The Benefits of Working with the Sony a6300 and Atomos Shogun

by Doug SanfordMarch 24^th, 2016

We found that the new  Sony Alpha a6300 mirrorless camera is delivering excellent 4K video at a very low price. Using an external recorder  we can use easier-to-handle and higher quality codecs. In this guest article cinema5D reader Doug Stanford describes his experience working with the Sony a6300 and Atomos Shogun.

— Intro by Sebastian Wöber

Check out our Sony a6300 Reviews:

• Sony a6300 vs. Sony a7S II Image Quality – How Good is it Really?
• Sony a6300 Low Light Test – a Mini a7S II for Much Less?
• Sony a6300 Review – Real-World Video & First Impressions
• Download & Grade Our Sony a6300 Footage – 4 Free LUTs by James Miller

The Sony a6300 and Atomos Shogun

As a full time freelance commercial/corporate comm/lifestyle DP living just outside of Washington DC, I’m always trying to balance competitive workflow and deliverables with gear that I can afford as I build my equipment arsenal a year and a half in to self-employment.

I’ve been searching for an upgrade to the 5D mark II that would give me a few key features that have been lacking for my clients – 4K acquisition, a robust codec for color work, external monitoring, and high frame rate shooting – but without making the leap to a cinema body just yet. With the announcement of the Sony a6300, it finally felt like the time to jump to an updated and more modular solution had arrived.

I selected the Sony a6300 and Atomos Shogun (Note that Atomos recently introduced the Ninja Flame), with the Metabones Speed Booster to be the foundation of my new rig, as I already owned a collection of Canon primes. My 2016 goal has been to grow my commercial/corporate comm business, but for now I still need to be able to shoot on location at weddings that can last 8-10 hours, so I used that as a baseline for my rig’s stamina. I rarely come back with more than five hours of footage from these, so 4x SanDisk Ultra II 480GB SDDs that can record just shy of 6 hours of 4K ProRes 422 (HQ) collectively seemed like the minimum I could stock.

Similarly, after running some tests it seems like both the Shogun and a6300 can run reliably for an hour and a half per battery (note: Atomos and Sony brand batteries – I have tried 2 varieties of off-brand batteries and found at best they’re about 20-30% quicker to burn out), so 4x Atomos 5200ma and 4x Sony NP-FW50 is the minimum I can bring with me to most likely make it through a full day.

Often times I’ve struggled with low-light during receptions and I’ve grown to love the field of view that a full frame sensor provides, so the Metabones Speed Booster was a no-brainer to add to the setup.

My media management has been challenging to think through – I work off a Glyph 4TB RAID drive that definitely can’t support holding on to too much footage at 4K ProRes 422 (HQ) sizes. My strategy (for now) has been to down-convert on ingest to 1080P ProRes 422 for self-produced projects that don’t demand extensive color treatments or 4K deliverables. I’ll retain the benefit of the supersampled 4k->1080P sharpness and noise but be working in post start-to-finish at the resolution that I’m delivering at 99% of the time.

Shooting with the a6300 and Shogun

I shot this video as my first trial run with the rig outside of the house. I wanted to put it through the challenges of high contrast, bright sunlight shooting and see if it could hold up and provide shadow detail without blowing the clouds – an issue I ran into constantly with the 5D mkII.

Processed with VSCOcam with f1 preset

Some notes on how this was shot: I used PP7 in it’s factory mode: S-log 2, no settings altered. I had a Tiffen 72mm Variable ND filter on the lens to keep my shutter speed at 1/50 and f-stop typically around F/2.8 and ISO 800 (natural ISO/lowest available in video mode).

There are some scenes where you’ll catch some of the unpleasant X-pattern from setting the ND filter to it’s max darkness, I’ve added an additional 0.6 ND B+W filter to my collection since shooting this to avoid having to crank the variable ND as high.

I unfortunately recorded nearly all the footage at 422 (LT) by not remembering to double check my settings before I headed out, but thankfully it didn’t have a lot of impact on the footage, which didn’t receive a very heavy grade in post. I shot in a lot of direct sunlight for about 45 minutes – simultaneously to the a6300 and Shogun, using the a6300 to trigger recording via HDMI – without running into any heat issues thankfully, but it was pretty cool out and I wasn’t shooting clips back-to-back.

