Time for a shootout?

[DBounce]

I’m sort of kicking around the idea of doing a shoutout between the 1DXMK2 and the Panasonic LUMIX GH5. It’s been a while since I’ve used the Canon for video... mostly the GH5 has been my weapon of choice. But I was never disappointed with the video from the Canon. However, for grading the 10bit footage of the GH5 seems to be more flexible.

What got me started on the shootout idea was a YouTuber who said he’s returning the Canon and keeping his 5DMk4. One of the reasons he states is ISO performance... something I have never had issue with on the 1DXMK2. I’ve managed with the limited ISO of the GH5, which The 1DXMk2 handily outperforms.

So here’s the question... do the shootout now, or wait till I get me hot little paws on the GH5S? I’ll be using a MB SB to keep things even. 

What do you think. Or is this one already put to bed? And I’m not just speaking about  low light.

Oh, also... if any of you have a C200 and one of these two... or both... I would love to see how they all compare 

 

[no_connection]

Do some different scenarios, like candle lit dinner, day at beach and whatnot, but match DOF and field of view. With and without SB.

Think more of what do I want to do and try doing that and see what setup breaks first and what you need to do to fix it. If it can be done.

[mkabi]

Can you also use EOSHD C-log or something similar and do a comparison with VLog and/or HLG - before and after grading.... side by side would be nice too.

Thanks in advance.

[DBounce]

1 hour ago, mkabi said:

Can you also use EOSHD C-log or something similar and do a comparison with VLog and/or HLG - before and after grading.... side by side would be nice too.

Thanks in advance.

I’ve kinda lost my faith in pseudo log profiles. Real logs are not merely color profiles. There are no real log profiles for the 1DXMK2. And I can tell you from my previous experience and grading 1DXMK2 footage vs 10bit VLog that the Canon footage breaks much sooner than VLog. For that reason I tend to pick a color profile and bake it in on the Canon. But IQ is a closer call between these two. It’s been awhile since I shot video with the Canon, so I’m also focusing on the shooting experience between the two after owning both for some time. The stills performance of the 1DXMK2 is so good it make even the ham-fisted look like a pro. The GH5 makes you work harder to get magic.

I’m pulling out the tripod 

[webrunner5]

Speaking of comparisons, I just happen to have one. Talk about baked in, well this sort of proves the point good or bad. Same color science in the bigger models as these consumer cameras. Canon is Canon no matter the price colors wise. etc.

 

[wildrym]

It would be great if you could compare dynamic range and latitute (overexposed/underexposed skintones brought back to normal exposure)

 

Also IBIS with no IS lens on gh5 vs OIS lens on canon

 

Thanks!

[mkabi]

14 hours ago, DBounce said:

I’ve kinda lost my faith in pseudo log profiles. Real logs are not merely color profiles. There are no real log profiles for the 1DXMK2. And I can tell you from my previous experience and grading 1DXMK2 footage vs 10bit VLog that the Canon footage breaks much sooner than VLog. For that reason I tend to pick a color profile and bake it in on the Canon. But IQ is a closer call between these two. It’s been awhile since I shot video with the Canon, so I’m also focusing on the shooting experience between the two after owning both for some time. The stills performance of the 1DXMK2 is so good it make even the ham-fisted look like a pro. The GH5 makes you work harder to get magic.

I’m pulling out the tripod 

Can you at least do Prolost Flat then?

[DBounce]

So this video represent an exploration into these cameras. I have never really taken the time to do a real side by side, other than when I first got the GH5 and wanted to see the FOV difference compared to the Canon. I have to say, on a 4K monitor, I can clearly see noise at all ISO on both cameras. The GH5 was color graded as it was shot using V-Log. Whilst I choose (as I almost always do) to use a baked-in color profile with the Canon. I had to work prettty hard to get the GH5’s colors close to the Canon... and even then, I can tell you I was not 100% happy. But that said, in the first few shots from the GH5, the color grade is truer to what my eyes could see in that room vs what the 1DXMk2 reproduced.

