Sony launches PXW-Z100 4K Handheld XDCAM

[Cliff Totten]

I would strongly suspect that Sony would cripple the HDMI on the AX1. It's important for them create some distance between the AX1 and the Z100 so I strongly suspect that it's 8 bit 4:2:0. (maybe 8 bit 4:2:2 if we are very lucky)

Although, I suppose it doesn't make a significant difference being that the video comes from this tiny single Bayer sensor. Heck, I suspect that even XAVC Long GOP 150Mbp/s, 8bit 4:2:0 already captures 99.99% of all this little sensor has to offer. Putting a 10bit 4:2:2 600 Mbp/s intraframe codec behind a sensor like this is overkill. It's like building a high tech Formula 1 race car but giving it a 1.6 liter Toyota Corolla engine.

Anyway, I was excited about this AX1/Z100 a week ago but that emotion is almost completely gone now.

My first 4K camera will be Sony's new super 35 "FS200" with XAVC-S Long GOP onboard. That's when I jump in the 4K pool. ;-)

I would give $100 right now to see Sony's three year 4k camera R&D plans. Whatever it is, I suspect that in 18 months, the Z100 will be considered the "bottom-end" of their 4K product fleet.

NAB 2014! I'm there!

[Meng Li]

One more question: in the future version of 4K cameras, like Z200 or further, for sure it would have interchangeable lenses. It is a nice benefit yet a questionable thing. The current Fujinon lenses are designed for 1080p,can it be fit for a 4K resolution? Like the one on PMW-300 may not be fit onto PXW-Z300 later due to the quality and purpose, etc. Another question is, is there any way to put Canon DSLR L lenses onto PMW-300 and how would be the quality of that?

[Jack Zhang]

Originally Posted by Cliff Totten

My first 4K camera will be Sony's new super 35 "FS200" with XAVC-S Long GOP onboard. That's when I jump in the 4K pool. ;-)!

I would have wanted that to be the FS700R, which would have recorded XAVC-S to SDXC UHS-II cards if there was more R&D time dedicated to it. You've got to remember, the FS200 may not have ND filters when and if that comes out.

[Ron Evans]

I expect the next XAVCS announcements to be a version of the EA50 or the VG30( or both since they are really the same though the extra space in the EA50 may be needed for the fan). Consumer /prosumer small sensor and larger sensor with XAVCS with PXW-Z100 and FS700R for the pros with XAVC.

Ron Evans

[Jack Zhang]

FS700R's already been announced. there's no XAVC support. The only codec I would see Sony putting on the next upgrade to the FS700 is XAVC-S. If standard XAVC is put on, that starts to cannibalize the F5 market. Plus, XAVC requires the UDF filesystem, something SD cards and Memory Sticks don't support.

The 700R came out too fast. With faster card technologies other than XQD like UHS-II and adding USB 3 to the FMUs, there is ample opportunity to add onboard 4K recording, yet they released it to early to add these technologies. Remember that it is a single slot camera, and the least physical change required is usually the best route, so UHS-II SD cards sound about right.

[Ron Evans]

XAVC isn't a recording option on the FS700R but Sony does provide a 4K solution with the IFR5. Difference is 12 bit compared to 16bit for the F5/F55 so there is a clear difference. It will be interesting as to where Sony creates the differentiators between models. It may well be that the average person will not see a difference in any of the outputs at the first generation stage anyway !!

Ron Evans

[Mark Fry]

Originally Posted by Ron Evans

I think the PXW-Z100 is more capable than either the PMW 100 or the PMW 160 just shooting 1920x1080 and forgetting about 4K. Just view 4K as a bonus with the present firmware.

I've seen this sort of thing repeated several times in this thread. Is Ron just repeating himself or is this a generally accepted expectation from the new 4k cameras? My recollection of the early HD cameras, particularly the Z1/FX1, was that the DV images they produced were not comparable to true DV cameras, since the in-camera down-scaling was pretty crude. (*) This has also been true of 1080-line images produced by DSLRs down-converting from native 3000 or 4000-line chips.

I can see it is handy to have this backward compatibility. However, unless someone has good information to the contrary, I would not expect a 4k camera to be as good at producing HD images as an HD camera.


(*) IIRC, there was one advantage to using the Z1 for DV - it was native 16:9 which could be an advantage over a PD150 and it's funny anamorphic lens.

[Matt Davis]

Originally Posted by Mark Fry

My recollection of the early HD cameras, particularly the Z1/FX1, was that the DV images they produced were not comparable to true DV cameras, since the in-camera down-scaling was pretty crude.

The in-camera downscale was a bit crude, but suck in HDV and scrunch it down using a good quality scaler, little bit of sharpening to dress it up after, and yes - it was very nice. It was the long-standing joke - how to make a nice SD DVD - start with HDV.

Of course, crude down-scales, and leaving the detail wound up, would guarantee a scratchy aliased picture. But that, and black stretch, and careful use of longer focal lengths and it would punch above its weight. I could put my pictures (HDV downsampled) up alongside a DSR570 and not look embarrassing.

[Mark Fry]

Originally Posted by Matt Davis

The in-camera downscale was a bit crude, but suck in HDV and scrunch it down using a good quality scaler, little bit of sharpening to dress it up after, and yes - it was very nice. It was the long-standing joke - how to make a nice SD DVD - start with HDV.

