Just got a new camera - Sony F35

macgregor said:
OK just to summarize the problem.
When recording to prores 444, the full range is clipped and only limited range is shown in post. Therefore, the tip of the highlights are clipped, and no matter what levels you select on resolve, or if you try lowering the gain, no chance to recover those.
This is not F35 fault, nor F3, it seems to be prores problem.
Before this, with prores HQ, while a limited range codec, you could select data levels in resolve, and be able to access those clipped highlights. But with prores 4444 that doesn't seem to work.
ALexa encodes their logC as limited range when recording to prores, just to avoid this. But Slog is a full range codec, and whites reach 104IRE.
Click to expand...

This is the reason I always record to DNxHD with the Hyperdeck. But if Prores 444 is not working the only other chance would be Cineform since DNxHD doesn't do 444.
 
saw avid whitepaper - avid dnxhd 444 exists - don't know if the Alexa can record to this. on the white paper it says its 10 bit, however.
 
back to 444 arguement - posted by Alister Chapman on dvinfo.net:
"OK. The question over what's happening to illegal levels was raised by a commenter on my blog and I've dug still deeper.

First of all: There is NO difference in the dynamic range of the Gemini recordings and the 422 recordings. This is what I had expected, but not what my NLE's (both FCP and CS5.5) were telling me.

The problem is this. The Gemini is RGB and when you drop RGB into an NLE's time line the levels are mapped to legal levels, i.e. nothing above 100% (unless you have superwhite mapping turned on in FCP). The Samurai and internal F3 recordings however go up to 109%. But the NLE maps the output to the monitor or frame grabs ignoring anything over 100%, so the illegal whites get clipped and the appearance is that the 422 recordings (illegal levels) have less range than the 444 RGB (legal levels).

My bad. Sorry if I've made anyone panic. I guess I'm going to need a decent dual link 444 monitor to really analyse 444 footage. With 422 I can just plug it in to my Transvideo and measure everything with that.
__________________"

haven't tested it out yet but maybe its mapping the 444 rgb to the 100 IRE levels, not keeping it in super white land.
 
Apparently, from what Sound Devices Tech told me, their 444 is REC709 only which would map it to 100. The phone-level tech wasn't able to comment as to why, but did say that they realize it's an issue.

Eddavid, I'm very curious if the Pix240 sees and reports the F35 as 12-bit when you switch the AJA multiplexer. When I called their tech support, the were a little baffled by my question, but started rifling through the docs and found that there's a dip switch to force 10 or 12. I'm suspect that there's a flag sent over SDI that is non-standard and non-Sony blessed gear will not see it. At this point, I have no substantiation other than comments from people who are "in the know".
 
Confirmed - with AJA multiplexer out from the F35 recording out at 444 rgb 12-bit - with dipswitch set to RGB and 12-bit - the PIX 240 is reporting the signal in at 12-bit!

12-bit 4444 files out of the F35 - BOOM!
 
eddavid said:
Confirmed - with AJA multiplexer out from the F35 recording out at 444 rgb 12-bit - with dipswitch set to RGB and 12-bit - the PIX 240 is reporting the signal in at 12-bit!

12-bit 4444 files out of the F35 - BOOM!
Click to expand...

So is the PIX240 the most affordable way to record 444 with the f35 or F3?
 
LeavingTheCandy said:
I've been working on some half-way scientific tests of the Pix240 on the F3 and F35. My findings are interesting and disappointing:

The setup was each camera going to Pix240 running Aurora (3.0 firmware), recording 422 and 444.

I used a profile chart, purposely over exposed to see where/how each camera clipped and where the superwhites ended. This is the setup with my target. Put some black in there to make sure black was as black as possible. (again, didn't pay much attention to mid-gray, etc).

