Panasonic GH5: The Mysterious Photo 6K and Sensor!

(Part 3)

The 6K Photo Mode is somehow mysterious for a lot of reasons but it gives us a lot of hints as for what the sensor can possibly be.

Well, in Panasonic’s product page it says:

“6K PHOTO” is a high speed burst shooting function that cuts a still image out of a 4:3 or 3:2 video footage with 18-megapixel
(Approx.6000 x 3000 effective pixel count) that 6K image manages.
The name ‘6K PHOTO’ is tentative.

So this is a Photo burst of 30 fps that takes 6000 x 3000 stills from 4:3 or 3:2 video footage. Yes, this is quite confusing, what does that even mean? Why 2:1 aspect ratio?

Maybe 18MP is about the limit in which the camera can work without relying on buffer memory, which makes sense. Panasonic said that the GH5 has an advanced Photo 4K that can shoot 4K burst at 60 fps, if each 4K is about 8~9MP, then 4K 60 fps — double the frame rate — would consume about the same processing power and data transfer bandwidth as for a 18MP at 30 fps.

The trick gives us the hint that the GH5 is probably doing full sensor readout, so possibly it means that 4K comes without the extra crop

Otherwise how else the GH5 would grab 18MP stills from video if not from full sensor readout? And since it’s a full sensor readout, why would Panasonic offer it for Photo 6K and not take advantage of it and providing 4K without crop?

I think that the reason Panasonic hasn’t done 4K without cropping is not because they were not capable of that but because they didn’t want to compromise with rolling shutter like it happens with some other cameras that output 4K from a full sensor readout.

So if they decide to adopt it with the GH5, it probably means that this sensor is pretty fast and both Photo 6K and 4K 60fps kind of attest to that.

If it’s taken from 4:3 or 3:2 images, it means that the sensor should have more than 18MP. If it’s a 4:3 sensor that has about 6000 pixels horizontally, it means it’s a 27MP sensor, if it’s a 3:2 sensor, it will have 24MP.

But I doubt Panasonic will put a 24MP~27MP sensor inside the GH5. As most rumors were correct, I think we can speculate a bit based on the rumors that were not revealed in the development announcement at Photokina 2016, like the fact that the GH5 would have a multi-aspect sensor.

Multi-Aspect Sensor

To grab 18MP stills of a 6000 x 3000 resolution and assuming it has the same 1.86x crop as in the GH1 and GH2, this would be a 24MP sensor but with 22MP Micro4/3 stills and that’s still more MP than I would expect from Panasonic, I would think it would be closer to 20MP instead — just like the rumors.

This is when the “Approx.6000 x 3000 effective pixel count” comes in.

So let’s do some mental exercise. Let’s start with the already known 20MP sensor from the Panasonic GX8 and let’s make a multi-aspect sensor from it.

GX8 20MP has 5184 x 3888 pixels, the multi-aspect sensor would have 5663 x 3888 pixels for a total of 22MP. If you take the the video aspect ratio of it you get about 5663 x 3185 pixels for 18MP.

So this might not be 6000 x 3000 pixels but this gives us exactly 18MP.

Curious, huh?

Differently from when there were the GH1 and GH2, today DCI 4K is another aspect-ratio option and it’s different from the usual 16:9 for video. The 1.86x crop and 18.9mm x 13.0mm sensor used in the GH1 and GH2 works for Full HD and even Ultra HD, but it wouldn’t be ideal for DCI 4K. And this gets trickier since I can’t know for sure how Panasonic would address this in a multi-aspect sensor.

One option is to keep it just like the GH1 and GH2 and the DCI 4K would have an extra crop compared to UHD, but I don’t think Panasonic would do that because that would deny the whole purpose of offering Cinema 4K.

A similar option would be to set the DCI 4K as the largest width, matching the full sensor width and then UHD would be a crop from that, keeping the same height. But that would also result in more crop than the original 1.86x.

But the most interesting option would be to make the GH5’s sensor actually larger than it was with the GH1 and GH2 to match the width of the DCI 4K aspect-ratio within the same image circle. So instead of something close to 18.9mm x 13.0mm sensor of the GH1 and GH2 it would more like 19.3mm x 13.0mm, so just a tiny bit wider.

On a side note, DCI 4K aspect-ratio is pretty close to the 2:1 aspect ratio of the 6000 x 3000 said in GH5’s page. And with this new width, the sensor would have about 5755 x 3888 pixels for a total of 22.4MP while the DCI 4K would come from a sensor readout of the 5755 x 3035 pixels area.

So, wouldn’t you say that both scenarios kind of fit the “Approx.6000 x 3000 effective pixel count”?

On the other hand, if you were to actually match this same multi-aspect sensor to 6000 pixels in width you would get basically 6000 x 4000 pixels!

Yeah, this multi-aspect sensor made to fit DCI 4K aspect ratio inside the same image circle is basically a 3:2 sensor.

