GH5S Sample Footage and Reviews

[jonpais]

O

[webrunner5]

2 hours ago, Mmmbeats said:

No, I'm quite happy to base my knowledge on industry stalwarts like Adam Wilt, thankyou.  And confident in sharing his contributions with others.  He's clearly a great deal better informed than your good self .

I'd love to know what you mean by 'all else' in your phrase above?  Surely you don't mean codec?  There's plenty of footage from the BMPCC for example recorded on Prores (obv. not RAW) that still displays high dynamic range.  You (and others) might have to revise your thinking a bit. 

You are Not going to get 13 stops on a BMPCC without using Raw! Silly stuff. You will get 10 to 11 without Raw. And every camera made in the last 5 years can do that.

And even in Raw if you don't have your exposure perfect, WB set to 40 IRE, nothing much over 100 IRE, you Aren't getting 13 stops either.

[Mmmbeats]

I agree there's a dip.  I've never done a side-by-side, but the little I shot, and watching others footage, I'd say you lose around a stop.  BMPCC footage in Prores easily bests GH4 / GH5 for DR in my experience.

[ntblowz]

GH5s is used on Season 3 of Frontier TV show for some low light scene. 

 

[Jim Giberti]

7 hours ago, webrunner5 said:

You are Not going to get 13 stops on a BMPCC without using Raw! Silly stuff. You will get 10 to 11 without Raw.

That's simply not true. We shoot all BM cameras and there's nothing like that difference between raw and ProRes - and I've tested the comparison for several shoots.

Well shot ProRes is virtually indistinguishable from raw (to an audience.)

Of course there are times (bright uncontrolled daylight) where the extra mojo of raw is great especially on skin- but it's nothing like the 2 or 3 stops you're suggesting.

It's why most commercial work is done in ProRes - great codec and marginal  difference in IQ - but a huge difference in workflow/deadlines.

The strength of raw is it's malleability in post with Resolve. But properly shot ProRes has it's advantages over raw for much/most pro work.

[Jonesy Jones]

Yah web runner just lost all credibility with me with that statement. 

[graphicnatured]

I don't see that big a difference either in my Pocket.

[mercer]

I’ve noticed a little difference with the BMPCC and BMMCC between Raw and ProRes. I would say 1-1.5 stops less in ProRes. Obviously the major difference is in film mode vs video mode and that could easily be 2-3 stops.

The advantage of Raw is in the color tonality. Living in the woods and shooting in rural areas, I notice it in tree bark the most. Raw will record the textural difference in a more 3D way.

This can easily be described as having more DR but I think it is the minuscule color tonality that Raw is capturing and not necessarily the traditional idea of Luma based DR from shades of whites to grays to blacks. This is also something more noticeable in clouds.

Years ago, there was a kid named Zach who used to post here and he was very happy with his t2i. As he should be, the t2i was a great camera back in the day. Well he decided to post an image in an attempt to decipher how many stops of DR he was getting using Tragic Lantern.

I think the consensus was 6-7 stops. 6-7 stops. This was from a camera that helped start the DSLR revolution... a legend in its own right. And now, only 5 or 6 years later, we’re arguing about what camera has 1 or 2 more stops of DR... hell the fact that any affordable camera can get over 10 stops is amazing.

[anonim]

@mercer

Nice. Just for the game... today at the morning I played with testing some reverse-log curve (in developing process) in my country house - how much of DR I can achieve. First one is original GH5/Natural profile shot, second one using this reverselog curve and graded... How many stops of DR it is? I don't know, but it doesn't look bad to me... these are stills grabed from clip in Resolve. (Of course, outdoor is snowish pure white

f4, ISO200, 180 degree + ND filter, Voigtlander 42.5mm

 

 

[mercer]

@anonim I’m on my phone right now, so it’s kinda hard to tell. I’m also not that big of a pixel peeper. I would assume the GH5 can do 10 or 11 stops, maybe 12 with vLog.

The guy who posted his test before, recorded a DR graph he got online. I don’t think it’s that accurate but it gets you in the ballpark if you’re worried about it.

To add to my earlier comment, I went back and looked at some of my old BMPCC/BMMCC footage this morning and I actually think there is only a half a stop to a stop difference between ProRes and Raw... really hardly noticeable at all.

[anonim]

@mercer

Oh no, I'm not worried at all... actually, I know that this has identical DR as Vlog, so, judging from posted examples, it had to be 12 stops. As I wrote, I'm testing capability of other picture profiles with reversing them in log state with accurate reverse log curve. I posted it just for the sake of the last DR question in topic. Actually, I've found that DR of Vlog curve is achievable with any other - I tried constructing and applying procedure first with Natural.... Looking at posted examples and comparing/judging from real circumstances - result is excellent.

[mercer]

1 hour ago, anonim said:

@mercer

Oh no, I'm not worried at all... actually, I know that this has identical DR as Vlog, so, judging from posted examples, it had to be 12 stops. As I wrote, I'm testing capability of other picture profiles with reversing them in log state with accurate reverse log curve. I posted it just for the sake of the last DR question in topic. Actually, I've found that DR of Vlog curve is achievable with any other - I tried constructing and applying procedure first with Natural.... Looking at posted examples and comparing/judging from real circumstances - result is excellent.

Interesting. Why not just shoot in vLog rather than in a standard profile and use the reverse method? It seems like there would be a lot of other variable... picture profile settings... etc. 

