5D Mark III raw versus Panasonic GH4

[themartist]

I'm confused at the desire for the sharpest image...i think a balance needs to be struck...when watch cinema films if I was watching the level of sharpness most of the gh4 videos that have been posted exhibit I would feel like i was watching a home movie...however the reason ML Raw has been a smash is it takes an almost unusable level of sharpness in the h.264 codec of the 5d3 and adds enough sharpness to match the red that neumann posted...we should look at it as what is optimal vs. "the sharpest" I think a comparison showing human subjects is much more useful than static shots, I never understood that.  We shoot people much more than buildings and trees.  I think the 5d3 raw would be much more appealing to the masses (shooters and viewers) than gh4 image (camcorder sharp).  

 

Nope. I hear these "video" comments when discussing cameras so often and they're really bizarrely off base. Please show me a camcorder that shoots your home videos at that resolution. Theres nothing "home video" about the GH4. The type of camcorder home videos you're thinking of have less resolution, more in focus, blown out highlights. Not remotely similar. The extra resolution is amazing. Its not a bad thing, not a video thing. You can always tone down detail through a whole array of techniques, however you cannot add resolution where there wasn't any. Where you see people using stock lenses on autofocus, without tripods or steadicams and without careful compositions - that is the shaky ugly home video you're seeing. It has nothing to do with the GH4 at all.

 

The tests to show off resolution show what is possible, they're not meant to be film-like. If you have an aerial shot of dolphins in a clear ocean you want that resolution. If you need to greenscreen, you need that resolution. Too much resolution? You simply soften in post. Its that simple, resolution is awesome and the Gh4 is amazing. The reason a focus on people hasn't been used in these tests is because character shots are normally shot wide open, which doesn't get to emphasize the resolution as well as bricks, trees and scenery does. 

 

Is the GH4 better than the 5D with raw? They're really close. Each have pro's and cons according to your workflow and depend perhaps on which lenses you own. Your 5DMK3 is a powerhouse. Seriously, time to put that bad boy to work. Have you tried ML raw yet? I know you mentioned before that you were hesitant to try it, but its worthwhile knowing if you can handle the extra hassle for the extra DR. At the end of the day editing is a huge part as well. The biggest advantage the 5D has over the GH4 is full frame, thats about it... but to me thats a big one. I direct commercials and do documentary work. If I was to be given a gift of a choice between the 5D and the GH4 - It would be a tough decision. I travel a lot, so a compact kit is important to me... but so is full frame, low light and top end stills. Decisions, decisions. I don't need a new camera immediately so I might wait for the A7s to decide. But if I owned a 5D, I wouldn't be pondering whether I made a good purchase or not.

 

You on the other hand, have mentioned that you're a school teacher and you've just bought the 5D in order to start doing documentary projects and are still new to both cameras and film. Honestly, you need to forget this obsession about which camera to get and the whole ML raw vs GH4 thing. Its really misdirected, as you have more than enough firepower to win an oscar. Rather buy some interesting lenses, a steadicam, follow focus, anamorphics... stuff like that to play around with. And then upload work for people to critique. Please be careful not to get too carried away at these comparisons at the cost of going out and filming.

 

[themartist]

I've always been of the opinion that hd, then 4k, then 8k etc have been quantifiable numbers which happen to be marketing friendly and that it didn't really matter.  I'm not getting a gh4 since I want full frame, but this test really does show how really great that extra resolution is.  If you want it lower res, mis focus slightly or whack a gaussian blur of 2pixels over the image and you have good old 1080p.  the issue is that a lot of users won't consider or implement all of the factors that make up a true cinema experience and the added resolution will only serve to multiply the obviousness of the failings and the pointlessness of the users decision to shoot 4k (who will likely be a consumer under the impression they are a professional, rather than a professional moving with the current trend).  It'll be when Andrew does a creative piece with the camera and his cookes or someone like Hugo Goudsward takes the camera and makes it work for him that we'll see the jump where the resolution really takes it up a notch.  

