GH4 Looking the Gift Horse in the Mouth

blazer003

Well-known member
First of all, I LOVE the GH4. I don't even have it and I already am 99.9% confident that it's going to be the most useful camera I've ever owned. It's nearly everything I want. Going from a GH2 to the GH4, its focus peaking, 4K video, improved AF, headphone jack, wireless control, cine profiles, I'm drooling. It's just a dream come true. Not to mention 4K RAW over HDMI if I need it in the future (assuming recorders come out, which I'm sure they will) and the YAGH if I need it sooner.

But as I was thinking about the 4K and the slight extra crop, and feeling only that slightest lament that my lenses will be a bit tighter, which honestly won't be a big deal (2.0 for 1080p vs. 2.3 for 4K,) I had the wishful thought: "What if they could have just used the whole sensor to record video?"

Still resolution (and resolution for the sensor) in 16x9 is 4,608 x 2,592. 4K Video is 4,096 x 2,160. I know, like my title states, it's looking a gift horse in the mouth to want more. I've (nearly, 1 month away) been given this wealth of video technology, and of course, I want to go further. And I'm well aware that while the difference in those numbers (4,680 vs 4,096) doesn't seem very high that it represents a whole lot of pixels of resolution data. 11.9 million pixels vs 8.8 million. 3 million or 135% more pixels to be exact. But the opportunity to have even more resolution to pan and scan, even when working in 4K (if I find myself doing that) along with not having extra crop at all would just go another step.

Could the processing power to handle that much be there? Maybe, maybe not. We all know that sometimes there is more horsepower than the developers choose to use (gahhem... GH2... gah-hem). Sorry, hairball.

Well, I can wish for the GH5... Wow, I'm just sad. I'm know there's a lot of others there with me. 8 bit vs 10 bit internal. Battery power to the YAGH. I guess it's just human nature to want more. Greedy bastards we are.

I can't wait for my GH4 to get here so I can stop pontificating this stuff and get out there and play with it!
 
There's a good chance that using the whole sensor and interpolating the pixel mismatch would result in more artifacts, moire, aliasing etc than the apparent 1:1 sampling induces. All things considered, I'll take the crop and use my speed booster.
 
Moire and Aliasing are the main reasons for the extra crop, and I think it's a great trade-off to reduce these two problems.
 
Actually I do agree with you. Why does the video have to conform to 1920x1080 and why does it have to be 4096x2160 and why 3840x2160? Is it an actual limitation of AVCHD and mp4 being multiples of 16? Couldn't it use MJPEG or DNG to get around that? Because I don't think anyone really minds, at this level, of having the option of strange dimensions if it means no aliasing/moire and our own choice to crop however we want in post. But yeah all good :D
 
Why does the video have to conform to 1920x1080 and why does it have to be 4096x2160 and why 3840x2160? Is it an actual limitation of AVCHD and mp4 being multiples of 16? Couldn't it use MJPEG or DNG to get around that? Because I don't think anyone really minds, at this level, of having the option of strange dimensions if it means no aliasing/moire and our own choice to crop however we want in post.

They are current video standards, that's why.
 
Actually I do agree with you. Why does the video have to conform to 1920x1080 and why does it have to be 4096x2160 and why 3840x2160? Is it an actual limitation of AVCHD and mp4 being multiples of 16? Couldn't it use MJPEG or DNG to get around that? Because I don't think anyone really minds, at this level, of having the option of strange dimensions if it means no aliasing/moire and our own choice to crop however we want in post. But yeah all good :D

Because 3840x2160 are direct ratios of 1920x1080 and results in easier scaling
 
But as I was thinking about the 4K and the slight extra crop, and feeling only that slightest lament that my lenses will be a bit tighter, which honestly won't be a big deal (2.0 for 1080p vs. 2.3 for 4K,) I had the wishful thought: "What if they could have just used the whole sensor to record video?"
Before you dive deeper into the rabbit hole, you might want to learn a bit more about what you'll actually find down there. If Panasonic used "the whole sensor", the GH4 would be recording RAW video, rather than an HDMI-compatible format. At that level, you don't have 4,608 x 2,592 full-color pixels, you have a Bayer Filter array of sensels in red, green, and blue monochromatic shades:

http://en.wikipedia.org/wiki/Bayer_filter

Compared to the challenge of upgrading to a 4K H.264 workflow, RAW video post-production is significantly more demanding. If that's what you're looking to tackle, you'll likely want to pass up the GH4 in favor of the Black Magic Cinema Camera.
 
