Sony PXW-X70 announced: Pro XDCAM version of AX100

[David Heath]

Originally Posted by Lawrence Bansbach

It might be interesting if the the bottleneck were the SD cards and Sony planned to enable 4Kp60 via the HDMI 2.0 port.

Forgive me if I'm misunderstanding you, but it's worth repeating again that in principle **FOR A LONG_GOP CODEC** doubling the frame rate does *NOT* mean doubling the bit rate (all else equal).

Halve the time difference between successive frames and expect less difference between them - which any coder can exploit. Typically, the TIME between I-frames may remain the same (typically 0.5sec), so the no of frames between I-frames will be 30 for 60p, 15 for 30p. Same total number of I-frames.

And since the I-frames tend to be much larger than difference frames, that's why 60p need not be a much higher data rate than 30p.

[Ron Evans]

On my FDR-AX1 30P, 3840x2160 has option for 60Mbps or 100Mbps and 60P is fixed at 150Mbps., 3840x2160 . 1920x1080 XAVC-S is 50Mbps. I think that since the 1920x1080 422 10bit is quoted at 50Mbps I would expect the data rates to be similar to the AX1? Maybe just like the AX100 30P 60Mbps.

Ron Evans

[Noa Put]

Originally Posted by Jos Svendsen

Is it a constanf F2.8 or F2.8-4.0 like AX100?

Acc to the spec sheet it has the same lens which is F2.8-F4.5

[Alister Chapman]

Originally Posted by Lawrence Bansbach

That's too bad. UHS-II speed class 3 SDXC/SDHC cards have a minimum write speed of 30 MBps (240 Mbps).

So allowing for the normal 200% overhead needed for reliable video that's a real world limit of about 120Mb/s unless you start specifying which cards can and can't be used. You only have to look at the issues GH4 users are having when they go above 100Mb/s to see that there is a big difference between what a card may claim to be able to do and how it behaves in the real world.

You need more than 3G SDi for uncompressed 4K. The FS700 manages to squeeze 4K over it's single SDI by using compressed raw. You can pass 4K 60p over HDMI 2.0.

The lens appears to be the same lens as the AX100. It's f2.8 to f4, 9.3 to 111mm.

[Noa Put]

Eventhough the specs say it's a f2.8 to 4.5 I just checked on my ax100 and acc to the info on screen it stays at f4.0 when fully zoomed in.

[Lawrence Bansbach]

Originally Posted by David Heath

Forgive me if I'm misunderstanding you, but it's worth repeating again that in principle **FOR A LONG_GOP CODEC** doubling the frame rate does *NOT* mean doubling the bit rate (all else equal).

Halve the time difference between successive frames and expect less difference between them - which any coder can exploit. Typically, the TIME between I-frames may remain the same (typically 0.5sec), so the no of frames between I-frames will be 30 for 60p, 15 for 30p. Same total number of I-frames.

And since the I-frames tend to be much larger than difference frames, that's why 60p need not be a much higher data rate than 30p.

Even if going from 30p to 60p doesn't increase the data rate much, going from 8- to 10-bit and 4:2:0 to 4:2:2 will increase it. The AX100's 4Kp30 data rate is 60 Mbps; anything more is likely to push it higher. Or if we extrapolate from the X70's 50-Mbps for HD, 4K is likely to increase it. If the 4K upgrade isn't XAVC-S but XAVC with a data rate closer to (or greater than) 100 Mbps, then it might outstrip an SD card's ability to record it reliably -- an issue Alister mentions above -- and require an external recorder.

[Ron Evans]

Maybe the 4K upgrade is not just software but hardware upgrade to XQD card slots etc !!! Justifying the cost.

Ron Evans

[Lawrence Bansbach]

Originally Posted by Alister Chapman

You need more than 3G SDi for uncompressed 4K. The FS700 manages to squeeze 4K over it's single SDI by using compressed raw. You can pass 4K 60p over HDMI 2.0.

Yes, of course. I had a brain spasm.

[Al Bergstein]

Still looking for low light performance on this. Stuff done with daylight seems just like all the rest done with daylight. The stuff is fine. But want to see low light performance.

[Josh Bass]

anyone know how this compares to an ex1 or ex1r? that was always my "when i can afford it" cam (used obviously).

[Andy Wilkinson]

Josh,

Alister mentions that it's not got the dynamic range of the PMW-300 (successor to the EX3 and EX1r type cams) in his initial review (he estimated at 10-11 stops area) and that it also suffers a bit from highlight clipping.

So in HD mode I would EXPECT it to be not as good - especially bearing in mind the much lower price point - but then because of that 4K ready sensor it may well be as sharp - or even sharper. My C100 with its single, big "4K sensor" is certainly very sharp.

So we're back to the "is dynamic range more important than sharpness" type debate. Mind you, it's current single sensor technology versus a 6-7 year old 3 sensor one. I'd always want more dynamic range - but that's just me.

I stress I'm an ex EX3 owner/current owner of the PMW-300 and C100 (and TM900 still going strong) - but have never seen/handled this PXW-X70 cam in the flesh so take this for what it's worth...

Pretty interested in it as a toe in the 4K world though - but doubt I'll buy it until 2015 (or something else as there will be other, perhaps better options by then).

[Ken Ross]

Originally Posted by Al Bergstein

Still looking for low light performance on this. Stuff done with daylight seems just like all the rest done with daylight. The stuff is fine. But want to see low light performance.

I'd be surprised if there's any significant differences between the AX100 low light and that of the X70.

[Mark Rosenzweig]

Originally Posted by Al Bergstein

Still looking for low light performance on this. Stuff done with daylight seems just like all the rest done with daylight. The stuff is fine. But want to see low light performance.

I agree with Ken that the low-light performance of the AX100 will be similar to that of the PXW-70 -same lens, sensor and processor.

Here is an AX100 4K low-light video:

Sony FDR AX100 4K Video in Dim Light: A Visit to Grand Central Terminal on Vimeo

Here is a 4K video taken at night comparing the AX100 and the GH4:

Panasonic GH4 and Sony FDR AX100 Low-Light Comparison in 4K 2: Streets at Night on Vimeo

[Mark Watson]

I see the price of the FDR-AX100 and HDR-CX900 dropping here in Japan. $1,580 and $1,350 in stores and less on-line. Yamada-denki is not known for low prices, yet they show 10% discount.

Mark

[Noa Put]

Originally Posted by Al Bergstein

Still looking for low light performance on this. Stuff done with daylight seems just like all the rest done with daylight. The stuff is fine. But want to see low light performance.

I did a non scientific but real world test yesterday at a wedding in a very dark candle lit only venue between a cx730, ax100, rx10 and gh4. I did not pixelpeep, just looked at the result from a normal viewing distance on a big screen.

I pushed each camera to it's max by selecting a shutter of 1/25th for all camera's, I had a f2.8 lens on the gh4 and did the test shot at 6400 iso, the rx10 was set at f2.8 and had a iso of 12800, the cx730 had a gain of 24db and the ax100 a gain of 30db

The difference between the cx730 and the ax100 was slightly in the advantage of the ax100 which had a bit brighter image but was a lot more detailed, the cx730 softens up a bit at max db while the ax100 retains much more detail

The gh4 was a bit brighter then the ax100 but appeared a bit less detailed.

The rx10 was the brightest of the bunch but also very noisy at 12800, at 8000 iso noise levels are acceptable and comparable with 6400 iso on the gh4