Article just posted by No Film School Red Epic vs Panasonic GH4

[AndreasOberg]

At this point I'm not even surprised when tests are incompetently botched. I just expect it. At least they had the good sense to include the source files so that someone who knows what they're doing can actually compare things.

Even my comparison is bad, the trees are waving in the wind like someone with their hair on fire--these were the least motion blurry frames I could find for each.

Nice picture
One thing we see clearly in the picture is that the very strong compression in the GH4. Of course the 4k uncompressed would look a lot better but requires more equipment.
There is also quite a lot of green chromatic aberration in the image from the lens. To be fair to the GH4 the image quality looks like it could be similar if we had an uncompressed image, it looks like there is an image hidden behind the compression.
Also is it really a good idea to show the 5k vs the 4k. Should we not render it to a 4k image for it to be better to compare, or?

Cheers
/Andreas

 

[Gavin Greenwalt]

Also is it really a good idea to show the 5k vs the 4k. Should we not render it to a 4k image for it to be better to compare, or?

Cheers
/Andreas

Note that they're both 100% aka 1:1. Yeah, if this was a real test for one the FRAMING should have been the same too. Honestly there is nothing really useful to be learned from this test it's so flawed. The "5k" isn't really 5k it's just a 4k crop from the Epic. So I should really change that label. They are both "4k". In fact if you zoomed way in with the 5D so that you only took a 1080p crop of leaves you would get the same thing sharpness. So the Epic actually has more resolution if you supersample it. But the only way to show that would be to intentionally degrade the GH4 footage.

 

[Aaron McLane]

We have been using our GH4 for UAV work and the 4k although highly compressed is not too bad. For all the UAV guys it's going to be a lot less painful crashing GH4's then Epics. I think the 1080 footage is better on the GH3 though not sure why.

 

[Luke Neumann]

Note that they're both 100% aka 1:1. Yeah, if this was a real test for one the FRAMING should have been the same too. Honestly there is nothing really useful to be learned from this test it's so flawed. The "5k" isn't really 5k it's just a 4k crop from the Epic. So I should really change that label. They are both "4k". In fact if you zoomed way in with the 5D so that you only took a 1080p crop of leaves you would get the same thing sharpness. So the Epic actually has more resolution if you supersample it. But the only way to show that would be to intentionally degrade the GH4 footage.

Disagreed. We learned all that we needed to from the test...the GH4 can cut together with Epic footage. We included more samples with different lenses and the difference is still there. If you turn down sharpening on the GH4 all the way and post sharpen the r3d they are pretty much identical from a "sharpness" standpoint alone. The simple and undeniable fact that blew us away is that a $1,700 camera could act as a decent B cam for your Epic and all for a 4K timeline. Good time to be a filmmaker, as they say.

 

[Luke Neumann]

If you watched the video example "a Coast Day" you can see how unpleasing the hyper sharpness can be, it lacks depth and feels brittle instead of much finer details that smooth out. Just pause the video on a shot with crashing waves ... it looks like an over sharpened jpeg. So yeah not enough depth, after 12+ bits of info to work with, only having 8 is torture. Though as always good ideas are king, but if you have an awesome idea it can only get better with a better camera system. One thing about all these 'comparisons' is that they are almost never unbiased. Look at the 400% crop, I mean they're not even the same shot, close up vs med-CU, and their matching contrast and white balance is a joke. I know my scarlet is sharper in 4k than their 5k epic shot. But it's all about the snobbery and not about honest reporting. For its price the GH4 might be a cheap b-cam, especially if you soften some of the sharp edges.

That was Go Pro footage...

 

[Luke Neumann]

You're welcome universe. We can now return to more productive debates like Star Wars vs Star Trek etc.

At this point I'm not even surprised when tests are botched. I just expect it. At least they had the good sense to include the source files so that someone who knows what they're doing can actually compare things. Even my comparison is bad, the trees are waving in the wind like someone with their hair on fire--these were the least motion blurry frames I could find for each, as they say Garbage In Garbage Out.