About that room: This was not a low-light test... It was really more of a “I’m in a normal room, with a bunch of people, and I want to try to get some footage without anything but practical lighting... because I did not bring a lighting kit to lunch”. Why? Because often times I take my camera along when we go out. So capturing footage inpromptu is a frequent situation that I find myself in.

Realization: I always thought the GH5 was clean at < ISO 1600, but after watching this video on a 4K monitor, in my office, I can tell you that is simply not the case. Under these “normal” lighting condition both cameras showed shadow noise at even ISO 500. I don’t know why I have never noticed this before, but I think that the better TVs and monitors become, the more apparent the shortcomings of our cameras will be.

Where am I now? The noise from the Canon and the GH5 was comparable at similar ISOs with the MBSB attached. Granted, the GH5 could gather a lot more light than the Canon with the focal reducer. But if set to similar ISOs the noise was close enough as made no difference. So what does this mean for the GH5S? Looking at these noise levels I am surprised that I somehow missed it prior to this. Perhaps it’s because I was only critical of my images when they were planned and lighted scenes? The room was not really that dark. The shots in the video are pretty representative of what my eyeballs could see. I’m a bit surprised that I could not get perfectly clean footage under fairly normal conditions. 

Did I do anything other than CG? Well, I should point out that the second to last scene (2.26) was using a film look, but honestly the noise is exactly the same. It looked bad, so I took it into Magic Bullet and threw a look on it. It was shot at ISO 2500 on the GH5 using a native Panasonic 12-35mm f2.8 version 2 lens @12mm. Another shot at (1.55) was more about delving into what a I could do creatively with it. That’s the scene shot at 70mm, where Hildy (the Doogans) looks at the camera. I add a film stock and denoiser at 15%. The rest of the footage is untouched, other than basic CG to match the 1DXMk2.

I think that native glass... unlite is not a great idea. And most of the “fast” glass with a few exceptions does not stay sharp unless stopped down a bit.

Anything to add? 

 

32 minutes ago, mkabi said:

Can you at least do Prolost Flat then?

I’ll check if I still have prolost on the camera. But I can tell you, I have tried more flat profile on the Canon than you can shake a stick at. The best they get you is a fairly light grade. If you want to bend and twist your footage, you need at least 10 bit. It is much much more robust. But for ultimate flexibility you need 12 bit raw. From my experience with 8 bit from Canon and Sony, they are more fragile when grading. 

[EthanAlexander]

Thanks for posting!

24 minutes ago, DBounce said:

Granted, the GH5 could gather a lot more light than the Canon with the focal reducer.

Not sure what you're getting at here... With the SBXL the crop factor is 1.28... So it's practically identical to the 1.3 4K from the 1DXII and therefore the light gathering would be exactly the same: With the adaptor you can think of it as putting a 15-45mm 1.8 on a M43 sensor or just think of it as putting the 24-70 2.8 on a sensor that's the same size as the 1DXII (APS-H in 4K). (The only thing you can't do is think of it as putting a 1.8 on an APS-H sensor.)

24 minutes ago, DBounce said:

I’ll check if I still have prolost on the camera.

No need to upload anything, it's just tweaking the camera settings on the Neutral PP:

[DBounce]

32 minutes ago, EthanAlexander said:

Thanks for posting!

Not sure what you're getting at here... With the SBXL the crop factor is 1.28... So it's practically identical to the 1.3 4K from the 1DXII and therefore the light gathering would be exactly the same: With the adaptor you can think of it as putting a 15-45mm 1.8 on a M43 sensor or just think of it as putting the 24-70 2.8 on a sensor that's the same size as the 1DXII (APS-H in 4K). (The only thing you can't do is think of it as putting a 1.8 on an APS-H sensor.)