 

just to clarify I completely disregarded the term 4k within a consumer camera, and the term gh4 (due to my dislike of anything smaller than full frame:)- until this test.  8 months ago the 5d3 raw hack turned everything on its head and now there is the little gh4 which doesnt need 10gb/min worth of cards to shoot with it.  superb IMO

 

Totally agree with you Ric. Looks like you and me might be waiting to hear about the A7s, though to be honest I'm ok waiting a good year until investing in my next body. After the GH1, GH2, nex 5 and Nex 7 I'd prefer upgrading to full frame. For now I'm just loving the progress thats happening and the speedbooster does the trick until my next splurge. I'll be following EOSHD daily, as I always did (before I decided to dive in and start posting)

[Stronz]

Thanks Andrew! Loving the GH4 so far, so many awesome features and customization options in the menu. Did a shoot recently with the Nikon D600, Canon 5DmkIII (unhacked) and the GH4. Some people don't see a difference between HD and 4K, well the Nikon and Cameras aren't even shooting true 1080p, they are mushy and ugly compared to the crisp detail on the GH4.

 

The raw hack on the 5DmkIII is nice but the workflow is so slow and cumbersome. It gives beautiful result but is not user friendly at all. The GH4 gets you very close while being very easy to use.

[themartist]

Thanks Andrew! Loving the GH4 so far, so many awesome features and customization options in the menu. Did a shoot recently with the Nikon D600, Canon 5DmkIII (unhacked) and the GH4. Some people don't see a difference between HD and 4K, well the Nikon and Cameras aren't even shooting true 1080p, they are mushy and ugly compared to the crisp detail on the GH4.

 

The raw hack on the 5DmkIII is nice but the workflow is so slow and cumbersome. It gives beautiful result but is not user friendly at all. The GH4 gets you very close while being very easy to use.

 

I have only once worked, editing some Red footage. If the ML raw is more cumbersome and heavy... I'd definitely not be interested... no matter the result.

[PAVP]

I have a Canon 50D and IMO ML RAW isn't so bad to deal with if you prepare for the workflow.  It's not so hard to convert the files to ProRes or DXnHD and work from there in 10bit.  IMO the image is just soft enough to be more flattering and the color is excellent.  I would love the GH4 down the line, but for right now there's no rush.  NONE of my clients would even appreciate the differences.  It all depends on what your needs are.  Basically we have some really good options for making quality images.

[Orangenz]

5 out of 30 students preferred the 5D when shown via a projector. One of those said that preference was because of the colours. For me, every scene on the GH4 looks better (projector or HD screen) with the one exception of the out of focus wall with tree at the very top. The GH4 shots are exactly the type of image I'm after. Will use the kit lens for a bit before thinking about speedbooster.

[Julian]

Nice test again Andrew :) Great work.

 

I know it's not the point of this test, but max 30 fps vs max. 96 fps does deserve a mention imo.

[hmcindie]

It was a good test in my opinion. 

 

Otherwise except the Epic was clearly focused on the tree. It's sharp and detailed, the closeup shot of the house is already out of focus. 

[KarimNassar]

gh4 footage looks good indeed

[Shield3]

No mention of low light? Big win for the 5D.

Yep.  Also keep in mind he graded this in Resolve; I've used Resolve and Adobe After effects with the ACR.  The latter does a much better job IMO with the raw dng files.  Then once you factor in how much better the 5d3 is for shooting sports with the 61 pt AI Servo AF, and also the fact there's no speed booster for EF glass - it's a no brainer if you already have the 5d3 + EF lenses.  If you're starting with nothing, it's not as easy of a choice.  But I like wide and full frame, so make mine Canon for now....

[Shield3]

Also, it's been my experience that the Canon just does a much better job with skin tones.  Of course, one cannot ignore the 96FPS of the Panny either.

If a speedbooster was released today for MFT - EF mount, I'd buy a GH4.  I don't want two "systems"...

[Shield3]

Already tested for those paying attention, it isn't a big win for the 5D actually!

 

 

I watched this, and clearly the 5d3 walked circles around the "quarter size" sensor.  I see a dramatic difference at ISO 6400 - do you not?  Might also be "fair" to not use the speed booster as many people getting a GH4 won't be using one.  Just sayin'.  Yes, you can use fast glass and special adapters to get the micro four thirds sensor to look good at ISO 1600-3200 - but let's see some tests straight out of the camera with each 24-70 equivalent @ F/2.8.  The 5d3 will crush the gh4 in low light, period.  You have 4x the sensor size - how can you expect anything else?  It's physics...

[AndrewM]

Looking at the test footage, both look really good, but... to put it in very unscientific terms, the Canon just seems to have "more colors." It seems like the main colors in the frame can be matched, but then the secondaries are kind of missing/washed out in the GH4. I'm guessing it is just a bit depth thing... Can't wait to see the 10-bit.

[core_38]

To me 5d3's raw format beats all gh4's prized features (4k, price, 96fps, lightweightness)

The look which you can pull out of the raw footage is priceless, imho.

