Panasonic GH7

[92F]

https://www.youtube.com/watch?v=kRRJ-Ilo1HM

[kye]

Some stunning footage at the start of this one...

 

[Davide DB]

27 minutes ago, kye said:

Some stunning footage at the start of this one...

 

I liked but what's in there can't be obtained with the GH5?

I see a nice but pretty heavy color grading.

[kye]

Panasonic must be pretty confident about the LogC implementation.  They got this guy to shoot a film with ARRI as A-cam (Mini LF) and GH7 as B-cam, and also specifically told him to try and break it (which he failed at, despite it being kicked into a brick wall and falling onto a concrete floor from shoulder height - twice - as well as taking other knocks).

and the short film itself:

 

[kye]

7 minutes ago, Davide DB said:

I liked but what's in there can't be obtained with the GH5?

I see a nice but pretty heavy color grading.

Not a clue.

Seems odd to criticise a video if it doesn't answer every question you ever had about something! 😂😂😂

But, it's ok, comparisons are coming.

[Davide DB]

On 6/6/2024 at 3:37 PM, IronFilm said:

  

 

It's raw. So of course you can't be using the full sensor with a scale down applied. Nah, raw means a crop of the sensor if you're not recording the full resolution. 

Guys forgive me for insisting on this raw thing. Tell me where I am wrong. I am here to learn.

In M43 I have a crop of about 2x due to the ratio of the FF sensor to the M43 sensor.

When I shoot in Prores or H265, the camera uses the whole horizontal part of the sensor and then scales the image to C4K or UHD resolution.

When I shoot in Raw the camera uses exactly the pixels needed to get the desired resolution otherwise it would not be raw.

I did not understand if in the two options the camera changes the crop factor of the lens.

Thank you in advance

 

[Davide DB]

1 minute ago, kye said:

Not a clue.

Seems odd to criticise a video if it doesn't answer every question you ever had about something! 😂😂😂

No but I don't criticize it.

I think it is impossible to judge cameras by videos. This is a bias that was born on social media and will never die. 

I have never heard about a film that it would have come out differently if they had used one camera or another and yet this BS on social never dies. 

The videos say more about the skill of the author than the camera.

I would like to see Roger Deakins with a GH2.... 😉

 

[sanveer]

20 minutes ago, Davide DB said:

No but I don't criticize it.

I think it is impossible to judge cameras by videos. This is a bias that was born on social media and will never die. 

I have never heard about a film that it would have come out differently if they had used one camera or another and yet this BS on social never dies. 

The videos say more about the skill of the author than the camera.

I would like to see Roger Deakins with a GH2.... 😉

 

If Panasonic paid him $1 Million, he would.

Also, I may be wrong, but its difficult to really judge the image unless one sees it on a much bigger screen, that had great HDR. On a smartphone screen it's pretty useless. On smaller monitors too, it's pretty difficult to judge.

 

[Davide DB]

2 minutes ago, sanveer said:

If Panasonic paid him $1 Million, he would

And for sure he could write anything in the video description...

[sanveer]

6 minutes ago, Davide DB said:

And for sure he could write anything in the video description...

Exactly 

[PannySVHS]

Highlights look awful in this video imho @kye I really like some very subtle hues in the footage and the rendering of textures. Some shots look 100% like Aja Cion, in a good way and in a bad one as well, some look akwardly and unpurposefully dark, some look promising. Mixed bag with interesting things to come hopefully.

[eatstoomuchjam]

1 hour ago, Davide DB said:

Guys forgive me for insisting on this raw thing. Tell me where I am wrong. I am here to learn.

In M43 I have a crop of about 2x due to the ratio of the FF sensor to the M43 sensor.

When I shoot in Prores or H265, the camera uses the whole horizontal part of the sensor and then scales the image to C4K or UHD resolution.

When I shoot in Raw the camera uses exactly the pixels needed to get the desired resolution otherwise it would not be raw.

I did not understand if in the two options the camera changes the crop factor of the lens.

Thank you in advance

 

To be clear, lenses don't have a crop factor.  Lenses project an image circle of a certain size.  When the image circle is smaller than is needed to cover a full camera sensor, you will get dark corners or in some extreme cases, the outline of a circle.
As long as the sensor is smaller than the image circle, though, you will get an image that covers the entire sensor.  If the sensor is 24x36mm (full frame), this would be referred to as a "1x" crop factor in modern terminology.

If you put a smaller sensor with the same lens or only capture a portion of the above sensor, the crop factor will increase.  We usually calculate this by calculating the diagonal lengths of the sensors and dividing them (or the areas of the sensors and dividing them).  So a sensor of around 22.3 x 14.9mm ends up with a diagonal about 1.5x that of 24x36 and Micro 4/3 is about 2x.

If you then choose to use only a portion of the Micro 4/3 sensor, the crop factor would increase again.  So in raw, going to 4k from 6k is usually an additional 1.5x crop and going to 2k from 6k will be closer to a 2x crop.  

[zlfan]

9 hours ago, ita149 said:

I would have apprecied a comparison between Open Gate and 4K with everything higher than a 8mm lens (like he did in his previous video), because saying the stabilization is the same while testing the IBIS only at this focal lenght is not serious.
In my tests, both the GH6 and G9II have worse stabilization in Open Gate with lenses > at 15mm, more jumps.
I understand he likes Panasonic a lot but owning all the recent Panasonic cameras, I can say it's a very oriented comparison to makes people believe the camera has no flaws.

In the same way, all these youtubers almost never compare the video IQ (details rendering) between 4K/6K/Open Gate or between the recent and the "old" gen Panasonic cameras, and yet there is a lot to be said. Each time we need to compare the cameras ourselves or wait the impression of people who bought the camera themselve.

have you tested em1 series and gh6 and g 9 2 in terms of ibis at focal length greater than 100mm, at least greater than 50 mm? I guess Olympus ibis still beats penny by a mile at long focal length, which is critical for manual focus lenses adapted to m43 bodies. 

