July 3, 2014

How Does the GH4 Stack Up Against the BMCC For Green Screen Work? Prepare to Be Surprised!

Common knowledge tells us that codec and compression are two of the most important factors (after lighting, of course) when it comes to green screen work and being able to pull a quality key from your source footage. Anybody who has ever had the opportunity to shoot green screens both with DSLRs and higher quality cameras with more robust codecs can attest to the fact that it's significantly easier to pull a key from the latter. However, seeing as the GH4 keeps blowing past people's expectations, it was only a matter of time before we started to see green screen comparisons to cameras with better internal codecs. The fine folks at Video Alchemy performed one such test recently, and ran the GH4 up against the BMCC's ProRes and CinemaDNG. The results are surprising, to say the least.

Before we get to the comparison video, we should talk about what the expectations would be for a test of this nature. Taking the GH4's 8-bit 4:2:0 4K output and running it up against the 1080p ProRes 4:2:2 and the 2.5K CinemaDNG from the Blackmagic Cinema Camera seems like a bit of a mismatch, especially considering that the BMCC is an incredibly sharp camera in its own right, but with far superior codecs. Based on this alone, it would be safe to assume that the BMCC would outperform the GH4, even despite the fact that it's not a 4K camera.

With that in mind, here's the test from Video Alchemy.

Right off the bat, it's clear that the GH4 holds an advantage in terms of overall image detail, which should come as no surprise considering that it almost holds up to the EPIC (just in terms of detail). However, what's really surprising is that the GH4 looks to pull a slightly better key than either of the BMCC examples, despite the fact that the 8-bit codec should be struggling around the edges of that afro. It's hard to say that the GH4 wins this battle outright because the key is slightly softer around the edges than either of the BMCC examples. However, you can't argue with the fact that the GH4 example looks absolutely fantastic.

Of course, the BMCC was no slouch in this competition either, and both the ProRes 422 and CinemaDNG pulled fairly solid keys with little work. Ultimately, both of these cameras are going to be viable options if you intend to shoot against a green screen with any frequency. Considering they're both in the same price range, the decision between the GH4 and BMCC should come down to whether or not you need 4K output or RAW versatility, and not the performance of the internal codecs.

What do you guys think of this test? Are you surprised that the GH4 performs so well despite its inferior 8-bit codec?

[via Video AlchemyWolfCrow]

Your Comment

51 Comments

Does anyone know how much dynamic range the gh4 has in its top 1080p mode vs it's down res 4k to 1080p?

July 3, 2014 at 9:26PM, Edited September 4, 11:56AM

4
Reply
Question Guy

There was an article about the GH4, with a Panasonic rep talking to folks from Zacuto, on NFS not too long ago. I think he said 12 stops?

July 3, 2014 at 10:37PM, Edited September 4, 11:56AM

0
Reply

unless GH4 will start to shoot in raw format, to get 12bit won't be possible

July 4, 2014 at 12:15AM, Edited September 4, 11:56AM

0
Reply
Alexey

I think you're confusing stops of light with color channel bit depth. A camera's ability to record 12 bits of color information per channel has virtually nothing to do with its ability to record a certain number of stops of exposure latitude.

July 4, 2014 at 3:01PM, Edited September 4, 11:56AM

4
Reply
Mr Blah

Matt, the Panasonic rep, said 11, though some testers thought it was a bit of a reach.
.
On the other hand, the YAGH attachment does allow a 10-bit recording (to AJA Ki Pro Quad, now sold for $3,500) but I haven't been able to find any of those videos yet. Once Atomos Shogun comes out in the fall, one would be able to record in 10-bit w/out the YAGH. Convergent Design, according to statements made by Mitch Gross at Broadcast Asia, also should be capable of taking the 10-bit HDMI signal out of GH4.

July 4, 2014 at 12:51AM, Edited September 4, 11:56AM

3
Reply
DLD

The dynamic range in 1080 should not be that different than in 4k , the 4k advantage would come from detail of the sharp resolution.

I still did not see a dynamic range video test but Dxo did a test based on the stills side and it was slightly better than the mark 3 from canon , so I would not be surprised if DR is around 12 - 12.5 ,

The Gh4 is winning me over day by day , paired with the shogun this thing will be a beast , I even believe that shogun may be able to do 10 bit slow motion as well as 4k

July 4, 2014 at 12:34AM, Edited September 4, 11:56AM

0
Reply
cjay

Not sure how this site got their info but they say 12.8 (EV)
http://snapsort.com/compare/Panasonic-Lumix-DMC-GH3-vs-Panasonic-Lumix-D...

