anything over 60fps isn’t practical for my work. Most of the commercial gigs I am on need slow motion and almost all of the music videos I work on use it. Heck, even documentaries use a ton of slow motion in their B-Roll now. Slow motion is EXPECTED on a product like this. Phones are shooting 250 fps in full HD now. I realize a cinema camera is a bit different then a phone, but the point I’m trying to make is that the world is accustomed to seeing slow motion now. It’s a major part of the visual medium. A $20k+ cinema camera needs to be able to be properly overcranked. I‚Äôm not sure why both Canon and Panasonic think this isn‚Äôt an important feature.


Leave a Comment!

Your email address will not be published. Required fields are marked *