G
Guest
Guest
Anyone who has shot fire at night knows that the fire becomes the main source. here it is swallowed by the ambient light of the Sun, just after sunset. the f2.8 is a lens spec, or Sony has their iso way way off.
Rumor: Sony A7s will have 4K video
- Thread starter vanvideo
- Start date
G
Guest
Guest
you and i wish, brother, you and i wish.
also, in a clear sky like that, there should be stars, not a swath of pale blue gradient, and purple tinged clouds. trust me, this scene could be shot with lower iso. may be some fudging on the specs or numbers.
Last edited:
Remember that this is scotland, which is fairly northern. In the summer it doesn't go fully dark. It doesn't specify whether an nd is used in this shot though so you may be partially correct.
Ian-T
Veteran
Kind of what went through my mind.
If there was an ND in the shot then that's definitely fudging specs, otherwise I don't think anyone would lie about the settings. I did consider there might have been an ND on the lens and Den forgot about it. I do it alllll the time.
G
Guest
Guest
It is more about source/contrast and cues of nature, than it is about clipping and dynamic range. Wouldn't matter if it had 30 stops of latitude. The ambient light is from the sun below the horizon. A light that most people can walk around by. The fire just warms up the man's face as he leans towards the fire pit, instead of being the light source. Even the human eye would never respond that way if it was complete night.
Last edited:
squig
Veteran
Lulz, I can't leave you kids alone for a second. I go away for a few hours and come back to conspiracy theories and 31ms rolling shutter. I'm not buying any of it, but I'm glad everyone's been turned off the A7s, now I won't have to wait in line and it might be a bit cheaper.
I'll test the rolling shutter and dynamic range as soon as I get my hands on one and send the data to Samuel for analysis. Those figures can be trusted because we've put a bunch of cameras through the same tests.
If the RS really is 31ms I'll have a cry.
I'll test the rolling shutter and dynamic range as soon as I get my hands on one and send the data to Samuel for analysis. Those figures can be trusted because we've put a bunch of cameras through the same tests.
If the RS really is 31ms I'll have a cry.
pulpfiction007
Veteran
The rolling shutter is pretty bad on both the new Sony cameras...get prepared to shed some tears, lol
Simon Shasha
Well-known member
For those with their heart set on the A7s, I hope Sony can somehow help reduce the rolling-shutter significantly via firmware.
Also, I don't know what lens is on that A7s in the video, but I'm going to guess it's the Sony 70-200mm F4? If so, and I'm not excusing the A7s here, but a lot of CMOS based cameras exhibit rolling-shutter like that on long lens...
Anyway, to each their own. Personally, I'm more interested in the GH4 for video, the A7s for my street-photography and use with old manual-glass.
Also, I don't know what lens is on that A7s in the video, but I'm going to guess it's the Sony 70-200mm F4? If so, and I'm not excusing the A7s here, but a lot of CMOS based cameras exhibit rolling-shutter like that on long lens...
Anyway, to each their own. Personally, I'm more interested in the GH4 for video, the A7s for my street-photography and use with old manual-glass.
squig
Veteran
I haven't done a shot list yet but my guess is 50% of my film will be shot on sticks, 20% steadicam, and 30% on a weighted/balanced shoulder rig. If I recall correctly the BMCC is 25ms? and so is the 5D MK2. I shot loads of stuff with the MK2 and rolling shutter wasn't an issue. I shot with the Nex-5n (29ms) and that was okay too with a bit more care.
I doubt it's as high as 31ms.
Samuel H
Major Contributor
(here) 5D2 is 25ms, BMC is 26ms, if the A7s is like those (which would be what my test would say if the NAB clip was at 30p), it's actually a sizeable improvement from the 29ms of my 5N. 31ms is bad (nearly like the D90), but over the last few years I have somehow learnt to move the camera carefully without thinking about it (that's probably why, before measuring it, I thought my 5N was better than my old T2i in this respect, when it's actually the other way around).
And I very much doubt RS can be improved through firmware. Firmware can increase the MHz at which the sensor operates, and this would reduce RS, but also generates more heat; if they're being very conservative with the low-level settings for the prototype cameras, this could happen; but I doubt it would be a big-enough change that we'd notice the difference. What we need for reduced RS is either a hardware change (more efficient heat removal through the body: quite unlikely) or a metal layer spin on the sensor design (a change in the fabrication guideline that tries to get rid of a problem, in this case leakage or some other issue that generates heat and forces a low MHz setting; it takes between two and six months to fabricate and test it, but it could be already underway; quite unlikely too).
What they could do through firmware is to offer a "fast operation" mode that has faster RS but generates a lot of heat and doesn't allow takes of more than, say, 30s, and demands a couple of minutes of cooling off after each one. Pair that with a conservative auto-off feature to avoid bricking cameras, and it may be a possibility even for a conservative manufacturer. It would be perfect for my needs, but since it would be a first in the industry I would tend to think this one is even more unlikely than any of the others.