I don't know why we would expect a very big difference between the two. They both have similar numbers of photosites. They both are bayer pattern. They both have OLPFs. I would be more surprised if there was a large difference between the two than anything. And when you actually look at the parts of the frame in focus it's pretty clear that they're pretty much the same resolution with just the GH4 being slightly softer (more out of focus? motion blurry?) and more heavily compressed and artifacty.

Since the framing is identical in your before and afters you scaled up the GH4 shot to match the framing of the Epic (it wasn't quite shot with the same window out of camera). So this is more like a 350-400% zoom of the GH4 and a 300% zoom of the Epic.

 

[Gavin Greenwalt]

Disagreed. We learned all that we needed to from the test...the GH4 can cut together with Epic footage.

If your stated test objectives were "The GH4 can intercut with RED footage reasonably well." You would be correct it would appear unsurprisingly that two 4k cameras under controlled circumstances will look similar. But that was not the stated goal of your test, the stated goal of the test was "Sharpness comparison" and your conclusions were that the GH4 was substantially sharper than an Epic and I'm sorry but your test was fatally flawed both in construction and execution. To go a step further, not only was your test flawed but you didn't even ultimately read your own flawed test results correctly and incorrectly deemed the GH4 as sharper. I'm glad the GH4 meets your demands, based on your tests it will not be useful to me but to each their own.

You even had me on your side before doing the comparison. When someone asked me to look at the footage I told them straight up "Yeah, the GH4 is probably a little sharper at 4k since it probably has a weaker OLPF." I was honestly surprised to be proven wrong about my biases.

Since the framing is identical in your before and afters you scaled up the GH4 shot to match the framing of the Epic (it wasn't quite shot with the same window out of camera). So this is more like a 350-400% zoom of the GH4 and a 300% zoom of the Epic.

Once again, wrong. Since the 4k GH4 frame just happens to almost perfectly match the FOV of a 4k extraction from the RED frame I didn't have to scale either image to line them up (test it for yourself). If you had matched FOV between the 5k RED and the 4K GH4 then yes I would have needed to scale the GH4... and the RED 5k would have appeared sharper than it did. But you effectively ignored the RED's resolution advantage making it a 4k camera to 4k camera comparison.

As I said in a previous post. If you had used a 100mm lens on the 5D it would also have matched perfectly since a 100% crop of the 1080p footage would also be 1:1 (if centered on the flowering tree).

 

[Luke Neumann]

If your stated test objectives were "The GH4 can intercut with RED footage reasonably well." You would be correct it would appear unsurprisingly that two 4k cameras under controlled circumstances will look similar. But that was not the stated goal of your test, the stated goal of the test was "Sharpness comparison" and your conclusions were that the GH4 was substantially sharper than an Epic and I'm sorry but your test was fatally flawed both in construction and execution. To go a step further, not only was your test flawed but you didn't even ultimately read your own flawed test results correctly and incorrectly deemed the GH4 as sharper. I'm glad the GH4 meets your demands, based on your tests it will not be useful to me but to each their own.

You even had me on your side before doing the comparison. When someone asked me to look at the footage I told them straight up "Yeah, the GH4 is probably a little sharper at 4k since it probably has a weaker OLPF." I was honestly surprised to be proven wrong about my biases.



Once again, wrong. Since the 4k GH4 frame just happens to almost perfectly match the FOV of a 4k extraction from the RED frame I didn't have to scale either image to line them up (test it for yourself). If you had matched FOV between the 5k RED and the 4K GH4 then yes I would have needed to scale the GH4... and the RED 5k would have appeared sharper than it did. But you effectively ignored the RED's resolution advantage making it a 4k camera to 4k camera comparison.

As I said in a previous post. If you had used a 100mm lens on the 5D it would also have matched perfectly since a 100% crop of the 1080p footage would also be 1:1 (if centered on the flowering tree).

