As for the Arizona falls clips the highlights stayed around 60-70 IRE so and actually most of the perceptual contrast was gained in the image by boosting them instead of lowering the shadows, which I rarely see.
A7s sample footage
- Thread starter squig
- Start date
As for the Arizona falls clips the highlights stayed around 60-70 IRE so and actually most of the perceptual contrast was gained in the image by boosting them instead of lowering the shadows, which I rarely see.
philiplipetz
Veteran
Extensive stills test shot in Brazil during World Cup. Impressive, even when compared with A7r. There is somethng magical about the image it gives. http://briansmith.com/sony-a7s-field-test/
Last edited:
squig
Veteran
Interesting to see the HDR shot vs the single exposure shot, there's not much difference. The magic of the A7s is the huge DR so closely replicates film, that's what sold me on it after seeing Den's Scotland shoot.
Sorry, but basing your notions of dynamic range on this test is very seriously flawed.
It very clearly explains that the test was made using internal media.
There is no way in heaven I would use internal recording on my F3.
It does not come even close to what can be had with an external recorder and SLOG.
Internal it is 35Mb/s. To a samurai in prores422HQ it is 220 Mb/s.
Internal recording to the A7S is 50Mb/s, so of course it's going to look better.
However, output that camera to a recorder you will get maybe 100Mb/s, if that.
Examining dynamic range without asking for the highest quality output the camera can produce is silly.
Put all those cameras on a recorder outputting the same high quality Prores for instance and you will see which camera fares better.
And why it costs more.
The higher data rate of the F3 to a recorder definitely shows more detail in the highlights as compared to internal for instance. Thus a step test reveals more dynamic range. Not a surprise.
This is why making assumptions about headroom with a variety of different internal compression schemes at varying bit rates makes no sense.
Further, it's not just about having information in the highlights and the shadows, it's about the quality of that information. You cannot get around the fact that the A7S is 8bit. End of the day grading will still be problematic and any of the other cameras will produce a higher quality output.
Last edited:
I agree with you in some respects (like after grading) but when speaking of DR and logcurves the ungraded visual image coming from the internal media of each camera I think it will give a pretty good indication of which ballpark they all play in. And about the A7s being 8bit: I think it looks to grade pretty well anyway
But ofcourse its not an F3.
paulcurtis
Veteran
I think this is a bad result for the A7s. It is showing that the blacks are very elevated compared to the FS700. The 700 isn't technically Slog2 there is an additional curve on there (see how much more constrastier it is). The extra contrast is a good thing in 8 bit because it is using the 256 levels better (bit more shadow range, bit more mid with a bit less highlight). The A7s appears to be so elevated that out of that 8 bit container a good 10 to 20% of it is not being used. I would want to see the blacks much blacker.
So it might be that the other profiles could offer a better gradable result.
Hopefully when it ships in the UK (maybe next week? does anyone have one here?) i might be able to tell for myself
cheers
Paul
ReelWorksMedia
Well-known member
this this is a big help. Thank you. Was pondering on when to swap out the 5d3. Looks like sooner than later.
paulcurtis
Veteran
Bit depth won't affect dynamic range. It will create an image that when you grade will band quicker but the black and the white point would be the same. The only time you'd really get more is RAW, and when you can do highlight recovery. Frankly you could do highlight recovery with Slog2 but that's a post process.
So the whites and the blacks from the FS700 are the same whether it's Cine4, Slog2, 4K2HD - only RAW gives you a little bit extra.
cheers
Paul
photochris
Well-known member
The face painted guy ISO6400 crop is awesome, very usable images at 51200 and in some situations 100,000+.
The single shot HDR shot shows the impressive DR, especially compared to Canon. I've owned the A99 and now the A7 and there's far more room to lift shadows and pull detail out of seemingly blown highlights with Sony raw's. The A7s looks to have raised that bar a bit.
Can someone point out the discussion on overexposing the SLOG2 on this camera and why one might think that's better than underexposing it?
From everything I've seen, personally, I've deduced that overexposing it is probably the quickest way to a very dead/dull image, and the camera pretty much wants you to underexpose it (or take advantage of it's S tag) to give you the best that it can offer for internal recording. I'm almost ready to bet that SLOG2, while being amazing, isn't the way to generate the most usable and most pleasing base from the camera after getting some very early stills from Roman.
