why can't sony keep up to the gfh5

garya

Member
If Panasonic can make a camera that does 4k video "internally", 10 bit, 422, 400 mbs, why can't Sony do it? Is it the sensor size between the two cameras that make it difficult for Sony? Or is it the suits at Sony marketing protecting their more expensive line of cameras to not lose sales to the lower cost a7sii? Or are they going to just ride it out as long as they can knowing the have the best low light video camera out there at that price range?
 
If Panasonic can make a camera that does 4k video "internally", 10 bit, 422, 400 mbs, why can't Sony do it?

I think Sony will, but it's not easy to pack enough processing power into an A7S size body and deal with the heat it generates. My guess is that the A7S III will bring 10-bit before the end of 2018.
 
possibly. the gh5 is not much bigger. End of 2018 is a looooooong time to wait. I saw rumors/leaks that it might have 10 bit but 420, 100 mbs thru hdmi for external recording. Once again, the gh5 has internal with 422, 400 mbs. the only thing that makes sense is the mft sensor instead of full frame. Also, the a7sii, if real, will be next year. As we all know, rumors are not always true and maybe Sony will surprise us with an announcement in December this year with a release date before 2018.
 
If we're going by the previous cycle, the a7SIII will be announced/released 2-3 months after the a7RIII.
 
If we're going by the previous cycle, the a7SIII will be announced/released 2-3 months after the a7RIII.

...I don't know if I would count on this. So far I've seen nothing official from Sony about the existence of an A7SIII, but lots of rumor sites have been predicting it's release all the way back to Spring 2017.
 
the gh5 is not much bigger.

...But neither is the GH4, which also has no overheat issues. Keeping a Full Frame sensor cool while processing a lot more bits does not sound like an easy task.

End of 2018 is a looooooong time to wait.

Yes it is, but Sony doesn't even have 4K 10-bit internal recording in their FS5 camera, so adding it to the A7SIII is going to be a big deal. ( unless Sony wants to keep 4K 10-bit as a FS7 only feature )

Once again, the gh5 has internal with 422, 400 mbs

This was a big deal for Panasonic to pull this off, and I don't think it's an easy thing for other camera brands to duplicate. ( Canon will probably never do this, Nikon maybe before 2020, and I expect Sony will have this in 2018 )
 
...I don't know if I would count on this. So far I've seen nothing official from Sony about the existence of an A7SIII, but lots of rumor sites have been predicting it's release all the way back to Spring 2017.

You never see anything official from Sony; it's almost always a surprise (besides the FS7II).

No one saw the a7SII coming. I used to be a big supporter of the rumor sites, but I am not anymore after constant mistakes this past year. I understand it's a tough job and they have their sources, but even those sources might sometimes not be in the loop or just completely wrong.

I'm only going by what Sony did with the a7RII and a7SII; June and September announcements.

Of course everything is subject to change, especially since the a7III hasn't been announced and the a7RIII was (out-of-order).
 
sony chose to work on other more important things first like picture profiles and video auto focus and photography. Nobody can see a different in 8 vs 10 bit
 
why can't Sony do it?

I don't think that they 'can't', I just think that they didn't expect they would have to. The GH5 was an anomaly when it was released with 10-bit 422 internal. No other DSLR or mirrorless camera had featured it before; in fact very few cameras under ~$10k had featured it at all.

Of course, the GH5 has changed what people now expect from all manufacturers. I'm sure Sony will adjust eventually, but keep in mind that these products often have a lengthy design schedule and are mapped out well in advance. They can't just rush a product to market to compete - they have to readjust their whole development roadmap which means it could be several years before they get something to market. Again, it's not because they couldn't - it's because it just hadn't been in their plans and they're starting on the back foot.
 
As a Lumix user for many years, the A7 series of cameras really have redefined what you can do with a 'small' mirrorless camera.

I don't care much for the colour or ease of use of them, nor lens size. (and for my use underwater, their manual white balance is shockingly poor)
But credit is due to the diminutive bodies sizes, with IBIS and full 35mm sized sensors.

I do get a feeling though that all this talk about the GH5s being the new low light king will really be laughed at when the A7sIII comes along. The A7sII it seems is still better than the GH5s at very high ISO and I would imagine Sony improving it still. But we are getting in to the realms of 'why' do we need a 512,000 ISO etc..

Exciting times ahead.
 
Last edited:
What with 250gb hard drives costing $1000 who would want all that data?

Sony still think it is 1995 hence low data rates on prosumer cams?
 
In the real-world tests that I've seen (most well known is this one), GH5 4k 10-bit 422 footage doesn't show much of an advantage vs Sony's XAVC-S 4k 8-bit 420. But the difference in DR is HUGE, with the Panasonic falling two to three stops short of the a6300 (link).

So you might as well ask: why can't Panasonic keep up with Sony?

And really: I don't see the point of asking any of these. They are just different. People with different tastes and priorities will choose one or the other. The Sonys match better what I want in a camera, but I know lots of people who have used both and prefer the Panasonics.
 
In the real-world tests that I've seen (most well known is this one), GH5 4k 10-bit 422 footage doesn't show much of an advantage vs Sony's XAVC-S 4k 8-bit 420.

The problem I have with Dave Dugdale's comparison is that he does not start with an A6300 image that shows banding, and then he should show us the SAME scene shot with the GH5 in 10-bit 4:2:2, where I would expect to see the 10-bit footage to show a clear advantage.

