Panasonic GH5: Slow Motion and ProRes/DNxHR possibilities

(Part 4)

two [EPIC] bits
9 min readOct 5, 2016

120 fps and 240 fps Slow Motion Possibilities

Sony already have slow motion with a pretty good quality, Panasonic needs to step up their slow motion game. 10-bit 4:2:2, 4K 60 fps, IBIS and 1080p 240 fps? These are killers specs, just make it happen.

Sony RX10 at 240 fps Slow Motion

There are two ways to achieve a great looking high speed slow motion.

Panasonic will need to do something similar to what Fuji did for their normal 1080p. Instead of a full sensor readout of all pixels, just skip a line so instead of reading 5700 x 3200, it would read 2850 x 1600 pixels and downsample it to 1920 x 1080 pixels.

This would still allow for plenty of details since it’s downsampling instead of the just usual line skipping and this is probably how Sony does 120 fps with the A6300. Plus this is 4x less resolution to read, which means that it could possibly read 4x faster the usual speed.

Assuming this sensor can read at 60fps, it means that this downsampling technique could possibly allow for 240fps at this 4x lower resolution readout. And if the Panasonic GH5’s sensor can’t do full sensor readout at 60 fps like the Olympus E-M1 MK II but only half of it, you can still achieve 240 fps by cropping to 1:1 pixel to exactly 1920 x 1080.

I think that 1080p 120 fps is quite guaranteed but hopefully Panasonic will push a bit and achieve 240 fps as well.

What would ProRes and HNxHR mean for consumers?

Most photo cameras doing 4K do so with a 8-bit 4:2:0 H.264 based codec at 100mbps (12.5MB/s). A ProRes HQ and DNxHR at their highest setting for 4K 30fps is up to 9x faster bit rate than that, so it’s very close to 900mbps.

The issue, as I’ve said before is that there is no SD Card formally guaranteeing such Minimum Sequential Write Speed, which is the measurement for video — as opposed to Maximum Sequential Write Speed and Bus Speed, which is more Photo related. But there are cards that seem capable of recording it, as the BlackMagic Video Assist 4K already records both ProRes and DNxHR.

To make it easier let’s see how much you would have to spend in cards to get 10 hours of 4K footage.

8-bit 4:2:2 H.264 IPB codec at 100mbps

1 hour is about 44GB, so 10 hours would be 440GB, that’s 7 cards of 64GB UHS-I U3, with each card costing about $30, you end up spending $210.

10-bit 4:2:2 ProRes HQ or DNxHR All-I codec at 900mbps

10 hours of footage requires almost 4TB of storage. A 64GB card can’t even record 10min of footage, which is quite limiting, so you certainly choose more storage. The most recommended card is the Lexar 2000x UHS-II and possibly the new Sandisk UHS-II, both rated at 300MB/s for Maximum Reading Speed. I think the maximum storage available is 128GB and it sells for about $200 — so basically all you would need to spend with a 100mbps but just in a single card.

So you would need 31 of these UHS-II 128GB high-speed cards and you would spend about $6200 just in media.

It means that compared to what you spend now, you will spend 30x more in cards and not just that, but compared to a codec compatible with the U3 standard of 240mbps, it means that you wouldn’t be able to keep the cards you already have.

And of course, with that same money you could buy at least 3x GH5.

Sure, you would change your shooting style and find more optimized ways to shoot and save up media, use logs and transferring footage to a laptop and external storage but not only that’s much more expensive, it’s much more troublesome and time consuming.

Anyone that already had to keep transferring files during production knows what I’m talking about, not only this is more prone to errors but if you don’t have the time to stop what you are doing to empty your media, it means that you will need somebody else to do it for you.

XQD Cards

XQD Cards could be another possibility. But in the mockup showed at Photokina 2016 the card’s door says SDXC, not that it means that much since it’s just a mockup but there were also rumors saying it was SD Cards, not XQD.

Right now, XQD Cards are cheaper than these new high-speed UHS-II but that can be just because there are less options for the high-speed UHS-II.

External Drive Back Up Solution

There are External Drives that backs up from SD Cards directly to it and they are reasonably affordable but even so with the limit of these high-speed UHS-II cards being 128GB, there is a lot of catch up to do if you want to record longer than 20 minutes.

Assuming the GH5 is a Dual Card Slot and both slots are UHS-II compatible, you would need at least 4 cards to make it work, plus the External Drive and that would total for at least $1000.

Even in that situation you would need to find a workaround if you want to record longer than the SD Cards can handle without stopping filming. And an ideal set would be with 3 or 4 cards of at least 256GB (40 minutes of recording each), preferable 512GB (80 minutes), which would make it more than $2000 —which is basically the rumored price of the GH5 just in storage.

Another issue is that there was a rumor stating that although it might be a Dual Card Slot, only one of the slots will be compatible with UHS-II, making it even less likely to record ProRes and DNxHR. But not just that, if you only have one slot to record, as long as 128GB is the biggest storage available, you will be limited to less than 20 minutes of recording before having to swap cards.

Would Panasonic offer 10-bit 4:2:2 with so many limitations and workarounds? Would Panasonic offer their biggest trump in video for just those that can afford spending another $1000 or $2000 in storage?

I’m not so sure about it.

Why am I saying all of this?

A lot of people say that they want ProRes HQ or DNxHR because they won’t need transcoding and that it will be much easier to edit without really giving enough thought about it.