ScreenShot2013-02-27at95951PM_zps2a42a451.png


Dropped the files into Resolve and used DATA LEVELS, not VIDEO LEVELS. I'm interested in what each was doing. I pulled the highlights back down to see how they clipped. The line between 896 and 1023 is at 940 (or top of REC709)
This is the result:

F3 at 422 ProResHQ
ScreenShot2013-02-27at95811PM_zpsb7100eb5.png


F3 at 10-bit ProRes4444
ScreenShot2013-02-27at95825PM_zps26f14900.png


F35 at 422 (ProRes HQ) Note that there's a little highlight detail is still there, but not as much as F3 (I kinda of expected this).
ScreenShot2013-02-28at80212AM_zps0b7769da.png



F35 at 444-10 bit (ProRes 4444)
ScreenShot2013-02-28at80230AM_zps0d99f5b6.png


Interestingly enough, 422 showed more expected behavior than 444. So, I placed a call to Sound Devices. Tech support told me that they only support REC709 (0-100) at this time and are exploring LOG and RAW options for a future time. I asked if this was a software or hardware issue and the tech did not know.

For tech-heads, since the F35 only has 1.5G HDSDI, I ran two outputs into a BlackMagic SDI multiplexer to get 3G out to the Pix. I ran the F3 (with paid SLOG) direct to Pix via 3G and via dual 1.5 into the multiplexer and saw no differences.

On an additional side note, the Pix reports seeing 10bit 444 coming from both devices. While the F3 only does 10bit, the F35 does 12, yet Pix reports as 10 no matter the setting between 10 and 12bit. I have yet to determine if this is simply the Pix not reporting what it sees correctly, or if Sony has embedded some flag into their SDI (proprietary, undocumented and secret of course) that requires their recorder on the back to actually honor it.

I have determined through some additional homework that the pins on the main-connectors on the F35 will not even POWER another device, unless an appropriate handshake is done, thus allowing it. (in other words, Sony has locked-down the interface I/O to play only with things they choose to play with). I do not have the resources to actually confirm this, so take it for what it's worth.

Any thoughts or comments are more than welcome here. Enjoy!
Click to expand...

Quite interesting tests. Are you still going to post the F35 vs F3 tests you mentioned a few pages back?

It would be very helpful.

I'm undecided if I should get a F3 with S-Log 444 or F35. As far as I could research here's where I'm standing.

F3, better low light, better dynamic range with S-Log, more flexible mount, light and smaller, more compressed codec option on top of 444 with XDCAM for quick and dirty jobs.

F35, slightly better highlight handling, 12 bit, winder color. But for 12 bit recording you need to add considerable extra costs.

But as they say the proof is in the pudding. So it would be great to see comparisons from those who have both. Have you given up on doing the tests? Thank you in advance.
 
I don't have scientific tests yet - I have real world tests.

The F35 feels much more like 35mm film - how it reacts to light, how the motion of it moves, how it handles sharpness, even its internal grain pattern.
I will show screengrabs when I can.

The F3 is still a really nice camera, but has a sharpness pattern that doesn't quite feel as organic as the F35.

Also the weight of the F35 and the balance of it is better for handheld. The shutter is quite beautiful and lovely .

12 bit recording is not that much more - I am getting it with the PIX 240 and the AJA Multiplexer. For the F3, I already would use the PIX 240 with s log. My 12-bit tests vs 10 bit pro res hq - now that's something I want to do actual scientific tests on but I would need a colorist to verify my findings.

A camera shouldn't be used based solely on tests. On tests, the Alexa probably beats the F35 hands down. But I still find the F35 personally a much more pleasing image.
 
Slightly off topic.... A question for those using the Pix 240 to capture both audio & video, any offset issues between the captured audio & video sync?

Thanks!

Dave
 
David W. Jones said:
Slightly off topic.... A question for those using the Pix 240 to capture both audio & video, any offset issues between the captured audio & video sync?

Thanks!

Dave
Click to expand...

Oh right, this is an important topic, depending on the frame rate, the pix should be set to a 2-3 frames delay, correct?
 
I think Eddavid has it right... there is so much more to a camera than the numbers it produces on a bench... there is it's "feel".

Panavision and Sony spent a huge amount of time making a camera that replicated 35mm film - the gold standard - and it does this in many ways - both good and bad. Right now, I wouldn't have anything else.
 
Agreed, but without numbers and test data, there's no way to explain why one camera looks different/better than another without some science.
Otherwise, cameras would still be steeped in myth and folklore like violin-making. What does a Stradivarius sound better than a Magginni? ;)
 
You must log in or register to reply here.
Back
Top