Blackmagic Forum

Just thought id share this video test between the gh5 and and pocket 4k

Couple other tests on his page as well.

"B" is for BMC no?

Blaine Suque wrote:Just thought id share this video test between the gh5 and and pocket 4k

Couple other tests on his page as well.

Thanks for the test... interesting results, is it right you recorded Prores LT (100mbps) vs All I (400mbps)?


Sent from my SM-N910F using Tapatalk

That's a great test to see the difference between the 2 cameras. Showing you that not even comparing ProRes LT to H264 All-I 400mbps can make up for the crippled info coming from the GH5S sensor ...and that not even writing it to some codec to upload to Vimeo can mask that difference in quality.

Bunk Timmer wrote:That's a great test to see the difference between the 2 cameras.

You're comparing adjustments to one baked-in h.264 image with another baked-image h.264 image -- impossible to draw any conclusions, which is the case with the test itself. The footage looks like log, or barely altered log. What's underneath, accessible to grading, nobody knows.

And compare these to similar lighting condition sample from Pocket v1 camera:

"Similar lighting conditions"? But one shot is log and one isn't? Or one is graded and one isn't?

At least Saresh was comparing the same shots, meaningless though his results are.

Now we're comparing completely different images and completely different workflows?

The OP was complaining on his website that the BMPCC 4K has lost the cinema look. He knows this because he compared a graded clip of the BMPCC with a film look applied, with log footage out of the BMPCC 4K.

Is this really the way to discover the differences between two cameras? How about a little rigor in the testing?

It is difficult to tell as the shots were different exposure and different framing that would be critical to really compare. The gamma is clearly different but it is possible on the GH5 and GH5S to change this in the Highlights and Shadow control. ( not available in Like 709, HLG or VLog-L for obvious reasons ) If the GH5S was shot in V Log L it has to be graded so could be made to look the same as the BMPCC4K as far as gamma was concerned. It was either not graded or not done to match gamma. BMPCC4K looks to be a great camera for the price though but not a competitor for my GH5's

Bunk Timmer wrote:That's a great test to see the difference between the 2 cameras. Showing you that not even comparing ProRes LT to H264 All-I 400mbps can make up for the crippled info coming from the GH5S sensor ...and that not even writing it to some codec to upload to Vimeo can mask that difference in quality.

I do not agree. The two cameras were not exposed the same way. That is up to the operator. The GH5S was shot in V Log-L it was not graded to bring out the mids correctly to match the BMPCC4K and it could have been done. Or the BMPCC4K graded to bring down the mids to the same level as the GH5S. I then expect your comparison may well have been very similar.The GH5 and GH5S have a Highlight and Shadows control that can manage the image in camera. If shot in V LoG-L then that is done in post.

John Paines wrote:You're comparing adjustments to one baked-in h.264 image with another baked-image h.264 image -- impossible to draw any conclusions, which is the case with the test itself. The footage looks like log, or barely altered log. What's underneath, accessible to grading, nobody knows.

I'm not sure what you are saying. For instance I named one of the two pictures 'BMPCC4K and the other GH5S. If it would be impossible to draw any conclusions I just took a gamble and possibly named the wrong footage BMPCC4K and GH5S ...Yet I'm pretty sure I know what is what and the only reason I'm sure is because I recognice the footprint from both the Prores files in the h.264 footage and the processed footage from the GH5 series in the h.624 footage. Once the original files are shared I don't expect to see a huge difference in result after applying the same curves.

SkierEvans wrote:I do not agree. The two cameras were not exposed the same way. That is up to the operator. The GH5S was shot in V Log-L it was not graded to bring out the mids correctly to match the BMPCC4K and it could have been done.

I thought the idea was to show that both cameras have the same dynamic range, but apparently Christian Plähn didn't succeed in your eyes. How about this, ask him to do it once more but this time correctly, then you do the grading and next I will point out what clip was shot on what camera.

