Bern Caughey
Veteran
^Thanks!
Help choosing between GH5 and a6500
- Thread starter mproust14
- Start date
Off Tom Antos's tests on YouTube, CineD is 8 stops, V-Log is 10 and a half'ish, maybe 11. GH4 photo Raw was tested by DXMark to be 12.7.
BTW, those tests were most awesome because, instead of shooting a chart, he placed ND filters over cutouts in front of a light panel and then compared the results on the histogram. That gave not only the total number of stops per camera - GH4, 5D MK III, Ursa Mini 4.6, etc. - but showed where each stop was.
Cary Knoop
Veteran
Here is a snapshot of the results:
Samuel H
Major Contributor
Very interesting test. I prefer the ones where you shoot a regular scene and then start overexposing and underexposing until some parts of the image break down, because these chart tests don't account for noise very well and can overstate the usable DR by several stops, but it's well done and very informative.
(another limitation could be that ND filters are not always as precise as we tend to think: several of my Tiffen screw-ins had errors of about a third of a stop, like, the 3-stops one was actually 2.7 stops; but I'd think Tom Antos uses better ND filters than I do so maybe there's no problem here)
Then, watching the very intersting video review, I see that on the G85 with cinelikeD he counts 8 stops and on the GH4 with vlog he counts 11, but also on the GH4 with cinelike D he counts 11, which would mean "no advantage from vlog". Looking at the video instead of the waveform, you could say it's 11 and 10 stops, so a 1-stop advantage for Vlog.
If you bring that to my test, it means the GH4 is still at least 2 stops behind the a7S in terms of dynamic range.
On the a6300 he finds 13.5 stops with slog3 and 12-to-13 with slog3 (he doesn't cound that last stop but I would say it's almost a useful stop; here's where the other kind of test would help, it would be a better way to see if that last stop is useful or useless). So, he also finds a Sony advantage of at least more than one stop for the Sony, going up to 2.5 stops of advantage if you're willing to use slog3 and denoise your footage in post (as I think would be mandatory on the GH4 anyway)
The biggest winner in that test, for me, was the BM 4.6k. I already knew it had awesome DR, but wow, does it shine here. He's counting 15 stops, I would stop at 14.5. But keep in mind that this is RAW: when you move to the real-world tests, it's the compression that kills a big chunk of DR in the shadows, and that's not going to happen to this one. It was the Pocket that convinced me that it was better to shoot a real-world kind of test than a chart, because charts tend to hide the very visible advantage of uncompressed RAW, particularly on the shadows. So: very impressive performance.
Join that with the overexposure test that Hurlbut shot (here), where he found the 4.6k footage would survive levels of overexposure previously safe only on the Alexa (5 stops!!), and you have me lusting for a 4.6k pocket. Please BM, you skipped the last NAB, make a big spash on this one <3
Edit: he's counting 10 stops for the 5D3, and I think that's too generous, the last one is almost the same as the background, I would stop at 9. That's at least 3 stops worse than the a6300, 4 if you go for slog3. And at least 5 stops behind the BM, not even counting the boost from RAW when you go to real-world shots. The 5D4 shoots 4k, but it doesn't seem it has any better DR than this, which means, for me, it's completely irrelevant. Man, did Canon fall behind in the video DSLR market...
Last edited:
Well, they began to ship 4.6 Mini just prior to the NAB'16 without the previously promised global shutter and have been catching up from there.
As to new product, I guess people might be seeking Pocket 4K or something with a higher ISO/better low light performance. On the other hand, BMD is not going to compete with A7s series or whatever else Sony can make based on its sensor and so it might as well stick with the "best bang for the buck" niche. In that sense, Mini 4.6 might be eligible for a price break.
PS. Early in that camera comparison video, Tom says that he tested his filters with a light meter and they checked out.
PPS. Canon has virtually abandoned - by agreeing with its nominal competition to abandon it - the sub-$10,000 video market. I figure that these abandoned niches get filled up at some point. The only question is whether it will be in the DSLR (unlikely) or in the cinema line.