I did my 4k 422 (LT)->1080P 422 conversion via Media Encoder CC on import, edited in Premiere, and sent it to Resolve for grading where I used a modified version of Casey Wilson’s REC709ish for A7s S-LOG2. My tweaked version has a lower contrast, a subtle warming in the shadows, and a bit of desaturation in the highlights and shadows, which I was using on the Shogun to preview while shooting. You can download my version HERE.

Throughout the grade I was impressed by how rich and vivid the colors became as soon as I loaded the LUT onto a corrector node. I played with exposure here and there and corrected UV hazing on a handful of shots, but for the most part the footage is pretty un-corrected. The especially impressive parts are the shots of ivy vines and the black bar gate blocking an alleyway, both of these pulled detail out of the shadows that blew me away.

Furthermore, the shots where I aim directly at the sun for flaring came out beautifully – no apparent banding or harsh steps out of blowout; the Speedbooster adding a few extra element flares that I’m fine with. Overall I can’t say enough about how happy I am with the sharpness and clarity of the image. Having had the opportunity to shoot on a C100 (with Atoms Ninja) and RED Scarlet with my lenses on a number of jobs in the last year I can say that the sharpness and noise levels of this camera are impressive compared to these workhorses.

The Trade-offs

There are of course some trade-offs to be made with a camera body that offers these features at such an affordable price point. The primary issue I have with the camera is the very noticeable rolling shutter, which has been written about by a lot of people at this point. Unfortunately the reviews are all being pretty fair, it would be challenging to use this camera entirely handheld and fast motion in the frame is going to suffer.

Initial testing with Premiere’s Rolling Shutter Repair effect is promising, but it feels like a bandaid to a fairly large problem. A secondary issue is overheating, which again, has been documented plenty in this body and other Sony mirrorless options. Using the Shogun seems to delay the issue but on a shoot today I finally had the camera shut off from overheating after about 3.5 hours of continuous use.

So the big question is, how important is it to record to the Shogun vs. internally in XAVC? Given that the Shogun is more than the cost of the camera itself (Alternative: Ninja Flame), it’s not necessarily an easy decision, but here’s what I’ve found from comparing a few shots simultaneously shot both ways:

Sony a6300 Internal vs. External Recording

• The a6300 takes significantly longer to overheat if you don’t record internally at all.
• The screen on the Shogun is much easier to view, though neither do amazing in bright sunlight.
• The Shogun adds a handful of precise and easy to use metering tools that can ensure you’re not blowing highlights, which can be a little unforgiving in S-LOG.
• The Shogun adds XLR in/out – making it finally simple to record and monitor your audio during interviews, and/or increasing your total record channels to 4 when combined with the camera’s stereo over HDMI feed.
• Despite being a nice 100mbit compression, the internal codec has it’s limitations. For scenes with little or subtle motion things hold up pretty well, but when you have a fast camera move, things fall apart in three areas: complex patterns, edges, and subtle gradients. Furthermore, beyond the compression artifacts you can see a visible difference in color information and an interesting contrast shift between the two versions, which I’m attributing mostly to the improved color space of the Shogun.

I’ve pulled a handful of stills from test shots taken around the house that exemplify this:

obvious color/contrast shift

Note the highly degraded detail in the gradients of the motion blur

note the loss of detail in the handle and speaker area

again, obvious color/contrast differences

note the loss of detail in the mic stand base’s center hole

note the loss of fine detail in the wire rack’s supports

color/contrast shift

while you can nitpick the loss of detail in the gradients of the bokeh, this is actually the worst part of the image I could find… the XAVC held up well in this scene, which I’ll chalk up to the lack of motion and the image information being mostly in the mid and highlight tones

color/contrast shift

a real loss of definition and smoothness in the bokeh

introduction of serious banding in the bokeh gradients, in the Shogun footage the sensor’s noise actually does a really nice job dithering these inside of the 8-bit HDMI feed, where the internal recording destroys the natural look of the Shogun footage with hard banding

Final Thoughts

If you’re looking to work in a capacity where you’re delivering footage to paying clients publishing to broadcast or projection, the Atomos Shogun (or Ninja Flame) can make this camera a viable option if you’re conscious of and compensating for the rolling shutter problems while shooting.

If you’re working on personal projects or on your own independent film endeavors you have more choice, and you can weigh the benefits of the added image fidelity against your working budgets to decide what you need. At the end of the day, I’ll be continuing to shoot at 4k ProRes 422(HQ) despite the more time consuming post workflow and physically cumbersome production rig, I’m really excited by the results and look forward to continuing to see what this set up is capable of.