July 4, 2014 at 11:59AM, Edited September 4, 11:56AM

2
Reply
John

DR doesn't depend on resolution, and in the GH4+cinegamma seems to be slightly higher than that of the 5D3+H264 (not sure how much), slightly lower than that of the C100+clog/widedr (about half a stop), vastly lower than that of the a7s+slog2 (at least 3 stops)

July 4, 2014 at 5:17AM, Edited September 4, 11:56AM

0
Reply

A7s?

July 4, 2014 at 7:55AM, Edited September 4, 11:56AM

0
Reply
Geno

Did you do side-by-side comparisons? So the A7S has more dynamic range than the ARRI Alexa?

July 4, 2014 at 9:27AM, Edited September 4, 11:56AM

3
Reply

DxO Mark have results for both cameras that can be compared side-by-side, though they only compare Raw photographs and not the video images. From those results, the a7S has 13.16 stops of latitude at ISO 80 and the GH4 has 12.8 stops at ISO 88. They both drop off from there, with the a7S maintaining its latitude more consistently across the board. For example, with both cameras set to ISO 400 the a7S has maintains 12.4 stops while the GH4 11.75.
I'd be interested to see what difference S-Log 2 makes, since using it locks the camera at ISO 3200 where, according to DxO Mark, the camera only has 10.5 stops of latitude.

So, no, neither camera comes close to the Alexa.

July 4, 2014 at 3:47PM, Edited September 4, 11:56AM

0
Reply
Mr Blah

@Mr blah

Dxo has dynamic range tests based on sensor for stills , but dynamic range in video is a whole different animal and information does not always equate to the same in video.

However with that being said the 8bit 4k seems to be on par or slightly better than the 5d3 raw which is about 11 to 11.5 in raw

The gh4 with 10 bit external may be 11.7 or 12 or hopefully better

One thing for sure is that the GH4 looks promising .

I do have a serious question for anyone who has experience with raw . does 10 bit vs 12 bit make a difference in terms if information in the blacks and shadows?

I guess what I'm really asking is if 10 bit just as good as 12 bit raw in terms of pushing footage and recovering in post . it just seeems that raw gives you that piece of mind and security , I know raw is not full proof but just wondering if 10 bit as sufficient as 12-14 bit raw ?

July 4, 2014 at 5:20PM, Edited September 4, 11:56AM

0
Reply
cjay

The ProRes version looked by far the best. The GH4 is clearly second to it. Of course it could all be because he didn't apply th same settings.

July 3, 2014 at 9:48PM, Edited September 4, 11:56AM

0
Reply
Cosmin Gurau

Especially since he didn't even try to sharpen the blackmagic camera, full well knowing that the images are unmodified, unlike the GH4, which is heavily sharpened in camera.

July 10, 2014 at 12:49AM, Edited September 4, 11:56AM

3
Reply
Jules

So the winner is..... a Blackmagic 4k!

July 3, 2014 at 10:04PM, Edited September 4, 11:56AM

0
Reply
Adam

Yea not sure why he used the BMCC rather then making an honest comparison of codecs with a 4k vs 4k by using the BM4K. BM4K keys much better then either of these in the video.

July 4, 2014 at 9:50AM, Edited September 4, 11:56AM

0
Reply
Derek

I find this comparison rather tilted. On a project such as this, you are dealing with pixels and how they respond to the keying program. More pixels will equal more definition. To compare something that is 4K to something that is HD is not a fair comparison at all. As great as everyone thinks the GH4 is, in six months they will be touting the next new thing. I find it just a bit boring.

July 3, 2014 at 10:06PM, Edited September 4, 11:56AM

2
Reply
Ronn Hague

I haven't seen the video yet, but as you conclude that the GH4 pulls a better key, that might be down to use of shutter speeds on the GH4, I often struggle with film cadence/motion blur that comes off the beautiful and stunning image produced from the bmpcc, for the money something special

July 3, 2014 at 10:06PM, Edited September 4, 11:56AM

3
Reply
kuban

I don't understand the point of that comparison? Apples to Oranges much? I mean Robert, you are really surprised that 4 times the resolution keys cleaner? really?

