4K upgrade

[David Heath]

Originally Posted by Matt Davis

In the grand scheme of things, I'd have preferred access to the 12 bit raw part of the FS700 more than the 4K resolution, but the subtle art of the Marketing Department is at work here - after all, if the FS700 could do even 10 bit out, what would that to to the F3?

But RAW implies what it says - the "raw" data from the chip, before any processing. It's a 12 bit word corresponding to the output value of each photosite.

So since the sensor in the FS700 is 4K, RAW from it is going to be 4K by default.

[Chris Medico]

Originally Posted by Robert Wall

Is this new information (about the R5 being required for the output of the 3G SDI to be anything that a recorder (either sony or 3rd party) can understand?)?

I expect 3rd party recording solutions to be available once the data format has been released.

[Alister Chapman]

Don't forget this is a bayer sensor (possibly the same sensor is going in the F5 and F55, with the F55 getting an improved filter array). So, if you want HD resolution images from the camera you need at least 2.5K raw. HD raw will not give HD resolution.

I have it on good authority that as well as the R5 there will be 3 other 3rd party recorder options for the FS700, but that there may be some differences in the way the signals are processed and recorded on the 3rd party recorders compared to the R5. My guess would be the 3 recorders being the Gemini, KiPro Quad and Cinedeck.

The F5/F55 will be 16 bit, while the FS700 will be 12 bit. But for most applications 12 bit will be good enough. Certainly is an interesting proposition compared to the F5.

[David Heath]

Originally Posted by Alister Chapman

Don't forget this is a bayer sensor (possibly the same sensor is going in the F5 and F55, with the F55 getting an improved filter array).

Don't forget the global v rolling shutter differences between the F55 and F5, which seems to point to far more fundamental differences between their sensors than the filter array - even though the dimensions, photosite counts etc are the same.

I'd say it's highly likely the FS700 and F5 share the same sensor, and the F55 has a derivative of it. The global shutter will mean a driver for each photosite with the F55 sensor.

I'm also not sure that it's strictly accurate to say the F55 filter array is "improved" over the F5 - more a case of "optimised for digital cinema as opposed to standard broadcast." As Mitch Gross says on the link I posted in another thread ( Sony’s PMW-F5 and F55: Defining CFA | CineTechnica ):

Why not make the CFA on the F55 standard for all sensors? The two cameras are for different uses and are aimed at different markets, but this is not a matter of Sony “crippling” a less expensive model camera but instead better defining it. Too much Color Space from a sensor is not a good thing. It means that the camera’s data stream is wasting processing and information space on colors that will never be reproduced. The primary use for the F5 camera is for broadcast work to be screened on LCD and OLED monitors. If the sensor’s CFA is more closely aligned to the Color Gamut of these output platforms, then the energies of the camera’s processing and codecs go towards accurately reproducing the image for the intended medium.

[Alister Chapman]

Lot's of speculation over the sensor/sensors. The published sensor specs for the FS700 and F5 are identical in every respect. I would be surprised if they were different. It's also conceivable that the global shutter function is simply disabled on the F5/FS700, although that does seem unlikely, so the F55 sensor probably is different.

To get the wider colour gamut of the F55 the CFA has to provide less out of band leakage while providing a broader pass band for each colour. This comes through the use of better colour filters. As the F55 gamut fully encompasses the F5 gamut, I think improvement is an appropriate term. Once you have a sensor capable of capturing a wider gamut it's relatively easy to then decrease or restrict that gamut through the DSP. After all the colour vectors are the same for both the F5 and F55, the only difference is that those vectors have a larger space on the F55. The colour response is not shifted or skewed, just broader. The F55 will have a broader colour palette than the F5 but the F5's colours will all be within the F55's pallet. When you connect the F5 or F55 to a conventional monitor the broader gamut of the F55 will simply be truncated, as a result they will most likely look almost exactly the same.

As when your after the very widest gamut you'll most likely be using the 16 bit raw recordings of the F5/F55, I don't think processing overheads or codec usage has any bearing on which colour space you use. Your simply taking raw pixel data. The better that data is, the better your final output, no matter what colour space your working in. I think this is less aligning cameras to different markets and more simply using a much cheaper CFA to make a cheaper camera possible, saving the more expensive CFA for the higher quality model.

[David Heath]

Originally Posted by Alister Chapman

Lot's of speculation over the sensor/sensors. The published sensor specs for the FS700 and F5 are identical in every respect. I would be surprised if they were different. It's also conceivable that the global shutter function is simply disabled on the F5/FS700, although that does seem unlikely, so the F55 sensor probably is different.

It may be conceivable that the F5/55 share the same sensor with global shutter disabled on the F5, but that should give them both equivalent sensitivity. (The extra circuitry for the global enabling drivers will still be present, and still blocking light from the photosites.) According to what Mitch writes, the F5 has better inherent sensitivity - which argues that on the F5 chip the circuitry for global shutter is absent. No global shutter - but better sensitivity. But highly likely there is commonality of design etc between the two.

As far as the CFA goes, then it may be largely down to semantics and definition of "improved". It's likely that such filters are more difficult to define accurately, hence more expensive, yet give no benefit for such as broadcast display. So a bit like paying extra for an uprated car engine, which then has to be detuned to the lower level for a specific use. What's the point? You may as well just use the cheaper engine.

When Mitch talks about the cameras data stream being wasted in processing colours that will never be produced, I think what he means is that if the system is capable of defining a finite range of values, better to use the entire range to define the wanted range - and hence lessen the step change between values.

It's also conceivable that more accurately defining the bandpass characteristics may even adversely affect the total transmitted light - hence impact slightly on sensitivity - but I'll confess I'm not sure about that. So it may be true that the wider gamut CFA of the F55 is indeed an "improvement" for a camera primarily designed for high end projected output, but at best an unneccessary expense, even a disadvantage, for a camera more intended for the broadcast market. I think it would be quite wrong for many F5 buyers to feel they are missing out by not getting the F55/65 CFA - the F5 may actually be better suited to their needs.

[Alister Chapman]

Well you'll probably be able to find out for yourself in Feb.

I'm not suggesting the F5 (or FS700) is using a sub standard CFA. Rather that the F55 is using a premium CFA and some of the guys at Sony have already indicated that the F55 CFA is very expensive to produce. As you say the CFA.on the F5 is more than adequate for most applications.

[Steven Blatter]

I think it would be great if the 4k upgrade on the FS700 would also record compressed 4K on-board like the Canon EOS-1D:

"The Canon EOS-1D is able to capture 4K (4096 x 2160), 24p (23.976) content directly to its on-board CF memory cards. It does this by using an 8-bit 4:2:2, high bit rate Motion JPEG compression scheme--allowing 4K content to be acquired in virtually any shooting environment, without an off-board external recorder."

Steven
Progressive Pictures - YouTube

[Jack Zhang]

That's not possible. The AVCHD encoder chip in the FS700 does not have enough power to encode even H.264 4K. The F5 and F55 are XAVC ready, and those would be able to record 300Mbps 4K in AVC.

[Sergiu Macarescu]

If I understand correctly the AXS-R5 recorder doesn't have an SDI port at all. That's why it needs a conversion unit. But the AJA Ki Pro Quad has an SDI input and can record raw 4k from the c500 with real time debayering for previews. I've read somewhere that Aja has announced the future support of the fs700. With any luck all we'll need will be the Ki Pro Quad(4k$) + several thubderbolt ssd(200$ each).