Out of the box, with no post work, the GH4 IS sharper. Sitting here going through a ton of footage on a shoot in Hawaii and I will stand by that statement until I see otherwise. So no, the test was not flawed. Did it meet your expectations? Who knows, thats not why we did the test. I get what you mean with the 4k windowing of the Red shot tho...makes sense. Good day.

 

[Gavin Greenwalt]

Out of the box, with no post work, the GH4 IS sharper.

Then you should post that footage instead of the GH4 footage that's softer than Epic, which is all you've done so far. Are you going to argue that your own GH4 footage that I posted above is sharper than the corresponding RED footage? Because I would say that's a hard case to make. And just saying it's sharper but offering plenty of evidence to the contrary doesn't give your position very much credence.

 

[Luke Neumann]

Then you should post that footage instead of the GH4 footage that's softer than Epic, which is all you've done so far. Are you going to argue that your own GH4 footage that I posted above is sharper than the corresponding RED footage? Because I would say that's a hard case to make. And just saying it's sharper but offering plenty of evidence to the contrary doesn't give your position very much credence.

Even if it is debatable the fact that it's even close is the entire point of the video. Sharp ass 4K from a $1,700 camera. End of story. We didn't even offer our opinion on the tests dude. If you doubt, buy your own and do your own tests. I think you'll come to the exact same conclusion. Sharp ass 4K, cheap b cam to your Epic. That's our opinion. Ours holds a little more weight since, you know, we are actually shooting with both cameras at the moment. You?

 

[Gavin Greenwalt]

Even if it is debatable the fact that it's even close is the entire point of the video. Sharp ass 4K from a $1,700 camera. End of story. We didn't even offer our opinion on the tests dude. If you doubt, buy your own and do your own tests. I think you'll come to the exact same conclusion. Sharp ass 4K, cheap b cam to your Epic. That's our opinion. Ours holds a little more weight since, you know, we are actually shooting with both cameras at the moment. You?

You're talking out of both sides of your mouth.

"We aren't making any claims dude!"
"The GH4 IS sharper."

I'm not disagreeing that it's sharp. It is after all 4kish. Just like RED is 4kish. People can make up their minds, but all they have to go off of is your test which is clearly softer in the GH4. "It's about the same" is not the same as "IS sharper". Nobody should care what my opinion is about the GH4. I don't have an opinion about the GH4, I have demonstrated that the RED is sharper in an image where you say it's softer. People should want to know factually what the case is and facts need examples. And now I'm starting to think that you aren't actually interested in methodically studying the results of your test, you just want to stir up trouble for self promotion and headlines--and so be it. And yes I am lumping you in with the No Film School author, because if it was my test and they were that grossly misinterpreting my work I would email them and have the issue a correction. There are plenty of blogs (including REDuser) who enjoy controversy. Me? I like learning things. If you have more footage you would like to compare I would be happy to look at it. I can't though evaluate and compare or learn anything from your subjective testimony, especially after all of the examples you've provided me have been diametrically opposite of what you claim them to represent.

 

[Luke Neumann]

You're talking out of both sides of your mouth.

"We aren't making any claims dude!"
"The GH4 IS sharper."

I'm not disagreeing that it's sharp. It is after all 4kish. Just like RED is 4kish. As to why your opinion holds more weight? People can make up their minds, but all they have to go off of is your test which is clearly softer in the GH4. "It's about the same" is not the same as "IS sharper". And now I'm starting to think that you aren't actually interested in methodically studying the results of your test, you just want to stir up trouble for self promotion and headlines--and so be it. There are plenty of blogs (including REDuser who enjoy controversy) Me? I like learning things. If you have more footage you would like to compare I would be happy to look at it. I can't though evaluate and compare or learn anything from your subjective testimony, especially after all of the examples you've provided me have been diametrically opposite of what they claim to represent.