Why not test the GH4 4K versus the BM4K. Same resolution, now you're looking at the difference the codec makes. (if any)

6K dragon footage, rendered out to 8bit is going to still key much cleaner than a 12bit RAW BMCC. Resolution is important for keying. Period.

All other things being equal, codec will matter.

July 3, 2014 at 10:08PM, Edited September 4, 11:56AM

4
Reply
sean

that's true. but at the end of the days, no matters who has more K... it's about good/bad cinematography

July 4, 2014 at 12:19AM, Edited September 4, 11:56AM

9
Reply
Alexey

As he mentioned he was comparing models due to the similarity in price.

July 5, 2014 at 7:58PM, Edited September 4, 11:56AM

0
Reply
Dan

i understand the comparison in the sense that they're both similarly priced cameras, but it seems like this comparison really should have been GH4K to BMPC4K, then you're going to see the difference the codec makes. he spent a lot of time highlighting the detail in the 4K image, but that stands to reason.

then once the results are in for the 4K v 4K, people can decide if it's worth spending the extra $2K on the BMPC4K, or if they'd rather stick with the GH4.

July 3, 2014 at 10:19PM, Edited September 4, 11:56AM

0
Reply
Ben Howling

The kind of person interested in a comparison like this doesn't have the extra $2K for the BMPC4K. I think it's worthwhile for people on a budget to see whether it's better to go for more-pixels-weak-codec or fewer-pixels-strong-codec, because that's the tradeoff at this price range. It seems clear that for keying, the benefit of working with 4K footage can't be overstated.

July 3, 2014 at 10:35PM, Edited September 4, 11:56AM

3
Reply

The bmcc 4k is only $3k

July 4, 2014 at 2:07AM, Edited September 4, 11:56AM

5
Reply
Theo Slawin

BM needed to lower that price from $4k when the GH4 came out. They were afraid, eh?

July 4, 2014 at 2:17AM, Edited September 4, 11:56AM

0
Reply
Geno

Afraid or competitive?

July 4, 2014 at 9:24AM, Edited September 4, 11:56AM

0
Reply

It was 4K for $4k. That was a big splash. Did you know about that? The GH4 changed all that. It's not just BM that had the GH4 strike the fear of god in them.

It's hard for some to view the camera they love in a less than favorable light. Yes, BM was afraid. Competitive would be a euphemism.

p.s. as a preemptive, "fear of god" is a saying, not a juxtaposition of the GH4 with God. Phew, headed off some commenters ;-)

July 4, 2014 at 11:18AM, Edited September 4, 11:56AM

8
Reply
Geno

Awesome. Outwardly projecting anti-BM fanboism, while demonstrating extreme GH4 fanboism all in one masterpiece self-foot-shooting comment. Way to self-discredit in the most entertaining way possible! Thanks for the laughs!

July 13, 2014 at 5:08PM, Edited September 4, 11:56AM

0
Reply
Jules

@geno bmd lowered the price of the original camera by $1000 dollars after about the same amount of time, it's just their strategy, nothing more.

July 4, 2014 at 1:12PM, Edited September 4, 11:56AM

5
Reply
Tim

Yeah, right.

July 4, 2014 at 2:11PM, Edited September 4, 11:56AM

0
Reply
Geno

I am right, thanks.

July 7, 2014 at 2:04PM, Edited September 4, 11:56AM

0
Reply
Tim

I wasn't aware the BMCC 4K had gotten that cheap, but that's still a ~90% premium to the GH4. There's definitely a class of filmmaker who shoots enough greenscreen footage to justify owning rather than renting, but doesn't have more than $2k to spend and doesn't have clients to pass the cost onto. If you're looking to rent the best camera you can afford for a 3-day shoot, there'd be no reason to consider any of these cameras.

July 4, 2014 at 4:08PM, Edited September 4, 11:56AM

0
Reply

This test does seem a bit skewed to me, comparing a 4k to a 2.5. If the GH4 is only "slightly" better than the BMCC for keying then I'd still go for the Blackmagic for it's color depth and flexibility to push and pull a less compressed format in post. Especially when comping shots.

It would be more interesting and useful to see the BMCC 4K compared to the GH4.

July 3, 2014 at 10:49PM, Edited September 4, 11:56AM

3
Reply
Nick

This test is a joke hahaha.....

