GH5 Panasonic GH5s EBU – Tech 3335 Assessment

According to Alan’s testing, the GH5s dynamic range is about 14.6 stops in HLG, which is said was a creditable performance. CineLike-D is about 13.9 stops and Like709 about 10.3 stops with the Knee switched off. However, all of this is using the 4:2:2 10-bit recording modes. Using any of the 8-bit recording modes will automatically reduce the dynamic range by at least a stop.
Click to expand...

Hmmmm. So the GH5s has 14.6 F-stops of dynamic range when shooting HLG... :-)
 
Those are some interesting ways of measuring dynamic range. First time I have ever heard of CineD outperforming V-log in terms of dynamic range.

The GH5S really is an almost perfect HD broadcast video camera replacement. Combined with the higher sensitivity and timecode support it is a solid broadcast camera for HD productions. When it comes to UHD I wonder what his benchmark is since there are not a ton of 3 chip UHD cameras out there. Netflix is pretty darn picky when it comes 4k standards and they are perfectly ok with a bayer sensor 4k camera. Considering how much the broadcast world is dragging its butt when it comes to 4k I would rather see some 4k now vs the snobbery of rejecting 4k cameras that get us 95% there.
 
My crappy internet barely lets me watch Netflix/HULU/VUDU/Crackle in 720p, what do I care about 4k right now? If the story is good, I'll gladly watch 720 vs. a poor story in 4k. They seem to forget some of this when they spec. minimums for projects. I understand why they want higher res. originals, but I think we are a long way from real sustainable 4k use in the home unless it is download only.

There are some issues with GH5s for HD production, but those issues will revolve around multi-camera productions and the control of those cameras.

And all that said, the images I get out of the lowly hacked GH2 are much cleaner than I get from our JVC Hm700chu cameras. No color fringing from a focus mismatch between the 3 chips being the most noticed issue.
 
Reading the comments on newsshooter some people seem to take issue with the results. I don't know enough to say anything intelligible, but it seems they measured by absolute dynamic range - which they didn't do with other cameras. I haven't compared to my EVA1 but I should take it out. That would be cool if true... but 13.5 stops in vlog seems optimistic.

I also have some questions as to some of their conclusions, such as writing that it is better to leave NR on or even turn it up.
 
Thomas Smet said:
The GH5S really is an almost perfect HD broadcast video camera replacement.
Click to expand...

And that it is fully agreed as it meets EBU Tier 1 for HD. Unfortunately if your end client is the BBC or most European broadcasters they are all signatories to the EBU standards whether we like, agree with it or not. Those standards are found in the following documents. Sadly the GH5S doesn't meet UHD Tier 1 or 2 as stated in Alan Roberts testing based on the existing standards.

http://dpp-assets.s3.amazonaws.com/...ecs/bbc/TechnicalDeliveryStandardsBBCFile.pdf

I draw your attention specifically to page 8 section 1.4.1. DSLR Cameras in the above doc.

https://tech.ebu.ch/docs/r/r118.pdf

I draw your attention to page 9 of the above document to understand the Tier minefield you have to navigate through to supply programming to these EBU broadcasters.

We supplied 13 episodes of a motor sport program to Granada TV in the UK some while back and the easiest way to meet all the requirement and avoid having to get approvals to use certain kit was to shoot it all on 3 chip XDCams which to boot are 8 -bit. Go figure. The only exception to that was for the use of GoPros for certain car / motorcycle shots.

Chris Young
 
Alan Roberts is the most renowned video camera testing expert in the world, writing white papers for BBC/EBU. Thing to keep in mind, measurements of db levels with this cam are only through HDMI port as you would use with an external recorder. I didn't see that much DR recorded internally, but it doesn't matter, Alan tests in accordance with EBU Tech 3335 which rates cameras on equal footing which is awesome. For perspective, he gave the C300MKII DR of 15 stops. Not only that, no resolution loss at up to 2500 ISO, impressive NR too, clean output. Whatever number you look at, GH5s is in the UHD game, the only DSLR/Mirrorless that is.

I just gave my camera a hug.
 
cyvideo said:
And that it is fully agreed as it meets EBU Tier 1 for HD. Unfortunately if your end client is the BBC or most European broadcasters they are all signatories to the EBU standards whether we like, agree with it or not. Those standards are found in the following documents. Sadly the GH5S doesn't meet UHD Tier 1 or 2 as stated in Alan Roberts testing based on the existing standards.

http://dpp-assets.s3.amazonaws.com/...ecs/bbc/TechnicalDeliveryStandardsBBCFile.pdf

I draw your attention specifically to page 8 section 1.4.1. DSLR Cameras in the above doc.

https://tech.ebu.ch/docs/r/r118.pdf

I draw your attention to page 9 of the above document to understand the Tier minefield you have to navigate through to supply programming to these EBU broadcasters.