So another possibility is that Panasonic uses a 3:2 aspect ratio 24MP sensor as the GH5’s multi-aspect sensor but with 22MP for the 4:3 area of the sensor. As I said, I’m a bit more skeptical with Panasonic going beyond 20MP for the 4:3 area, so for sake of argument I will use the 22.4MP multi-aspect sensor as my main example for the rest of this article.

To output the 10-bit 4:2:2 4K the GH5 would have to make this full sensor readout of 18MP and downsample it to 4096 x 2160, right?

But the GH5 could simply turn this same 18MP into 30 fps stills burst instead of downsampling and video encoding it, which would be exactly what the marketing department is calling “Photo 6K”.

The key word is “approximately” and of course, all of this counting that the GH5’s sensor is indeed a multi-aspect one like the some rumors were talking about, something that is yet to be confirmed.

It also should be noted that a 16:9 6K is 5760 pixels wide, which is basically what I got from this expanded 20MP in multi aspect, so this would be a 6K version of Ultra HD, which is what Photo 4K is.

You know what’s funny, the Olympus E-M1 MK II!

Again!

Olympus E-M1 MK II, the camera next door

The Olympus E-M1 MK II is capable of shooting 60 fps burst of RAW images from the sensor’s full readout — this basically means that this is not the same sensor as the GX8 and Pen-F, because those sensors were only capable of half of that frame rate readout.

This almost means 5K 60 fps RAW video burst and I’m surprised nobody is talking about that — like people played with the idea of 4K 60 fps RAW from the Nikon V1.

This is a very powerful Sensor and Engine combo. I don’t know for how long the E-M1 MK II can sustain that but Olympus claim to have 2x more RAW buffer capacity and 3x faster internal data writing compared to the original E-M1. It should also be noted that the Pen-F that also have a 20MP — but probably not the same one — and it already has a bigger buffer than the original E-M1.

So I imagine that you can get 1 or 2 seconds of burst, which can be turned into high quality 5K for 2.5~5 seconds at 24 fps.

More importantly, this is very similar to Panasonic’s Photo 6K, the biggest difference is just their approach.

Olympus chose full sensor readout for RAW images at 60 fps burst, while Panasonic chose image grabs at 30 fps for continuous shooting.

It should also be noted that this sensor should be capable of incredible slow motion! I can’t say if 4K 60 fps will also be without crop, but either way, both cameras should be able of 240 fps in 1080p if they want — even with some extra crop.

All these similarities make me wonder if Panasonic and Olympus are not sharing the same sensor or sensor technology.

Are Panasonic GH5 and Olympus E-M1 MK II based on the same sensor technology?

The GH4 and E-M1 were based on the same sensor, it wouldn’t be odd if the E-M1 MK II and the GH5 were also based on the same sensor, which could explain why the E-M1 MK II has such improved 4K specs. I don’t know, in one way or another, Panasonic and Olympus always end up sharing sensors.

The old 12MP Micro4/3 sensor by Panasonic, the 16MP in the GH3 and E-M5 by Sony, the 16MP in the E-M1 and GH4 by Panasonic, the 20MP of the GX8 and Pen-F by Sony. Sharing is the best way to reduce costs, even more in a flagship sensor that will be mainly used just for the flagship camera for a long time before shared with other line ups.

It would be nice if Panasonic could finally improve AF by adopting on sensor PDAF (Phase Detection AutoFocus) to work together with CDAF (Contrast Detection AutoFocus) just like everybody else already does — Olympus, Sony and Fujifilm, they all have on sensor PDAF. The improved AF and fast speed readout seems to be one of the things Olympus really wanted to highlight in their presentation.

As I said before, everybody has stepped up their AF tracking game and if Olympus’ claims are true, the only one that hasn’t still is Panasonic, so it’s about time.

Olympus might have little to no interest in using multi-aspect sensor since it wouldn’t be worth for them to work extra for something that is important mainly for video applications. But they wouldn’t refuse a partnership development if that could actually reduce the costs, right?

So let’s speculate a bit more.

Sensor Partnership Development possibilities:

There are three options that would help reduce the costs of the sensors.

1. If it’s the basically the same sensor, which is a high possibility if the GH5 doesn’t have a multi-aspect sensor.
2. If GH5 has a multi-aspect sensor and the Olympus E-M1 MK II’s sensor will be based on the same sensor but only activate the 4:3 portion of it or if both use the full sensor
3. They are not the same, the Panasonic GH5 will use a multi-aspect one and the Olympus E-M1 MK II will use a normal Micro4/3 sensor but they are based on the same technology and they were developed together

It was said in rumors that the E-M1 sensor was designed by Olympus and it would be produced by Sony. I think there is a possibility that Panasonic and Olympus could had developed the technology together and either Sony will produce both sensors or Sony will produce Olympus’ sensor while Panasonic will produce their own sensor. But this really depends on what will make the sensors different or similar to each other.