July 3, 2014 at 11:47PM, Edited September 4, 11:56AM

0
Reply
Aram

+1

July 4, 2014 at 12:19AM, Edited September 4, 11:56AM

4
Reply
Alexey

Comparisons like this are useful, but honestly, the bad green screen work that I see in corporate and nonprofit videos (in the relatively low-budget world I often work in) has nothing to do with the camera. It's not the codec or the compression. It's not even (usually) the lighting. It's sloppy use of software. You can pull amazing keys from just about any semi-decently-shot footage if you have Keylight...but it's hard to use. So I'm not trying to pour cold water on the conversation here, but it seems to me that in practical terms, keying well is often more about the person doing the keying than about the camera used.

July 4, 2014 at 12:50AM, Edited September 4, 11:56AM

0
Reply
Ben

Yes, he's not really modelling the best technique here, is he? Apart from anything else, the phrase "chroma key" belongs in the era of analogue video.

July 4, 2014 at 4:48AM, Edited September 4, 11:56AM

1
Reply
Graham Kay

Oh come on, stop this pixel to pixel comparisons at last.

July 4, 2014 at 1:12AM, Edited September 4, 11:56AM

0
Reply
einar

Talk about amateur keying work.Where is your core matte? nobody in the industry keys this way. When I work on a shot I separate the the edges I work on to the middle, that way you can pull the edge without the core disappearing. However I use Nuke not After Effects.

July 4, 2014 at 3:44AM, Edited September 4, 11:56AM

7
Reply
Shaun Fontaine

Blackmagic Production camera in it's current form is unfit for purpose due to the fixed pattern noise issue, we sent ours back and got a Cinema camera instead.

Blackmagic really need to find a pair and admit they fucked up big time but obviously that make them liable for compensation.

July 4, 2014 at 5:19AM, Edited September 4, 11:56AM

6
Reply
Filthy Punt

I thought they fixed that issue with an update, or was that just the pocket cam.

July 4, 2014 at 7:40AM, Edited September 4, 11:56AM

0
Reply
Shaun Fontaine

What a lame comparison. The whole comparison turned to - "look at the shine on the hair" .. ridiculous

July 4, 2014 at 11:43AM, Edited September 4, 11:56AM

0
Reply
Pietro

GH4 reviews come fast and furiously. Here's the brand new one from Philip Bloom.

http://vimeo.com/99538435

July 4, 2014 at 11:48AM, Edited September 4, 11:56AM

0
Reply
DLD

Can someone please confirm BMCC have resolved all the QA issues mentioned often by users?

July 4, 2014 at 3:00PM, Edited September 4, 11:56AM

3
Reply
Altus Firh

What issues? It's all in your head.
.
Sincerely, Grant Petty.

July 4, 2014 at 11:22PM, Edited September 4, 11:56AM

0
Reply
DLD

July 7, 2014 at 2:39PM, Edited September 4, 11:56AM

5
Reply
Alton Firth

you should try on the BMCC 4K i think that was an unfare test, i dont know why everyone it's giving the GH4 all the wins in all test, it's not so great......... i always see unfare test with the GH4 against lower cameras... if it was put against a BMCC 4K the BMCC would have won the match for shure

July 5, 2014 at 2:54PM, Edited September 4, 11:56AM

6
Reply
mateo sedano

It was a test with these 2 models because of price, it's not unfair.

July 5, 2014 at 8:12PM, Edited September 4, 11:56AM

0
Reply
Dan

New After Effects upgrade with key cleaner and advanced spill suppressor has vastly improved quality of keying and made it easier to use. I agree with person who said proper use of software is more important than camera.

July 10, 2014 at 4:51PM, Edited September 4, 11:56AM

0
Reply
Jeremy

Man, this GH thing again. it is not a 10 bit. 4:2:2 200mb/s 4kp25 or 4kp50 10 bit 4:2:0, or 300-400mb/s. It is still a toy. The only problem with the GH is that Red never went head to head with them with more professional specs, shame on them all. At least they could now do a 4k dragon Scarlet video photography all in one camera with full auto push button and selectable calibration.

July 11, 2014 at 10:30AM, Edited September 4, 11:56AM

6
Reply
Wayne Morellini

Pretty useful "toy" though to have rattling around in your kit bag...

July 16, 2014 at 11:59AM, Edited September 4, 11:56AM

0
Reply
Will Lodge