We supplied 13 episodes of a motor sport program to Granada TV in the UK some while back and the easiest way to meet all the requirement and avoid having to get approvals to use certain kit was to shoot it all on 3 chip XDCams which to boot are 8 -bit. Go figure. The only exception to that was for the use of GoPros for certain car / motorcycle shots.

Chris Young
Click to expand...

The most widely used 4K/UHD cam in the BBC/Sky inventory is the Sony PMW-F55, used for Ryder Cup, World Cup live UHD broadcast. It would not meet the current UHD1 Tier 1 either. The tiers are only recommendations, not mandates and Roberts says in his report GH5s is okay for UHD use if not the prime camera, which no DSLR/Mirrorless will ever be the prime camera in the broadcast/cinema world.
 
He also didn't say it failed on resolution, rather just the sensor count because by his math which he acknowledges "could be an over simplification," the 4x3 sensor size at 10.25 effective megapixels would appear to amount to a horizontal width of only 3704 pixels. If this really mattered, Pansonic could clarify the image is 1:1 pixel for pixel acquisition and not upscaled which they may have done already and otherwise seems unnecessary in light of the intended market. It's amazing it even received this EBU/BBC attention. Alan and EBU have actually tested comparatively few cameras since Tech 3335 was adopted.
 
Also, he measured the sensor width at 4:3, when shooting current TV format, 16:9 should have been used, which uses more of the sensor’s 18.96mm width.
Cheers
 
Whatever his credentials, he made mistakes about the sensor size and pixel count, and his assertion about dynamic range isn't real-world applicable.
 
Yes he made an error assuming the 16:9 was a crop of the 4:3 The 4:3 of both GH5 and GH5S is of course the same 16:9 of course is not. A major error in my mind when the whole discussion about the differences was key when the camera was intrduced
 
Yes, unlike the GH4/5, the GH5S maintains its native resolution on most aspect ratios, except for 1:1. The diagonal measurement is constant from 17:9, 16:9, and 4:3, with the area widths nd length being adjusted to keep the total pixel count at 4K. To me this is one of the biggest advantages of the GH5S, constant FOV across these aspect ratios.
Cheers
 
Last edited:
joe12south said:
Whatever his credentials, he made mistakes about the sensor size and pixel count, and his assertion about dynamic range isn't real-world applicable.
Click to expand...

It’s real world if it’s used for comparison to other cameras tested to the same tech 3335 standard, which puts it within a half stop of C300MKII 15 stop DR. And considering it measured 13-14 whether it was HLG, V-Log/L or cine-D, or that its sharpness was as good at 2500 as 800, or that NR improved it instead of the other way around, it’s an outstanding achievement.
 
From the article:

Alan, however, notes that these dynamic range figures are a little speculative and subjective. He stresses that it is up to the user to decide whether the performance at the extreme ends of the dynamic range are acceptable . . .
Click to expand...

Why are we doing this? There is an international standard for measuring dynamic range, ISO 21550. The definition is when signal equals noise. This requires no human preferences, only raw numbers from cold, hard test equipment.

Reporting what you think is the "usable" dynamic range, based on your "subjective" taste, is a recipe for chaos. Each camera will have several different numbers, spanning a stop or more, depending on whose blog you stumble across.

Going by ISO 21550 will yield a number that is higher than anyone's definition of "usable" dynamic range. But here's why that's okay: None of you took a light meter to the various kinds of locations you shoot at and measured the actual dynamic range, then said to yourself, "Hmm, I need a camera with 12.7 stops." No, all of you did the reverse. You thumbed through dozens of stills and videos from various cameras, found one you liked, read its dynamic range measurement, and said that's what you want. "Wow, the Arri Alexa has really nice dynamic range. Let's see, it has 14 stops. Therefore I want a camera that people say has 14 stops." Or you use it for comparing two cameras. "I can't decide between Camera A or Camera B. Oh, Camera A has 12 stops, but Camera B has 12.5. Since they're equal in every other way, I'll go with Camera B."

