GH5 Available for Pre-Order at $2k, amazing features.

Guys - sorry if this has been discussed and even resolved already, but what's the word on Speed Booster EF->MFT 0.71 or 0.64 compatibility with GH5? Are both going to work? TIA

Piotr
 
Very nice in-depth GH5 tests over at cinema5D, here.

* The internal codec looks awful. There's no point in having a 10-bit codec if it is this bad.
* DR is about two stops short of the a7S II (I was expecting at least three, so that's actually a good result).
* Still noisier than I would like.
* The rolling shutter is pretty fast (certainly A LOT faster than the Sony a6x00 or a7 series in 4k mode).
* When noise is not an issue, it's very sharp and detailed (but we all knew it would be).
 
Samuel H said:
* The internal codec looks awful. There's no point in having a 10-bit codec if it is this bad.
Click to expand...
I would be careful in blaming this all on the codec.
I know it is the national sport for some people on the internet to blame everything on H.264 but I prefer to see an actual A-B comparison confirming that all those flaws go away if recorded by an external recorder.

What is possible in my mind is that the 10bit is mostly an 8 or perhaps 9 bit signal while the rest is noise dithered to 10 bit. It is great to have 1024 discrete values for each pixel but if there is noise the signal may contain no more than 256 or 512 actual values, You can dither a digital signal to a million bits but it won't make any difference, the noise floor, and no I am not confusing this with the dynamic range, determines the actual bit depth of the signal.

By the way I think the over the top criticism on the internet towards Cinema5D is uncalled for.
Believers will continue to believe even if the facts run counter to their expectations and they prefer to find blame somewhere else!
 
Last edited:
I don't know why there is such a strong uproar over the Cinema5D review

A key selling point of this camera is being able to record 4K 10bit internally and at this point it doesn't appear as if there is any benefit to Panny's implementation of 10 bit on the GH5.

Panny knew they were sending a camera for evaluation and Cinema5D did their job and reviewed the camera. At this point, there does not appear to be a significant benefit to using 10 bit on this camera. This puts the onus on Panny to deliver more than just what is stated on their spec sheet.

Now, if Panny sent the camera and expressly told Cinema5D that they were not to review the camera, THEN, I would say that Cinema 5D is in the wrong both ethically and morally but I haven't seen or heard anything to that effect. There have been numerous GH5 reviews with image examples, so I have to believe that Panny is sending these cameras out with full knowledge that they will be tested in the field and those tests are being made for public consumption.

The review also made allowances for improvement and acknowledged that they were working with a pre-production version of the camera. Personally, I think we should be grateful that these limitations were highlighted as they may allow the end consumer to receive a better implemented 10 bit.
 
Cary Knoop said:
Believers will continue to believe even if the facts run counter to their expectations and they prefer to find blame somewhere else!
Click to expand...

That's an entirely simplistic point of view, and only serves to identify you as the same type of person you are criticizing, just on the other side of the aisle.
 
I've just seen the tip of the iceberg regarding all the drama around this article. I really don't care too much.

I don't really care if it's the codec or the 10-bit implementation or whatever: all I know so far is that, unless C5D did something wrong, images recorded internally are far from what they should be. In particular: this is not what 10-bit is supposed to look like.

I certainly don't blame H.264 on this: XAVC-S is H.264 as well, and it looks A LOT better than this. H.264 is basically a set of rules for decoding an image, you can do a good or a bad implementation at the time of encoding, and that will result on a good or a bad image when decoded. Same for 10-bit: it's a simple concept, but you can screw your implementation and make it not work.

To me, it looks like a bad codec issue, but well, maybe it's a compatibility issue from the editing software they used. We'll know soon enough.

And all the comments saying the test was an act of evil just sound like fanboy stuff. I just read a few and they reminded me of the old days when I visited PC hardware forums, with all those kids fighting ATI vs NVIDIA or intel vs AMD. Pretty irrational. Much like sports in general, actually: "my team is better than yours! your team sucks!" I'm very happy dvxuser is not like that.
 
Samuel H said:
To me, it looks like a bad codec issue, but well, maybe it's a compatibility issue from the editing software they used. We'll know soon enough.
Click to expand...

+1 :thumbsup:

I am surprised that the C5D folks did not investigate either of these things further ( or test with an external recorder ), as new cameras often bring new tech formats and our software has to be updated to catch up. ( i.e. remember when h.265 first showed up and nobody could edit it properly )
 
TheDingo said:
+1 :thumbsup:

I am surprised that the C5D folks did not investigate either of these things further ( or test with an external recorder ), as new cameras often bring new tech formats and our software has to be updated to catch up. ( i.e. remember when h.265 first showed up and nobody could edit it properly )
Click to expand...
The whole point, at least to me, of an GH5 is its portability and now with IBIS its handheld capabilities, if it turns out you need something like an Atomos to get decent 10 bit log output the whole point of having a GH5 goes mostly out of the door. I mean if you want to use the GH5 primarily with a solid tripod and external recording you might as well save some extra money and get a 15 stop dynamic range Blackmagic 4.6k URSA!
 
Cary Knoop said:
if it turns out you need something like an Atomos to get decent 10 bit log output the whole point of having a GH5 goes mostly out of the door.
Click to expand...

Adding an external recorder to the C5D test would help them to determine if the GH5 h.264 CODEC is the weak link, and if it is then Panasonic might have to retool their CODEC before the GH5 is released.

i.e. maybe the GH5 standard 4K IPB bit-rate needs to be bumped from 150 Mbps to 200 Mpbs
 
TheDingo said:
Adding an external recorder to the C5D test would help them to determine if the GH5 h.264 CODEC is the weak link, and if it is then Panasonic might have to retool their CODEC before the GH5 is released.

i.e. maybe the GH5 standard 4K IPB bit-rate needs to be bumped from 150 Mbps to 200 Mpbs
Click to expand...
That's a good point and boosting the bitrate would be a welcome thing!
 
Cary Knoop said:
The whole point, at least to me, of an GH5 is its portability and now with IBIS its handheld capabilities, if it turns out you need something like an Atomos to get decent 10 bit log output the whole point of having a GH5 goes mostly out of the door. I mean if you want to use the GH5 primarily with a solid tripod and external recording you might as well save some extra money and get a 15 stop dynamic range Blackmagic 4.6k URSA!
Click to expand...

An Ursa is 3 times the cost of a GH5 before you buy a battery, lens or viewfinder for it. You are in the same space as C5D. Be real, compare at the same price point and feature set. I will save you the trouble...There isn't anything. I would hope that an Ursa, FS7 and C700 would perform better. I look forward to an explanation of the 10bit issue too and do not fault C5D for this reporting. But we need to compare things bearing in mind the economics. A GH5 with Atomos Inferno, lens and extra batteries is still less than the body of an Ursa. To make an Ursa work one needs to spend more than the cost of the GH5 system on top of the base cost. A bare bones Ursa system with battery and viewfinder without lens will come in close to $9K then you have to buy a lens. If you decide to put it on a tripod you will not need the viewfinder so save $1500. But now you have to compete with the GH5 and Inferno at $4K. I guess you view $4K as just a little extra cost. It is twice in my calculations. Like choosing between a Kia and a Cadillac.
 
You must log in or register to reply here.
Back
Top