Some people are already dealing with such large files and using external recorders, but most of them record directly to SSD, which is much cheaper than these brand new UHS-II cards.

The whole point is, are you really prepared for that? Are you prepared to spend more than what you will pay for the GH5 just in media alone. You might not need the fastest UHS-II card out there or even 10 hours of shooting capability but be sure, you will still have to buy these new cards — and it won’t be cheap.

Some could argue that there are more compressed flavors of ProRes like ProRes LT, which can still manage a good video quality but much smaller files. But then, wouldn’t Panasonic be better off using their own codec at similar bit rate rather than paying Apple for ProRes?

Another thing that makes me think that ProRes and DNxHR are not very likely to happen is because Panasonic has its own Intra codec for 4K 10-bit 4:2:2, AVC Ultra. I already mentioned it before, it’s a H.264 based codec and at its highest settings I think it achieves 800mbps, which is already slightly less than the other option.

I’m pretty sure that if Panasonic was to choose a high bit rate codec, they would simply tweak a codec they already have instead of having to pay for ProRes and DNxHR.

Think about it, if the GH5 outputs 10-bit 4:2:2 via HDMI, the cost of the 4K recorder plus some SSDs will end up costing less or quite close to what you would spend with just SD cards — if the Panasonic adopted ProRes and HNxHR — and you would still be able to use the recorder as a monitor while you can’t really do that with expensive SD cards.

And since the GH4 already does 10-bit 4:2:2 via HDMI output and you already use an external recorder, if the GH5’s internal recording consists of ProRes HQ and DNxHR and spending just as much and an external recorder, what are the big advantages of internal recording over the external recorder? Back up media? ProRes Proxy? Some recorders already do that, so basically there is not much if any advantage at all.

Because at this budget, only those that could afford an external recorder in the first place will have the money to invest in these expensive high speed UHS-II cards.

Panasonic certainly takes all of that into account when considering adopting such high bit rate codecs for their consumer camera. So ProRes and DNxHR are still a possibility, I just don’t think it’s the most likely to happen or even the best solution.

I don’t think Panasonic will bring 10-bit 4:2:2 for just a few people that are willing to deal with such massive files and pay for these expensive media solutions. Most likely Panasonic will want that as many people as possible can shoot in 10-bit 4:2:2.

What good is your camera’s best feat if just a minority will be able to enjoy it?

Final thoughts

It’s cool that most of the speculation and analysis I made back in August still makes sense and part of that was already confirmed in this pre-announcement.

Overall Panasonic pre-announced just as an insurance, just in case some other company announces some desirable video camera, the word is out there and people already know that the GH5 is coming. It might look selfish but at the same time, who would be happy buying one camera and then a few weeks or months later you find out that there was this even better camera that you didn’t see it coming?

The pre-announcement consisted of just a single slide with 4 feats, so it’s clear that the GH5’s specs are far from final. With shipment at least 6 or 8 months away, Panasonic still have a lot of time to finalize the specs of the GH5.

Panasonic pre-announced it to hook people up but also to scan and feel what other manufacturers are offering, also possibly studying the feedback on the Internet about the GH5. Panasonic knows that people wanted IBIS for the GH4 and pretty much all recent posts or videos talking about the GH5 is addressing the lack of IBIS as part of the GH5’s feats in Panasonic’s presentation and press-release.

It seems that the G85 — the US version — will be able to do unlimited recording in 4K and this tells me 2 things:

  • Best Case Scenario: the G85 is capable of that with IBIS without any limitation
  • Worst Case Scenario: the G85 will have some time limit for recording 4K with IBIS

But even the worst case scenario still implies that even if you need to turn the IBIS off, the camera can still have IBIS inside. So IF Panasonic really faces some sort of issue with 10-bit 4:2:2 — even though I didn’t find much of an evidence that makes me believe that Panasonic will have ANY overheating issue — they can still turn IBIS off like they do with LX10 without the need of simply strip IBIS off the camera, so at least the user has the choice.

Right now Panasonic is working on the GH5 and trying to pack as many feats as they can until the end of the year. I hope they can put IBIS to work, improve the tracking AF and some more. Having Nick Driftwood working with them is pretty encouraging regarding the codec and its quality.

I still hope for H.265 Intra codec, which I find much more interesting than a H.264 Long GOP one. They will both have about the same file size, but with HEVC getting more and more support from both Hardware and Software, an Intra codec will always be better to edit than a Long GOP one.

Although I still think that there are more chances of Panasonic adopting H.264 Long GOP than any other option.

For now, it seems to be a wait game and hopefully there will be more rumors that will corroborate to the GH5 having IBIS so people can feel more at ease about it. Of course, I also want more details other than IBIS, the sensor and codec are also pretty intriguing mysteries for now and pretty important ones.

Overall I’m quite optimistic about the GH5 and looking forward for the final announcement. The Panasonic GH5’s product page says “Early 2017”, so either the announcement is coming later this year with shipment for early 2017 or as I speculated before, it might be announced at CES 2017.

Oh, Panasonic, how about Auto ISO in manual mode or minimum shutter speed and all the “minor improvements” that can make life much better?

Part 1: Overheating Gate and IBIS rant!

Part 2: Codecs and AF improvements Speculation

Part 3: The Mysterious Photo 6K and Mysterious Sensor!

Extra: The Sensor has to be Multi-Aspect!

--

--