We need a stable number for each camera, so that they can be compared. Camera X has an ISO 21550 measurement of 12 stops. Camera Y has an ISO 21550 measurement of 13 stops. I know neither of those are totally clean all the way down, but one thing I do know: whatever I end up thinking is the usable dynamic range of Camera X, I know Camera Y will have 1 stop more. Or again, after watching various videos, I like the cameras that have an ISO 21550 measurement of at least 13 stops, so when I am shopping I will try to buy one that has an ISO 21550 measurement of 13 stops.

This supposed diligence in trying to report the "usable" dynamic range of a given camera has no practical benefit and only muddies the waters.
 
Tom Roper said:
Alan and EBU have actually tested comparatively few cameras since Tech 3335 was adopted.
Click to expand...

Probably because he no longer works for the BBC. He retired some years ago. He now mainly consults and occasionally you will see reviews of his in Zerb the GTC magazine. GTC is the Guild of Television Cameramen in the UK for those of you not familiar with the GTC. Getting back to the standards. I once worked for the BBC. The standards whilst generally observed internally are primarily used for outside producers and program suppliers as a benchmark to maintain a degree of quality control.

I remember some years ago one of the commissioning editor at the Beeb commenting along the lines of when you get around 180 program submissions a month you have to have a common standard to throw at externals and not deviate from that otherwise it would be mayhem if you allowed one production to use such and such kit and didn't allow another to use the same. With big broadcasters there is always a certain amount of political bureaucratic oversight that you have to work under. The standards are also used as a technical filtering tool on on external submissions. In the first step it narrows down the program ideas and submissions they have to look at. If the subs meet the tech standards then what the commissioning editors look at and evaluate are the artistic, educational, entertainment and cultural values and aspects of the program submission. The commissioning editors who approve the programs don't and won't get involved with the tech side of the productions. They just need to be told by tech QA that yes this submission meets the BBC / EBU requirements. One of the reasons for that is that the Beeb on-sell many productions and are involved with many co-productions and need to assure future potential program buyers that yes this program meets EBU R118 3335. In other words it is the big picture they look at and one of the prime considerations is commercial on-sell viability.

http://www.gtc.org.uk/

http://www.bbc.co.uk/commissioning/tv/articles/who-we-are-how-we-commission

Chris Young
 
combatentropy said:
From the article:



Why are we doing this? There is an international standard for measuring dynamic range, ISO 21550. The definition is when signal equals noise. This requires no human preferences, only raw numbers from cold, hard test equipment.

Reporting what you think is the "usable" dynamic range, based on your "subjective" taste, is a recipe for chaos. Each camera will have several different numbers, spanning a stop or more, depending on whose blog you stumble across.

Going by ISO 21550 will yield a number that is higher than anyone's definition of "usable" dynamic range. But here's why that's okay: None of you took a light meter to the various kinds of locations you shoot at and measured the actual dynamic range, then said to yourself, "Hmm, I need a camera with 12.7 stops." No, all of you did the reverse. You thumbed through dozens of stills and videos from various cameras, found one you liked, read its dynamic range measurement, and said that's what you want. "Wow, the Arri Alexa has really nice dynamic range. Let's see, it has 14 stops. Therefore I want a camera that people say has 14 stops." Or you use it for comparing two cameras. "I can't decide between Camera A or Camera B. Oh, Camera A has 12 stops, but Camera B has 12.5. Since they're equal in every other way, I'll go with Camera B."

We need a stable number for each camera, so that they can be compared. Camera X has an ISO 21550 measurement of 12 stops. Camera Y has an ISO 21550 measurement of 13 stops. I know neither of those are totally clean all the way down, but one thing I do know: whatever I end up thinking is the usable dynamic range of Camera X, I know Camera Y will have 1 stop more. Or again, after watching various videos, I like the cameras that have an ISO 21550 measurement of at least 13 stops, so when I am shopping I will try to buy one that has an ISO 21550 measurement of 13 stops.

This supposed diligence in trying to report the "usable" dynamic range of a given camera has no practical benefit and only muddies the waters.
Click to expand...

We're doing this because there is no trusted source of definitive measurements against an objective measurement. (Similar situation exists in comparing projector contrast.) If you'd like to take the time to measure every camera in use, I'm sure we'd all be happy to use that number. ;-)

The other reason is because "usable" is well, usable. Everyone who has used the GH5s knows they can't get 14 stops of range out of it. Never. Not even 13. Probably 12, if they didn't underexpose. 11, yeah, safely.
 
You must log in or register to reply here.
Back
Top