No need to upload anything, it's just tweaking the camera settings on the Neutral PP:

When you attach a full frame lens via SB the camera automatically increases the available f-stop range. You gain an additional stop of light with the MBSB. There are tons of articles about this. Here is one newsshooter. 

Yes, forgot that about prolost settings... there are so many pseudo logs out there.

[EthanAlexander]

2 minutes ago, DBounce said:

When you attach a full frame lens via SB the camera automatically increases the available f-stop range. You gain an additional stop of light with the MBSB. There are tons of articles about this. Here is one newsshooter. 

I'm sorry but this is still not an accurate way of looking at this comparison. The two cameras you are comparing are receiving the exact same amount of light when the lens is wide open, whether it's read as a 2.8 on the Canon or 1.8 on the Panasonic.

ISO's are normalized between sensor sizes (think Sunny-16 rule) so that exposures can be matched, but the amount of light hitting the sensor on your GH5 using the SBXL is exactly the same as is hitting the 1DXII. Think about it... How can there ever be more light coming through the lens than there is when it's set wide open to 2.8? Using the speedbooster just directs that light at a smaller sensor, but it's the same light that's hitting the bigger sensor. The GH5 is reading it as f/1.8, because it now is (a 15-45mm 1.8), but that doesn't mean that the camera is receiving more light, just that the exposure settings will be different - this due to the normalized ISO's I was talking about.

[DBounce]

21 minutes ago, EthanAlexander said:

I'm sorry but this is still not an accurate way of looking at this comparison. The two cameras you are comparing are receiving the exact same amount of light when the lens is wide open, whether it's read as a 2.8 on the Canon or 1.8 on the Panasonic.

ISO's are normalized between sensor sizes (think Sunny-16 rule) so that exposures can be matched, but the amount of light hitting the sensor on your GH5 using the SBXL is exactly the same as is hitting the 1DXII. Think about it... How can there ever be more light coming through the lens than there is when it's set wide open to 2.8? Using the speedbooster just directs that light at a smaller sensor, but it's the same light that's hitting the bigger sensor. The GH5 is reading it as f/1.8, because it now is (a 15-45mm 1.8), but that doesn't mean that the camera is receiving more light, just that the exposure settings will be different - this due to the normalized ISO's I was talking about.

Do you suppose there is a reason Metabones calls it a “SpeedBooster”?

[EthanAlexander]

14 minutes ago, DBounce said:

Do you suppose there is a reason Metabones calls it a “SpeedBooster”?

Look, I'm just trying to educate other readers because I spent a long time misunderstanding what speedboosters actually do. At one point I actually thought that speedboosting FF lenses to smaller sensors meant I'd get more light and better bokeh than using an actual FF camera 

So, to answer your condescending question: As I've said - You can think of putting a 24-70mm 2.8 with SBXL on the GH5 as EITHER

• a 15-45mm 1.8 on a M43 sensor OR
• a 24-70mm 2.8 on an APSH sensor 

(Or, for shits and giggles, you can think of it as a 30-90mm 3.6 on FF)

But you CAN'T think of it as putting a 1.8 on any sensor larger than M43. (This is the fundamental misunderstanding right here that has led to confusion on countless threads.)

Therefore, the light hitting the two cameras is exactly the same wide open.

The whole reason I was telling you this in the first place is because you can actually do the comparison between the two cameras with the GH5 at lower ISOs and see how the noise performance compares at equivalence (Because of the squaring rule, I'm not sure the math, but there's a calc on the web, I'm sure). This is because of the normalized ISOs I was referring to.

[mkabi]

20 minutes ago, EthanAlexander said:

 Therefore, the light hitting the two cameras is exactly the same wide open.

 

I have to agree on this one...

Put it another way...

By adding a speedbooster it’s not increasing the diameter of the barrel of the lens... so whatever light the barrel of the lens is letting through is what’s coming to the sensor or speedbooster (and the speed booster doesn't add more light unless there is some sort of light built into it)

[DBounce]

1 minute ago, mkabi said:

I have to agree on this one...

Put it another way...