No he did not succeed. They look different because they were not graded to the same degree. As you mentioned the original files graded could look the same. Not surprising if they use the same sensor. It was this comment of yours that I disagree with "the crippled info coming from the GH5S sensor"

Alas, I do not think the GH5S and Pocket 4K share the same sensor, similar, yes, both from Sony, but not the same. The two base ISOs are different, the overall size is different. Inferring the GH5S and BPCC4K have the same sensor, is like saying the AF100 and GH2/3 had the same sensor, they did not.
Cheers

SkierEvans wrote:It was this comment of yours that I disagree with "the crippled info coming from the GH5S sensor"

I'm sorry. I will use the term 'not optimal processed signal' from now on.
I can only guess what the reason is for this signal, heat? protecting their higher end camera's? something else? I do know that it is possible to produce a more optimal processed signal for consumer cameras. Look at Fujifilms XT3. The footage will quantize the same way the Panasonic footage will quantize ...but then without the dancing pixel blocks all over your footage.
The BM cameras are simply in an other legue and Arri ...well Arri footage doesn't even notice you were trying to break it.

Denny Smith wrote:Alas, I do not think the GH5S and Pocket 4K sharemthe same sensor, similar, yes, both from Sony, but not the same. The two base ISOs are different, the overall size is different. Inferring the GH5S and BPCC4K have the same sensor, is like saying the AF100 and GH2/3 had the same sensor, they did not.
Cheers

So many people also assume that the sensor IS the camera... which isn't at all accurate. Just look at how different the results were between the AJA Cion, the IOIndustries Flare4K, and the Apertus cameras. Those actually are confirmed to have the same sensor, yet produce significantly different looks... much like Kodak negative film produces a different look from what Fuji neg film and what Agfa neg film produce.

If you read about the development process being the AXIOM Beta, you start to get a sense for how much goes into translating the data from the sensor's data bus into image data, even raw.

But most people don't pay any attention to that... and instead assume that it's just a matter of stuffing a sensor in a box.

Bunk Timmer wrote:The footage will quantize the same way the Panasonic footage will quantize ...but then without the dancing pixel blocks all over your footage.

Where do you see the pixel issue. I have not seen that on any of my videos from my GH5 or GH5S.

SkierEvans wrote:Where do you see the pixel issue. I have not seen that on any of my videos from my GH5 or GH5S.

In that case the GH5/S is the perfect camera for you.

https://m.youtube.com/watch?v=RWa6AK31R ... fullscreen skip over to 3.21 it's a really poor upload but hopefully it will show you what I'm talking about. I first mentioned it on the ungraded footage Mr Reid of EOSHD shared on the internet after the 400Mbps upgrade of the GH5 and got banned for pointing it out. It's the shot with the two horses standing and the blue one having dancing pixel clouds all over it's flank. It stood out to me on a mild grade of the original footage. The Eva 1 has the same problem when you crank the speed up to 240 fps. The sky of the skate boarding kid turns into some kind of blitzkrieg. You might think you don't notice stuff like that in a normal grade but to me it's the reason the GHxx series will always look digital. Once you have seen it it you can't unsee it ...Anyhow that's what I mean with the 'footprint'. Since the 'B' clips have the same footprint to me it's the footage coming from GH5S. I would be really surpriced if it turned out to be the other way around.

Here is one more test. For my opinion those tests are strange. He expose clips to similar midtones range with totally clipped sky highlights, so all cameras look more-less the same. But even so Pocket v1 have more dynamic range, you can see some details in extreme reflections and also red lights at the background have some details and looks smoother.
Pocket v1 have visible Moire



And noise test:

Bunk Timmer wrote:I would be really surprised if it turned out to be the other way around.

So much for my arrogance. Pretty embarrassing. 'A' turns out be the GH5S and 'B' the BMPCC4K cam. Yet I stick with my 'footprint' theory. Just doesn't seem to work too well when you're getting sloppy, arrogant and work with re-coded footage just as John Paines told me. I crawl back under my stone for while.

Dmitry Shijan wrote:But even so Pocket v1 have more dynamic range, you can see some details in extreme reflections and also red lights at the background have some details and looks smoother.