We didn't offer our opinion IN THE VIDEO. That was implied, I thought. After using it more, yes, I know that the GH4 is sharper out of the camera. You are debating it so clearly it's "debatable". I'm NOT debating it. I'm saying it is sharper but that's just my opinion on it.

Stir up trouble and self promotion? Yeah, not sharing any footage with you, bud. Spend your own money on the GH4 and try it out for yourself. It's the only way to "test" anyways. That's all we did. Guess we committed the cardinal sin of posting said tests online.

Good. Day.

 

[Gavin Greenwalt]

Folks to answer Luke's question, no if you can't accept feedback (especially criticism backed with clear illustration) then no you shouldn't be posting tests online. Testing is hard. Testing is really hard if you can't incorporate feedback and improve your methodology.

For future generations here are some golden rules on how to perform a resolution test:

1) Keep the framing the same. This is like Rule #1.
2) Stop down, but not so far that you get into diffraction softening.
3) Check your focus. Generally it helps to have a flat resolution target (Painting or printed chart).
4) Keep your resolution subject whether that's photographic or a chart in the center of frame.

 

[Luke Neumann]

Good. Day.

Meaning, I'm not coming back on here to read your responses. My side of the argument has been stated. Any further questions should be addressed on your own with your own cameras. Toots.

 

[Luke Neumann]

Folks to answer Luke's question, no if you can't accept feedback (especially criticism backed with clear illustration) then no you shouldn't be posting tests online. Testing is hard. Testing is really hard if you can't incorporate feedback and improve your methodology.

For future generations here are some golden rules on how to perform a resolution test:

1) Keep the framing the same. This is like Rule #1.
2) Stop down, but not so far that you get into diffraction softening.
3) Check your focus. Generally it helps to have a flat resolution target (Painting or printed chart).
4) Keep your resolution subject whether that's photographic or a chart in the center of frame.

5) Don't have a clear bias (6,422 posts and counting) towards one camera in the test. MOST IMPORTANT RULE OF ALL.

Now I'm done PM sent. Let's meet up in Seattle for some testing!!

 

[Phil Holland]

 

[Gavin Greenwalt]

5) Don't have a clear bias (6,422 posts and counting) towards one camera in the test. MOST IMPORTANT RULE OF ALL.

Now I'm done PM sent. Let's meet up in Seattle for some testing!!

Ask around. RED Fanboy and Gavin are not generally phrases that go together. I hold the dubious honor of being one of only a handful of REDusers who Jim Jannard has threatened to sue "into oblivion".

 

[Luke Neumann]

Folks to answer Luke's question, no if you can't accept feedback (especially criticism backed with clear illustration) then no you shouldn't be posting tests online. Testing is hard. Testing is really hard if you can't incorporate feedback and improve your methodology.

For future generations here are some golden rules on how to perform a resolution test:

1) Keep the framing the same. This is like Rule #1.
2) Stop down, but not so far that you get into diffraction softening.
3) Check your focus. Generally it helps to have a flat resolution target (Painting or printed chart).
4) Keep your resolution subject whether that's photographic or a chart in the center of frame.

Come on buddy. Don't go silent on me now. Respond to my email. I will come up to Seattle, bring my GH4 and Epic. YOU run the tests, we film BTS, we post it all online. Sound like a plan?? That way everyone will be 100% sure that the tests are as unbiased as possible. A true RED fanboy gets to man the controls. No better way to solve this IMO.

 

[Luke Neumann]

Ask around. RED Fanboy and Gavin are not generally phrases that go together. I hold the dubious honor of being one of only a handful of REDusers who Jim Jannard has threatened to sue "into oblivion".

OK, well either way. I doubt, given your stance, that you would favor one camera over the other. Let's do this!

 

[Luke Neumann]

Ask around. RED Fanboy and Gavin are not generally phrases that go together. I hold the dubious honor of being one of only a handful of REDusers who Jim Jannard has threatened to sue "into oblivion".

I'm not far from you. Just down in Oregon. Would only take a day.