[Lux Shots]

Yeah, this ^! In addition, your luma may be in tack, but the chroma is DOA.

I'm busy prepping for two separate shoots for two different customers and two different locations this Saturday. But I will try to do a proper dynamic range test with the GH5S on Sunday.

[anonim]

@mercer @Lux Shots

It was just an addition to mercer's nice comment about DR capability of modern cameras... otherwise I wouldn't post it here.

Concrete answers: 1) I have no Vlog in GH5 - more precisely, for me it is proven that its DR range is achievable by reproducing reverse software manipulation (i.e. that's the same as work of Balazer and others) - just one reason for me to paid extra cost is because of brilliant monitoring solution with Atomos recorders; 2) Vlog has its specific color reproduction, but it is also just arbitrary matter of taste, in fact there's no unanimous opinions about it as the most preferable solution 3) In posted example I liked to achieve some W. Andersson look - trying not just to retrieve color data, but also to test pushing them further in grading... So there's not at all any conclusion - maybe just that, at least with GH5, level of DR-wise grade-ability and color tweaking possibilities is pretty high (enough for serious usage) and funny once when proper base-log values are achieved by any methodology... (Hat down for results of @Sage)

[webrunner5]

18 hours ago, Jonesy Jones said:

Yah web runner just lost all credibility with me with that statement. 

Both of these articles State that the BMPCC Has 9 Stops of DR. Even less in ProRes. Even if you believe 13 stops, which it isn't, they show ProRes is even less. I rest my case. If you people really believe the bullshit that company's try to pass for fact I have a Bridge I am selling cheap.

I am not knocking a BMPCC, I had one once. They are a special camera. But most of the cameras Really don't have the true DR they state they do. And all the cameras that have super DR, Reds, Arri's are getting it in Raw not some other Codec.

http://www.similaar.com/foto/pocket-dr/bmpcc.html

https://www.focus-numerique.com/camescope/tests/blackmagic-design-bmpc-951.html

[mercer]

50 minutes ago, anonim said:

@mercer @Lux Shots

It was just an addition to mercer's nice comment about DR capability of modern cameras... otherwise I wouldn't post it here.

Concrete answers: 1) I have no Vlog in GH5 - more precisely, for me it is proven that its DR range is achievable by reproducing reverse software manipulation (i.e. that's the same as work of Balazer and others) - just one reason for me to paid extra cost is because of brilliant monitoring solution with Atomos recorders; 2) Vlog has its specific color reproduction, but it is also just arbitrary matter of taste, in fact there's no unanimous opinions about it as the most preferable solution 3) In posted example I liked to achieve some W. Andersson look - trying not just to retrieve color data, but also to test pushing them further in grading... So there's not at all any conclusion - maybe just that, at least with GH5, level of DR-wise grade-ability and color tweaking possibilities is pretty high (enough for serious usage) and funny once when proper base-log values are achieved by any methodology... (Hat down for results of @Sage)

Honestly, your first reasoning... that you don’t have vLog, is the perfect response. I was always a fan of CineLikeD. I thought there was always adequate DR and enough flatness for some grading. You really had to nail your WB or correct for the orange... neither was a horribly difficult task in post. And with 10bit, it should be a breeze or not even an issue.

[AaronChicago]

48 minutes ago, webrunner5 said:

Both of these articles State that the BMPCC Has 9 Stops of DR. Even less in ProRes. Even if you believe 13 stops, which it isn't, they show ProRes is even less. I rest my case. If you people really believe the bullshit that company's try to pass for fact I have a Bridge I am selling cheap.

I am not knocking a BMPCC, I had one once. They are a special camera. But most of the cameras Really don't have the true DR they state they do. And all the cameras that have super DR, Reds, Arri's are getting it in Raw not some other Codec.

http://www.similaar.com/foto/pocket-dr/bmpcc.html

https://www.focus-numerique.com/camescope/tests/blackmagic-design-bmpc-951.html

These tests can be accurate, and honestly more accurate than boasted DR. BUT, you have to look at the perspective relative to other cameras. These are "usable DR". They rate the 5DII at 6 stops. I'm sure they'd rate the GH5 at 8, and the Ursa Mini at 11. Usable DR is a bit subjective.

[webrunner5]

14 minutes ago, AaronChicago said:

These tests can be accurate, and honestly more accurate than boasted DR. BUT, you have to look at the perspective relative to other cameras. These are "usable DR". They rate the 5DII at 6 stops. I'm sure they'd rate the GH5 at 8, and the Ursa Mini at 11. Usable DR is a bit subjective.

Oh I will agree the BMPCC probably has about the most DR of any camera that is affordable. Although the Sony A7s in S Log 2 is calmed to be even more DR than the BMPCC. I will admit I think even with the links that does seem low for the BMPCC. If it really is that low than a Canon T2i must have 5 stops, and I have seen some tests that is not far from true.

 

[graphicnatured]

Last week I shot a vid using both Ursa Mini Pro and BMPCC on a gimbal. Both prores, and the Ursa seemed to have so much more DR it was way more of a difference than I anticipated.

[Mmmbeats]

The original comment that I thought was debatable was that FF has more impact on picture quality than bit depth (8-bit compared to 10-bit); that's what I found contentious.  But I do think it's become a bit of a derail at this stage.