Of course, crude down-scales, and leaving the detail wound up, would guarantee a scratchy aliased picture. But that, and black stretch, and careful use of longer focal lengths and it would punch above its weight. I could put my pictures (HDV downsampled) up alongside a DSR570 and not look embarrassing.

Oh, I agree. Shoot and edit in HD then down-scale carefully with good software right at the end of the process, and you end up with really nice 16:9 DVDs. But don't expect a camera tuned to give good HD to also record good DV direct to tape/disc/card/whatever.

By analogy, I'd be very surprised if the AX1 or Z100 will record HD images that match an HD camera of the same size/price - which I think is what Ron Evans is expecting. On the other hand, if you can afford to shoot and edit 4k, and down-convert to 1080 lines right at the end, it should make for a pretty good Blu-ray production. In fact, until a decent 4k delivery tool comes along, that's probaly what most people will have to do.

[Ron Evans]

Yes I have repeated myself on different threads---sorry. As to downscaling. The in camera downscaling has never been very good. I used to rent DVC200 before I got my FX1 and never rented after that as the FX1 was better in my experience. The FX1 setup to shoot DV rather than HDV was also good. It was just the output downscale that was not very good.

I never use in camera scaling always shoot in the HD format and then downscale using TMPGenc after editing in HD project. I expect the same from the 4K cameras and will see in a few months.

Sensor technology for consumers is continually driven by costs and size. For the professional the goal is performance. Every so often the developments in consumer tech overtake the pro market and I think this is now what is happening and drives the pro market to even more extremes to get the advantage. The net effect is that the low and mid pro lines are vulnerable to cost and performance when compared to the consumer products.

I am disappointed that these new 4K cameras do not have touch spot focus. Focus is difficult enough for HD now and with 4K will be even more difficult which has been mentioned already. Autofoucs is an answer but I for one would like to identify the point in the frame I want in focus. This is pretty standard now on all consumer cameras yet sadly missing from cameras that would really benefit from the feature. Focusing is much more difficult on my NX5U than the NX30U for example.

Ron Evans

[Jack Zhang]

Do remember the sensor tech is different this time around with it being a single sensor rather than 3, if you're comparing the Z1 downscaling workflow to the Z100 downscaling workflow.

I fear because the pixels are so small, too much heat or unintentional laser beam contact would easily cause some pixels to die easily.

[Ron Evans]

These 4K cameras are using Sony sensor already used to shoot HD if it is the one in my HX30V which shoots great HD video as good as my CX700 or NX30U. My main interest though is to shoot 4K and crop a HD image from it just like I did with the FX1 to get a SD image. It worked very well for me then and I fully expect it to work that way again. We will all have to wait and see but I suspect they will also shoot very nice HD. Will it be as good as a PMW200 or PMW300 ? I don't think so as they will have a sensitivity advantage. Will the PXW-Z100 be as good as PMW100 or PMW160 I think it will. We will have to wait and see.

Ron Evans

[Roshdi Alkadri]

There's always a catch to these "prosumer" cameras as they call them. They won't put up a $4500 camera to compete with their higher end cameras (obviously). I think the 4k will be "Good" but where it might excel is downsampling to HD for final output.

[Piotr Wozniacki]

Sorry for the naive question, but what's the technical difference between "downsampling" (which everyone says will be great from 4k to HD), and "downrezzing" (which I know for a fact is never quite good)?

On a similar note: when the FS700 first came out, everyone was raving how its higher pixel count enables downsampling to a HD picture actually better than that from an FS100.... Why?

[David Heath]

Originally Posted by Piotr Wozniacki

Sorry for the naive question, but what's the technical difference between "downsampling" (which everyone says will be great from 4k to HD), and "downrezzing" (which I know for a fact is never quite good)?

Not sure there is an accurately defined exact meaning, but "downrezzing" is normally taken to mean forming a high res signal from the chip and down resolving that, whereas "downsampling" means such as taking a group of photosites together and forming the lower resolution signal directly from that.

How good/bad either will be depends on circumstance. Downsampling works well in the case of something like the Canon C300 - but that's because the chip dimensions are an exact multiple of the desired output (1920x1080). Blocks of 2x2 photosites can be taken together to form pixels on a one to one basis.

Apply the same technique in a DSLR, and it's far more difficult. (And not as good.)

On a similar note: when the FS700 first came out, everyone was raving how its higher pixel count enables downsampling to a HD picture actually better than that from an FS100.... Why?

Well, I've always thought there was a bit of a mystery here, and if you look at charts, the actual performance of the two is quite similar. Best I've seen are from Adam Wilt - Review: Sony NEX-FS100 “Super35” LSS AVCHD Camcorder by Adam Wilt for the FS100, and High Speed and Low Light with the NEX-FS700 by Adam Wilt for the FS700 . Note especially the vert/hor resolution imbalances in each case.

In fact, examine them very closely and the similarities between them are far more than any similarity either has with the F3 chart that Adam also prints - in spite of the general view that the F3/FS100 share the same chip. My own view is that the FS100/FS700 have the same (3840x2160) chip - Sony were trying it out in the FS100, but didn't want to let the characteristics be known because the obvious next question would be "so when's the 4k version coming then!?"