As you'll see, I'm happy to jump on the negativity train for a few days, until the price is settled
And I very much doubt RS can be improved through firmware. Firmware can increase the MHz at which the sensor operates, and this would reduce RS, but also generates more heat; if they're being very conservative with the low-level settings for the prototype cameras, this could happen; but I doubt it would be a big-enough change that we'd notice the difference. What we need for reduced RS is either a hardware change (more efficient heat removal through the body: quite unlikely) or a metal layer spin on the sensor design (a change in the fabrication guideline that tries to get rid of a problem, in this case leakage or some other issue that generates heat and forces a low MHz setting; it takes between two and six months to fabricate and test it, but it could be already underway; quite unlikely too).
What they could do through firmware is to offer a "fast operation" mode that has faster RS but generates a lot of heat and doesn't allow takes of more than, say, 30s, and demands a couple of minutes of cooling off after each one. Pair that with a conservative auto-off feature to avoid bricking cameras, and it may be a possibility even for a conservative manufacturer. It would be perfect for my needs, but since it would be a first in the industry I would tend to think this one is even more unlikely than any of the others.
As you'll see, I'm happy to jump on the negativity train for a few days, until the price is settled
Last edited:
Nirv
Carbonite Member
It's easy to say you can deal with the large amount of rolling shutter by moving the camera carefully and of course all camera moves should be motivated and performed as skilfully as possible but at the end of the day no matter how careful you are I suspect you'll get plenty with the A7s.
I've shot many a scene on my FS100 on my blackbird stabiliser and during moments of panning while orbiting there's RS coming into the background and that's with the FS100 so I can imagine the A7s being a lot worse in a similar move. The point is you can mitigate RS as much as possible but you're going to find yourself in a situation where the camera lets itself down and you better hope if it's a professional situation the producer doesn't kick you down the road. The RS on the FS100 was so minimal the only time I was concerned was when shooting for a 3d insert and even then the error was minimal. Sony should be coming to market with cameras that equal or better FS100 RS not go backwards.
Anyway, I suspect this is the first use of a sensor in a whole line of new cameras and maybe in a real camera form factor it can be clocked higher and cooled more effectively such that it is a very compelling product. I'd like to see a FS100 body with this sensor, NDs, 10bit compressed XAVC record onboard and 10bit out. Sony, you'd have my money tomorrow.
I think the term "deal breaker" could be relevant if accompanied by, "if it really looks like this on a normal lens and if you do a lot of fast moves and action work where you move your camera quickly."
We produce a lot of action and narrative and TV spot work on sticks, jibs, dollys, sliders and shoulder rigs, and all of the shots are carefully moved - even when producing in quick changing environments.
I just don't get jello issues that matter in the finished piece with any camera we use - and we use a lot of cameras.
Again, I don't know any professional that moves a camera like in that weird, zoom and wiggle at the monitor thing I saw.
There's absolutely nothing about that that I could relate to professional work and therefore no way I would base any opinion on it.
I'm guessing it's probably similar to a 5D or 1Dc and I've never had jello affect any work I've done with either of them with any style we use.
It's like the global shutter on the BM4k.
Sure, I'd prefer less RS, but I wouldn't exchange it for the loss in DR of my other BM cameras.
I can control RS but I can't add DR to a camera.
We produce a lot of action and narrative and TV spot work on sticks, jibs, dollys, sliders and shoulder rigs, and all of the shots are carefully moved - even when producing in quick changing environments.
I just don't get jello issues that matter in the finished piece with any camera we use - and we use a lot of cameras.
Again, I don't know any professional that moves a camera like in that weird, zoom and wiggle at the monitor thing I saw.
There's absolutely nothing about that that I could relate to professional work and therefore no way I would base any opinion on it.
I'm guessing it's probably similar to a 5D or 1Dc and I've never had jello affect any work I've done with either of them with any style we use.
It's like the global shutter on the BM4k.
Sure, I'd prefer less RS, but I wouldn't exchange it for the loss in DR of my other BM cameras.
I can control RS but I can't add DR to a camera.
Andrew Gupta
Active member
I am outraged that Sony can't make ISO 400,000 usable (for the foreign audience -- that's sarcasm).
I'm not an expert on CMOS sensor read speeds. So, my question to the gurus of this forum is -- is it possible that read speeds are slower on 4k than with 1080? As we enter this very first generation of 4k cameras -- should I be expecting slower CMOS sensor read speeds?
As a side note -- we are in the very first generation of these 4k cameras. Think back to the issues with the earliest HD cameras. But -- if that low light of the A7s is a glimpse into the future -- well -- the future is both bright and interesting.
Samuel H
Major Contributor
And this one seems to have more RS than any of them, by a landslide in many cases (e.g. wrt the 5D3). I think the only one worse than this one is the D90.
But yes, I get that if you don't move the camera like a maniac this is not a problem. Scotland still looks great.
Samuel H
Major Contributor
Yes, that's possible. It depends on the sensor design. If it has in-sensor binning with full read-out (in this case 2-by-2-to-1 binning), this is quite likely. The cmosis cmv12000 has that, but most sensors don't (if this one doesn't, then it has to read everything, taking as much time as for 4K recording, and then the processor downsamples the image).
I can honestly say, in the studio or on location, I've never had to use that caveat with any crew <g>.
"OK, before we start I want to remind everyone, as a rule, as long as no one moves cameras or supports or for that matter lighting or talent, like maniacs, we won't have a problem."