In previous articles I speculated not only about both of them using the same sensor technology but also Panasonic and Olympus sharing some other technology. I understand that in some ways it seems like a conflict of interest since they are competing with each other, but maybe this is one of those “common enemy” alliance, in which they help each other just so Micro4/3 can be more competitive with other systems — with bigger sensors.

There are some older rumors talking about Olympus difficulties in bringing 4K to their cameras. We already know that Olympus has this 5-axis IBIS since the original E-M5 and they made it work with video soon after with a firmware update and that’s something that Panasonic was able to achieve with the GX80.

Not the exactly same period, but basically for as long as Olympus had IBIS and video working together and Panasonic didn’t, Panasonic had 4K in their cameras and Olympus didn’t.

Maybe a year ago or so when they were both in earlier stages of development, Olympus helped Panasonic with the basics of the IBIS tech, so Panasonic had the time to experiment and implement IBIS in the GX80 and now in the G80. At the same time, Panasonic could had helped Olympus with video encoding and also providing the codec technology for the E-M1 MK II.

It’s a bit of a stretch but maybe Panasonic also helped Olympus with their OIS and Dual IS technology, which turned out to be what Olympus now call “Sync IS” and that helped Olympus claim the 6.5 stops of stabilization. I wonder if Olympus would also share their PDAF technology with Panasonic, either way, as I’ve mentioned before, Panasonic really need to start working with PDAF or at least provide an equivalent solution for tracking AF.

Oh, and I’ve seen more tests of continuous tracking AF from the X-T2 and it seems just as good or even better than Dual Pixel AF from Canon and that’s a HUGE accomplishment. That’s an edge Canon had over all mirrorless system and that might be changing. Sure, much more tests need to be done before we are going there but it does look quite promising.

Olympus is known for Photography while Panasonic for Videography, so as long as they can still keep their edge in their respective segment, working together is not so troublesome and that’s why the E-M1 MK II does 60 fps RAW and the Panasonic will do Photo 6K, that would explain the difference in their approach regarding the sensor’s full readout capabilities.

E-M1 MK II Improved Video Capabilities

As I speculated in the first part of this Photokina 2016: Panasonic GH5 series, I just found it interesting that all of sudden Olympus comes up with a 240mbps 4K codec — and this bit rate seems to be DCI 4K only and UHD is 100mbps according to the product’s page.

Something that so far neither Panasonic or Sony have offered yet, which is a bit weird since Olympus was by far the last place in video department —even behind Fuji that has stepped up their game bringing 4K with the X-T2.

Even more interesting though was the addition of All-I 200mbps 1080p option for the E-M1 MK II as that seems the same codec that Panasonic uses in the GH4 — and now with the FZ2000. Improving 1080p with All-I was certainly not a priority so why would Olympus have it? Sony might be able to, but even they don’t offer that option, right now it’s just Panasonic. So Panasonic might really be helping Olympus with video after all.

Olympus hasn’t disclosed any info about the 240mbps codec besides the fact that it’s an IPB encoding, but I imagine that it won’t be 10-bit 4:2:2, otherwise they would had said so. Which makes sense, if Panasonic and Olympus were to share technology, Panasonic wouldn’t hand their trump card like that — same way Olympus still has an edge when it comes to IBIS, the E-M1 MK II being at least one stop better than the new Panasonic G80.

In the Part 2 I speculated about possible codecs the Panasonic GH5 could use but if we follow this new path, in which Panasonic and Olympus have some sort of deal, this 240mbps 4K codec could be a hint of what may come to the GH5.

But I wonder if this is 8-bit 4:2:0 or a 8-bit 4:2:2 encoding, sure, 10-bit is still above that but not by much, would Panasonic go as far as helping Olympus with any flavor of 4:2:2 encoding be it 10-bit or 8-bit? I don’t think they would, but it all depends on what Olympus could had possibly offered back or even demanded in exchange for their tech, so 8-bit 4:2:2 is still a possibility.

If Panasonic and Olympus are really working together, we may see this same 240mbps in Panasonic’s GH5, I’m not saying this will be the main codec, but it might be there just as an option. Couldn’t Panasonic offer this as their mid codec and then offer higher bit rate for those wanting more?

It’s an interesting way of upgrading and still diversifying, so you still offer a better codec within something U3 cards can take and then offer more for those willing invest in more expensive media, that way you don’t narrow down too much your potential consumers.

I’ve seen many people saying they want ProRes HQ and DNxHR but I don’t think that most people really thought throughout. I don’t think people realize how much their workflow would change or even something far more troublesome, how much more they would have to spend to be able to shoot the same amount of footage. More on this in the final article.

Part 1: Overheating Gate and IBIS rant!

Part 2: Codecs and AF improvements Speculation

Part 4: Slow Motion, ProRes/HNxHR and Final Thoughts

Extra: The Sensor has to be Multi-Aspect!