But when you doesn't need raw then of course gh4 is a winner

[hmcindie]

Yep.  Also keep in mind he graded this in Resolve; I've used Resolve and Adobe After effects with the ACR.  The latter does a much better job IMO with the raw dng files.  

 

Only because ACR adds in a ton of sharpening afterwards, unless you dial it down.

[johnnymossville]

You can do something about sharpness in post,  a mushy soft image will always be a mushy soft image. Try sharpening it in post and it gets funky.  GH4 is leaps and bounds ahead of the 5DMIII at this point in every usable way except extreme low light.   Canon's turn to do something interesting for video in DSLR's. 

[jgharding]

14-bit colour is nicer than 8-bit colour

 

raw looks nicer than 4:2:0

 

5D3 highlight rolloff is smoother, there's more dynamic range. It's great in low light.

 

however

 

5D3 raw requires huge fast memory cards that are expensive, lots of fast hard-disk space that's very expensive

 

5D3 raw requires long-winded workflow with various awkward workarounds as it's not Red-style compressed raw

 

add to that the GH4s 4K and GH4 is a much more sensible choice for most shooters out there.

 

If final image quality at all costs is what's needed, the 5D3 raw is beautiful... for student film, for example, or personal project. Place any real-world restrictions on it and it's harder to get away with when all's said and done.

 

I've seen some amazing 5D3 raw films, but I think far more work will actually be done with GH4.

 

It's an indie film-maker's dream!

[HurtinMinorKey]

Wow, this thread is fired up.  I really think the GH4 looks good, maybe surprisingly so, but then so does the GH3. Both cameras can hang with the 5D3R raw if you put them in the right setting and use them in the right way.  I see the GH4 as a nice improvement over the last generation, but it's lack of raw support means it's not for me. 

[MattH]

I watched this, and clearly the 5d3 walked circles around the "quarter size" sensor.  I see a dramatic difference at ISO 6400 - do you not?  Might also be "fair" to not use the speed booster as many people getting a GH4 won't be using one.  Just sayin'.  Yes, you can use fast glass and special adapters to get the micro four thirds sensor to look good at ISO 1600-3200 - but let's see some tests straight out of the camera with each 24-70 equivalent @ F/2.8.  The 5d3 will crush the gh4 in low light, period.  You have 4x the sensor size - how can you expect anything else?  It's physics...

 

Many people wont be using the raw hack on the 5d3 either, in which case dont even bother because it looses hands down, Period! (to use your word).  The 5d3 in raw is utilising it to the utmost of its ability, so the GH4 should be utilised to the utmost of its ability for a fair comparison.

In this example the speed booster does not give the GH4 an advantage because the lens on it is at 1.8 vs the 1.2 on the 5d3.  The speed booster merely equalises them.

 

The 5d has a larger sensor size, but not all of its sensor is read during raw video mode: possibly only 2 of its 22 megapixels, where as the gh4 uses every pixel in 4k crop mode, so the comparison isn't simple.   Its physics that pi is 3 if you measure it wrong.

I'm not saying that the 5d3 isn't better in low light but your arguments are unsound and your tone is pretty douchey.

[jgharding]

Many people wont be using the raw hack on the 5d3 either, in which case dont even bother because it looses hands down, Period! (to use your word).  The 5d3 in raw is utilising it to the utmost of its ability, so the GH4 should be utilised to the utmost of its ability for a fair comparison.

In this example the speed booster does not give the GH4 an advantage because the lens on it is at 1.8 vs the 1.2 on the 5d3.  The speed booster merely equalises them.

 

The 5d has a larger sensor size, but not all of its sensor is read during raw video mode: possibly only 2 of its 22 megapixels, where as the gh4 uses every pixel in 4k crop mode, so the comparison isn't simple.   Its physics that pi is 3 if you measure it wrong.

I'm not saying that the 5d3 isn't better in low light but your arguments are unsound and your tone is pretty douchey.

 

I'd agree,

 

One can't complain about a Speed Booster being used on GH4 if a 3rd-party hack is used to make the 5D3 record raw.

 

This isn't an abstract test of two out-of-the-box cameras, it's a comparison of two customised film-making solutions, each of which uses specialised hardware and software to get the best footage from the camera body in question.

 

The physical size of the 5D3 sensor does make it great in low-light, but since other factors also come into play on most shoots the test provides valid and interesting information that assists with equipment choices, especially if trading low-light performance against other considerations such as resolution and body weight, for example.