I don't see how you keep on concluding that the Pocket V1 has more dynamic range when the tests I've seen haven't shown that.

I still want to know his thinking why he used the codecs he used to make a unscientific test and draw conclusions from them.







Sent from my SM-N910F using Tapatalk

Perhaps it was due to the beating he took on the P4K vs Alexa thread, but now Wolfcrow is firing back with 50 Reasons to Pick the GH5 over the P4K. Whatever.

Of course for $$$ you can buy his guide to the P4K....

Chris Leutger wrote:Perhaps it was due to the beating he took on the P4K vs Alexa thread, but now Wolfcrow is firing back with 50 Reasons to Pick the GH5 over the P4K. Whatever.

Of course for $$$ you can buy his guide to the P4K....

It's a clickbait title and I wondered the same thing at first, but the substance of the video is actually solid. If you watch it, you'll see that he starts off by explaining that the image quality out of Pocket 4K is superior. Then he lists a bunch of menu features unique to the GH5. While I prefer the simplicity of the BMD camera UI and digital cinema philosophy, other folks may prefer the video centric feature set and more complicated menu setup of the GH5.

Jamie LeJeune wrote:It's a clickbait title and I wondered the same thing at first, but the substance of the video is actually solid.

That's actually a bit of a surprise after the last thread... hopefully BMD will be more discerning about who it sends review models to in the future. I wish I'd been one, but at least my colleagues at RSN got a chance to review one.

Jamie LeJeune wrote:

Chris Leutger wrote:Perhaps it was due to the beating he took on the P4K vs Alexa thread, but now Wolfcrow is firing back with 50 Reasons to Pick the GH5 over the P4K. Whatever.

Of course for $$$ you can buy his guide to the P4K....

It's a clickbait title and I wondered the same thing at first, but the substance of the video is actually solid. If you watch it, you'll see that he starts off by explaining that the image quality out of Pocket 4K is superior. Then he lists a bunch of menu features unique to the GH5. While I prefer the simplicity of the BMD camera UI and digital cinema philosophy, other folks may prefer the video centric feature set and more complicated menu setup of the GH5.

Yeah, some ticks on his list are ok, but there are again several errors about the PCC4k and you just can't trust this guy as he does not care about constructive criticism and does no follow up addressing those errors as any other professional reviewer would do.

If it’s on YouTube, chances are good it’s click bait. Once YouTube began monetizing clicks, it’s about making a living moving from one hot product to another. And most cinematography equipment reviewers to their credit likely need to have other sources of income from doing creative work in the field, but for many young people they hope to make a career of this sort of thing and will never use the equipment for anything other than reviewing it. Let’s move along to people that have a higher purpose...

Fortunately I didn't waste any time reading the nonsense.
Too busy setting up and shooting with the camera.

It's a genuinely break through tool - pound for pound best camera I've used and we've used a lot.

Robert Niessner wrote:Yeah, some ticks on his list are ok, but there are again several errors about the PCC4k and you just can't trust this guy as he does not care about constructive criticism and does no follow up addressing those errors as any other professional reviewer would do.

Then I'm glad I didn't waste any time watching it

I'll probably have one of these in my toolkit sooner or later (probably later, for reasons that aren't BMD related), and I'm pretty certain that I'll be satisfied with the image quality.

Jim Giberti wrote:Fortunately I didn't waste any time reading the nonsense.
Too busy setting up and shooting with the camera.

It's a genuinely break through tool - pound for pound best camera I've used and we've used a lot.

You like the camera then. Which is your favourate part of the tool?

Sent from my SM-N910F using Tapatalk

I have to fess up that I didn't pay any attention to it after watching a couple other of his videos about the new Pocket. I feel like he's somewhere between genuinely trying to figure things out and building his brand. But watching him do a couple reviews, put out his guide and then putting out the GH5 video seemed...exploitative of the situation. I agree with Robert's assessment. And yeah, "clicks" just...ugh.