From everything I've seen, personally, I've deduced that overexposing it is probably the quickest way to a very dead/dull image, and the camera pretty much wants you to underexpose it (or take advantage of it's S tag) to give you the best that it can offer for internal recording. I'm almost ready to bet that SLOG2, while being amazing, isn't the way to generate the most usable and most pleasing base from the camera after getting some very early stills from Roman.
squig
Veteran
Kholi have you checked out the Arizona XAVC S files? The 60p file (C0006) is underexposed and looks like **** https://vimeo.com/99082568
I would imagine the 60P looking pretty bad. What about the non-offspeed material? I'm downloading it now, though.
I can't say that I would overexpose from native, it may be the 8-bit or just the nature of SLOG combined with this system, the top end wants to go extremely fast and I don't get the sense that there're three stops of overexposure latitude. THat's of course just me, I would really like to get my hands on it to explore PP4 and PP5 primarily.
Did he ever tell you if he shot APS-C or Full Frame? Anyone know how that affects compression?
squig
Veteran
I posed the question, no reply as yet but he did shoot it with a Canon L 24-70 and the aliasing is only really showing up on the 60p shot. There's aliasing in the APS-C crop: https://vimeo.com/98672067 24p 1080p looks to be just about aliasing and moire free.
Samuel H
Major Contributor
I get what you are saying. This is just a case of different users, different needs. Anybody planning to use an external recorder should look at tests that use an external recorder. But I plan to record internally, so I looked for tests that record internally
I understand that I'm leaving some IQ on the table, but I like the simplicity and the small footprint, so if IQ is good enough I'll just save the $$$ and go straight to SD cards. If there's a ninja3 with 4K recording for $750 in 2015, I may get one; then I will worry about DR with external recorders.
squig
Veteran
Hopefully I'll get mine next week. I picked up a Sound Devices Pix220 to see how it looks recording the 422 out to prores4444. How do peeps with the F5/55 expose?
Yeah I sawr that. Just downloaded the files and had a look at them, I would have to say that not only is this about what I expect from an 8-bit source, it's actually better than I would ever hope for. haha. THe noise is just going to be part of the image, and I don't find it offensive at all. These examples and Roman's put this camera ahead of every other DSLR out there for me, or rather at least on the line of shooting 5D3 RAW and GH2 Hack (still so special).
I get what you are saying. This is just a case of different users, different needs. Anybody planning to use an external recorder should look at tests that use an external recorder. But I plan to record internally, so I looked for tests that record internally
I understand that I'm leaving some IQ on the table, but I like the simplicity and the small footprint, so if IQ is good enough I'll just save the $$$ and go straight to SD cards. If there's a ninja3 with 4K recording for $750 in 2015, I may get one; then I will worry about DR with external recorders.
I agree... I kind of feel like throwing external recorders on begin to defeat the purpose of just rolling with a camera. It's a shame that it doesn't shoot straight 4K internally.
dfreerider
Well-known member
I think he did use the APS-C crop on some clips, since he said it was very useful. But no idea which clips are which.
Probably not relevant but the flower test in crop mode is shot in AVCHD. Would like to see some crop footage shot XAVC-S in case that makes any difference.
I also just ran a really quick and crude NR on the Arizona stuff, it looks great. Again, nothing that I wouldn't expect going into the format, and it's seriously looking a lot better than most of the options that have dropped very recently. I'm pretty impressed with this thing... maybe not with the price though. =P
Not sure if Sony can claim 13 in video, but numbers are a bit arbitrary at this point... if it looks good when you grab all of the information that you need, or you don't have to work too insanely hard to use what you have, then it's good enough.
Edit edit: I think you guys might want to look into IR filtration, additional. I don't know if that's what I'm seeing in some of these arizona shots but whatever it is takes a toll on greens. It also might just be me.
Not sure if Sony can claim 13 in video, but numbers are a bit arbitrary at this point... if it looks good when you grab all of the information that you need, or you don't have to work too insanely hard to use what you have, then it's good enough.
Edit edit: I think you guys might want to look into IR filtration, additional. I don't know if that's what I'm seeing in some of these arizona shots but whatever it is takes a toll on greens. It also might just be me.
ReelWorksMedia
Well-known member
Wondering if the A7s has any type of view assistant, like the C100, when shooting Slog to help simulate a corrected image on the live view screen. Just makes real life live shoot environments more pleasing to shoot in. Would like to know that....or if just monitoring the WFM is really the best method and to not trust your eye as much.
That feature in the C100 is very nice.
That feature in the C100 is very nice.
squig
Veteran
Would you want to expose and focus with S-log on a 40 day shoot? I think the LUT on the Shogun is gonna be pretty essential. I'm hoping I can tweak my pix220 or AC7 enough to get a decent image until the Shogun ships. It does look great for 8bit but like any 8bit image it's a bit fragile and it's gonna be super important to nail the sweet spot in camera.