The first thing I noticed when I first got my GH5 and used it on an actual job, was how much better the 10-bit color was, especially when shooting under the horrible fluorescent lights of a hospital. Getting good skin-tones was significantly easier with the 10-bit footage.

But the difference in DR is HUGE, with the Panasonic falling two to three stops short of the a6300

This is Tom Antos's DR test of the GH4 internal 8-bit recording. From my own experience and from other GH5 reviews, the GH5 produces a better DR than the GH4 does with an external ATOMOS recorder. Despite appearing to have the same DR on paper, there's a full stop of additional DR with GH5 10-bit internal LOG over the GH4 8-bit internal LOG. With some of the real-world tests (not shooting charts) of the SAME scene shot with the GH5 in 10-bit V-Log L mode versus the A6500 in 8-bit S-LOG mode, to my eye the visual DR from each camera was a toss-up. ( but I prefer Panasonic color over Sony color )

Now we've got a new Panasonic GH5s camera which looks like it's going to be the mirrorless camera to beat at 2500 to 12800 ISO, so I look forward to seeing some real-world tests with several cameras shooting the same scene under the same lighting to see how the GH5s stacks up. It should be interesting.
 
8-bit vs 10-bit: you complain that he's starting from a video that doesn't show banding to begin with, but I've been using these cameras for three years now and I've never had footage that shows banding to begin with (and even after grading, it's never been a big issue for me, the codec breaking apart in dark red areas is my biggest issue so far regarding these issues) (if I use LUTs, yes, it appears inmediately; but I use curves and manual adjustments instead). But if you don't like that comparison, you can try this one as well, they really try to break Sony's codec here.

DR: you're right, I forgot that was the GH4. But then all the tests I saw said GH5 DR wasn't any better than GH4 when recording internally. I don't want to use an external recorder (only used it once, and I really disliked the experience, it slowed me down way more than I anticipated). Yet you clearly don't mind. So we're back at my final point: diferent tastes lead to different products. It's good that we are at such a point: many great tools, for different tastes and situations.

(And yes, I would also like to see side-by-side tests with real world stuff and not just charts; seeing the patch in the chart and actually being able to recover usable info on stuff that is that dark are two very different things; but if it's black to begin with, there's no chance)
 
I think most banding is caused by compression not lack of colour data.

So an 800mbs 8bit would wipe the floor with a 60mbs 10bit

Therefore one may conclude that paper specs are in the main irrelevant and side to side testing would be the leveller?
 
I think Sony is under presure because of GH series. Now almost everybody thinks that 8 bit codecs sucks. Even though those guys dont have sufficient monitor and have no idea what proper color grading is. But they still demand 10bit codecs for couple of bucks. Sony's decision to go with full frame makes it more complicated compare to m43. I suppose Sony itself dont know if they want to bring 10bit to alpha series because of protection of FS series.

Simply 10 bit (same for IBIS and AF) is big thing marketing wise.
 
Panasonic has kept a step ahead of Sony on certain video features like internal 4K, 10-bit HDMI out, and now internal 10-bit. I predict that the A7SIII will output 10-bit UHD through HDMI, but will still be 8-bit with internal recording.

Regarding LUTs causing banding, here's a tutorial by Matt Scott on deconstructing a LUT.
 
8-bit vs 10-bit: you complain that he's starting from a video that doesn't show banding to begin with, but I've been using these cameras for three years now and I've never had footage that shows banding to begin with

I've been shooting 8-bit until I upgraded to the GH5 (15+ years of shooting), and most of the time I would not see banding with any of my shots, but occasionally with mono-color long gradients (usually walls of buildings where the light slowly faded across the wall) I would see some banding. This is the type of subject where I would expect 10-bit image capture to show less or no banding compared to 8-bit.

But then all the tests I saw said GH5 DR wasn't any better than GH4 when recording internally.

I only saw this in all 8-bit reviews. The GH5 has cleaner blacks than the GH4 which gives you the full 12 stops of DR, where the GH4 has too much noise in the blacks to go beyond 11-stops.

I don't want to use an external recorder (only used it once, and I really disliked the experience, it slowed me down way more than I anticipated).

From my own experience there's no DR advantage to using an external recorder with the GH5, but you get an extra stop from the GH4 because of noise in 8-bit V-LOG L recordings. ( external recorders are still useful for other reasons like 1000 nit HDR display, high bit-rate ProRes recording, ability to capture to SSD drives, etc... )

So we're back at my final point: different tastes lead to different products. It's good that we are at such a point: many great tools, for different tastes and situations.

Yes, I totally agree here. Every piece of gear has pros and cons, so you just have to find what works best for your work.

I also feel that maximum DR has been over rated by people comparing cameras using charts, when they should be comparing real world footage of the same scene. I honestly don't expect to see a lot of difference between 12 and 13 stops of DR in normal shooting. Highlight handling is more important to me than maximum DR, as I don't mind less DR if the highlights have a natural roll-off. ( this is the one thing that film was always great at )
 
Simply 10 bit (same for IBIS and AF) is big thing marketing wise.

I would not agree with this statement when it comes to color correction or grading. My first jobs shooting GH5 10-bit footage under crappy hospital fluorescent lighting, showed a huge improvement when trying to correct color, especially for skin-tones. From this experience I would not want to go back to shooting 8-bit because of the better color 10-bit footage gives me. ( I've been playing with the GH5 V-LOG L ---> ARRI ALEXA color transform LUTs, and hope to soon try them on an actual job )
 
Back
Top