[Davide DB]

2 hours ago, eatstoomuchjam said:

To be clear, lenses don't have a crop factor.  Lenses project an image circle of a certain size.  When the image circle is smaller than is needed to cover a full camera sensor, you will get dark corners or in some extreme cases, the outline of a circle.
As long as the sensor is smaller than the image circle, though, you will get an image that covers the entire sensor.  If the sensor is 24x36mm (full frame), this would be referred to as a "1x" crop factor in modern terminology.

If you put a smaller sensor with the same lens or only capture a portion of the above sensor, the crop factor will increase.  We usually calculate this by calculating the diagonal lengths of the sensors and dividing them (or the areas of the sensors and dividing them).  So a sensor of around 22.3 x 14.9mm ends up with a diagonal about 1.5x that of 24x36 and Micro 4/3 is about 2x.

If you then choose to use only a portion of the Micro 4/3 sensor, the crop factor would increase again.  So in raw, going to 4k from 6k is usually an additional 1.5x crop and going to 2k from 6k will be closer to a 2x crop.  

 

So my 8-18m @ 8mm become 22.3mm.

Hummm this is a deal breaker for wide angle underwater use...

[eatstoomuchjam]

If you want to shoot raw, yes.  If there's a non-windowed ProRes or H.265 mode, though, they should be about like a 16mm FOV on ff.

[Davide DB]

For narrative shots I use a lot the 12-35mm in a 6" dome port. 

For ambient wide angle shots I use the 7-14mm or 8-18mm on the wide end. Rule #1 underwater is to reduce water/distance.

The most useful combo is the (vintage) Pana 14-42mm with a special wet lens which converts the lens FOV from 75°-29° to 130°-50° and it focus basically on the lens.

With this extra crop everything becomes too narrow.

 

 

[ac6000cw]

3 hours ago, Davide DB said:

So my 8-18m @ 8mm become 22.3mm

If you using an 8mm lens on your GH5, it will be the equivalent of about an 11mm lens on the GH7 in DCI 4k RAW mode. Open Gate 5.7k RAW will have the same FOV as there is no extra crop in that mode (so an 8mm lens will still be 8mm).

[zlfan]

5 hours ago, Davide DB said:

 

So my 8-18m @ 8mm become 22.3mm.

Hummm this is a deal breaker for wide angle underwater use...

if you shoot 5.8k raw, there is no extra crop, 8 mm is 16mm ff. 

[zlfan]

if you use c4k raw mode on gh7, coupling with a Zeiss s16 pl 10-100mm, it is about 28-280mm ff. there should be no vignette on the tele side. on the wide side, very possibly there is no vignette neither. 

[kye]

9 hours ago, PannySVHS said:

Highlights look awful in this video imho @kye I really like some very subtle hues in the footage and the rendering of textures. Some shots look 100% like Aja Cion, in a good way and in a bad one as well, some look akwardly and unpurposefully dark, some look promising. Mixed bag with interesting things to come hopefully.

I guess mileage varies depending on how you look at the footage.

What I see in it is:

• pretty heavy grade done but footage looks consistent and hues look right, including skin-tones
  this indicates that the sensor and profile behave well in the grade and don't get pulled around in unnatural ways - this is a lot less common than you'd imagine and many of the grades I've done end up making different tonal or colour regions of the footage don't look coherent with each other..  
   
• saturated edges look fine
  this is very difficult because doing a strong subtractive saturation operation is very challenging as it stretches tones apart significantly based on their hue.  this is especially difficult on edges where if you have an edge of a coloured object against a neutral object then in the space of a few pixels you go from saturation being low to being high, now this wouldn't normally be a challenge because maybe they're at a similar luma level but after the subtractive saturation operation the saturated pixel is now significantly more saturated but more importantly its also now dramatically darker.  this will ruthlessly reveal compression artefacts on edges, and is the primary reason that I have to go B&W and blur the absolute crap out of any footage I shoot on the cheap action cameras etc for cheap camera challenges
   
• greens look good
  shots like the dog at 0:54 show deep saturated greens in the foliage in the background, which have clearly been significantly altered.  in real life the graduations in greens amongst footage like this are very subtle, and yet the colours in the video look dense and the opposite of stretched low-quality footage, they have a density that is reminiscent of the things that the 5D with ML does very well (but I'm not saying this is at that same level - this video isn't enough to judge that)
  anyone who has declared war on yellow-ish greens like I have and tried to grade them back into looking lush (or even just not dead-looking) will know the whole thing exists on a knife edge.  the fact that you're cooling the shadows and warming the highlights (which this grade does quite considerably) also stretches the greens too
   
• shots look to have usable DR in outside scenes in full-sun on surfaces that would be blindingly-white
   
• there is an effortlessness to the colours
  when you look at film there is an effortless coherence to the whole thing, like it was always going to be this way that it would be silly for it to be any other way - the colours in film look like this was their destiny - you don't question it.  this coherence is incredibly rare for digital and especially smaller sensors when not given enough light, so although the outside sunshine pics are great, it's the inside shots with the deep shadows that show off a coherency into the lower registers of the tonal ranges.  the colours don't cheapen, they don't thin, they don't shift, or become glaring or dull, etc.

etc.

Hollywood colourists can work low-level miracles, but will be the first to tell you that if the footage is lacking then there's a limit to what they can do - more would be asking for high-level miracles!  Either this footage is good and the colourist was a super-star, but considering this is a YT video, I'm leaning on it being a good-to-very-good grade on top of very capable footage.

The more I colour grade my own footage, the more I see when I look into the footage of others.