First of all, I LOVE the GH4. I don't even have it and I already am 99.9% confident that it's going to be the most useful camera I've ever owned. It's nearly everything I want. Going from a GH2 to the GH4, its focus peaking, 4K video, improved AF, headphone jack, wireless control, cine profiles, I'm drooling. It's just a dream come true. Not to mention 4K RAW over HDMI if I need it in the future (assuming recorders come out, which I'm sure they will) and the YAGH if I need it sooner.

But as I was thinking about the 4K and the slight extra crop, and feeling only that slightest lament that my lenses will be a bit tighter, which honestly won't be a big deal (2.0 for 1080p vs. 2.3 for 4K,) I had the wishful thought: "What if they could have just used the whole sensor to record video?"

Still resolution (and resolution for the sensor) in 16x9 is 4,608 x 2,592. 4K Video is 4,096 x 2,160. I know, like my title states, it's looking a gift horse in the mouth to want more. I've (nearly, 1 month away) been given this wealth of video technology, and of course, I want to go further. And I'm well aware that while the difference in those numbers (4,680 vs 4,096) doesn't seem very high that it represents a whole lot of pixels of resolution data. 11.9 million pixels vs 8.8 million. 3 million or 135% more pixels to be exact. But the opportunity to have even more resolution to pan and scan, even when working in 4K (if I find myself doing that) along with not having extra crop at all would just go another step.

Could the processing power to handle that much be there? Maybe, maybe not. We all know that sometimes there is more horsepower than the developers choose to use (gahhem... GH2... gah-hem). Sorry, hairball.

Well, I can wish for the GH5... Wow, I'm just sad. I'm know there's a lot of others there with me. 8 bit vs 10 bit internal. Battery power to the YAGH. I guess it's just human nature to want more. Greedy bastards we are.

I can't wait for my GH4 to get here so I can stop pontificating this stuff and get out there and play with it!

No 4K RAW from the GH4, only 10bit 422.
 
They are current video standards, that's why.

Well... yes. Just curious rather as to any technical limitation that may stop them, or if they're just not finding enabling it because people either dismiss it as too strange and don't suggest it. I find having choices of 4:3, 3:2, 1:1, as we do for still photos, quite liberating. It'd be wonderful in cinematic applications, for example using anamorphic lenses or going for a classic look without a crop etc.
 
I agree completely. Personally I'd love a 4.6K mode. There is plenty of precedent for it - Arri allows full-gate mode on the Alexa, RED allows odd aspect ratios, etc.

Andrew Reid made that exact point to Panasonic folks directly (thanks Andrew), mentioned half-way down in this article:
http://www.eoshd.com/content/11934/panasonic-gh4-preview

Heck... I'd also love a mode that uses more of the height of the sensor so that you can use anamorphic lenses to their full too.

On the other hand... I totally understand why Panasonic didn't just give an option to record the entire sensor. They maybe are mass-producing those 4K encoding chips - and they couldn't customize them for each camera. Easiest way to get 100% bug-tested encoding is to make those chips only accept the current 4K standards and test the heck out of them at those specific formats. If that's what they needed to hit the price point, I'm cool with it.

That said... a hack would be awesome :)

Bruce Allen
www.boacinema.com
 
There's a good chance that using the whole sensor and interpolating the pixel mismatch would result in more artifacts, moire, aliasing etc than the apparent 1:1 sampling induces. All things considered, I'll take the crop and use my speed booster.

I wasn't clear in my original post, but I was talking about recording internally (not out of the HDMI, nececssarily, as I realize that's a standardized format and connection) at 4.6K and using every pixel direct, just as it is doing for 4K, but using every pixel on the sensor and getting that extra resolution and eliminating the extra crop. Just a random bemusing thought I had.

That said, for anyone that says it's a "glass half full" mentality, you're very right, of course. But I'd like to point out it's that same "glass half full," "wish it could be better," mentality that gave us hacks for the GH1 and GH2.
 
Back
Top