By adding a speedbooster it’s not increasing the diameter of the barrel of the lens... so whatever light the barrel of the lens is letting through is what’s coming to the sensor or speedbooster.

The light is the same, but the intensity (concentration) is different. You do not see any difference in DOF. Think of it this way, normal sunlight will not catch a newspaper on fire, but concentrate it with a magnifying glass and that same normal sunlight will cause the paper to burn. For the camera sensor it is getting more light than it normally would, because the concentration is higher. End result is you can work at lower ISOs, but you gain no DOF advantage. 

This is not theory... I see this whenever I use my SpeedBoosters.

[EthanAlexander]

8 minutes ago, DBounce said:

For the camera sensor it is getting more light than it normally would, because the concentration is higher. End result is you can work at lower ISOs, but you gain no DOF advantage. 

Ok, maybe we've been saying the same thing after all, just using different language... 

I'd be curious to see the two compared with the lower ISOs on the GH5.

I think, based on quick math, that ISO 640 on the GH5 could be compared to 1600 on the 1D with same lens settings. If you're up for it, this would be an interesting comparison.

[mkabi]

28 minutes ago, DBounce said:

The light is the same, but the intensity (concentration) is different. You do not see any difference in DOF. Think of it this way, normal sunlight will not catch a newspaper on fire, but concentrate it with a magnifying glass and that same normal sunlight will cause the paper to burn. For the camera sensor it is getting more light than it normally would, because the concentration is higher. End result is you can work at lower ISOs, but you gain no DOF advantage. 

This is not theory... I see this whenever I use my SpeedBoosters.

Agreed. Like a magnifying glass... I don't know how it will affect the sensor in the long run though - as you said like burning paper.

So no DOF difference? When adding a speedbooster it only gives you additional FOV and ISO, no DOF change?

[DBounce]

12 minutes ago, mkabi said:

Agreed. Like a magnifying glass... I don't know how it will affect the sensor in the long run though - as you said like burning paper.

So no DOF difference? When adding a speedbooster it only gives you additional FOV and ISO, no DOF change?

Correct, DOF stays the same... and honestly that is a lucky coincidence, since AF is unusable on the GH5 for the most part. I would say in a cinema camera too much DOF can work against you. Heck I dial it down frequently in stills because you cannot get a sense of where the image was taken, and oftentimes when telling a story (weather in stills or motion) you want the viewer to understand context.

18 minutes ago, EthanAlexander said:

Ok, maybe we've been saying the same thing after all, just using different language... 

I'd be curious to see the two compared with the lower ISOs on the GH5.

I think, based on quick math, that ISO 640 on the GH5 could be compared to 1600 on the 1D with same lens settings. If you're up for it, this would be an interesting comparison.

Interested, I’ll throw something together and chuck it up on YouTube. 

As I sit here waiting for the GH5S I have high hopes for this camera. So far I like what I see. Though @Neumann Films video seems to be a bit off in the motion cadence department... clean imagery, but the cadence feels like 30p not 24p. We will know soon enough what Panasonic has given us. 

 

[mkabi]

2 hours ago, DBounce said:

Correct, DOF stay the same... and honestly that is a lucky coincidence, since AF is unusable on the GH5 for the most part. I would say in a cinema camera too much DOF can work against you. Heck I dial it down frequently in stills because you cannot get a sense of where the image was taken, and oftentimes when telling a story (weather in stills or motion) you want the viewer to understand where the subject is.

Yeah, I have to agree that too much DOF isn't good for video.... For me... Even when you get it right, it feels like a person is standing in front of a blurred green screen. You either over do it to make it look like a dream sequence or under do it enough for the audience to know the point of focus without distraction.

+I agree on the 30p feel of Neumann Films work.

[Neumann Films]

Which video feels like 30p? Quite a few shots in the official one were at 30 and transformed to 24, so that should have no effect on motion cadence.

The low light test and then our Vlog used 24p and 180 shutter.