New Sony XDCAM - the PMW-100

[Dave Blackhurst]

Originally Posted by Chris Barcellos

Yeap, they do. Even my lowly VG20 shoots in 1920x1080 at 60p.

Not to mention even some of the Sony PS cams do... TX100v/200, HX100v/200v, HX9 & whatever the replacements for that one are...

Actually 1080/60p is pretty impressive from the TX100 considering it's just a point and shoot full auto proposition!

Pretty much any higher end new release Sony has 60p, the new Alphas also have 24p, along with 60p. I believe the NEX cameras do as well.

60p has it's attractions, there's a definite "crispness" about it over 60i. Definite "being there" feeling when played back on a device that can properly handle it, and you can stop and freeze a frame in time...

[Ron Evans]

Originally Posted by Alister Chapman

According to the literature it's a single sensor with 1920 x 1080 pixels which is very disappointing as this means bayer or clear-vid, either way the resolution is likely to be some way below that of an EX1 or EX3.

It says effective so no way of knowing if that is the real sensor or after interpolation. Same issue with the NX30 which says 6650k gross 6140k actual, whatever that means !!! That's about 3 times 1920x1080. Be interesting to see how these cameras compare. On the surface I think the NX30 with a Nanoflash may be a better deal !!!

Ron Evans

[Andrew Bower]

Originally Posted by Dave Blackhurst

Not to mention even some of the Sony PS cams do... TX100v/200, HX100v/200v, HX9 & whatever the replacements for that one are...

Actually 1080/60p is pretty impressive from the TX100 considering it's just a point and shoot full auto proposition!

Pretty much any higher end new release Sony has 60p, the new Alphas also have 24p, along with 60p. I believe the NEX cameras do as well.

60p has it's attractions, there's a definite "crispness" about it over 60i. Definite "being there" feeling when played back on a device that can properly handle it, and you can stop and freeze a frame in time...

60p is also better when trying to show the humidity's affect on truss rods in older neck-through basses...

HA, just kidding. Good to see your post Dave. Long time no speak/type.
I might need to talk with you regarding my bass again. I figure it needs to see the guru once every 10 years or so...
:)

...and while I am excited to see Sony coming out with new cameras. I was really hoping for an EX5 or an NX5 with bigger chips or an FS100 kit with power zoom and a zoom rocker handle like the FS700.
But I guess we still have a few more days...

[Alister Chapman]

Effective Pixels does not normally refer to post de-bayered pixel count. I'm quite sure that if the sensor had a larger pixel count Sony would be boasting about it, after all many people will buy based on getting the largest number of whatever it is.

Even if they are combining several smaller photo sites into a single pixel, the pixel count is still too low for full resolution HD with a single sensor. So as has been suggested by others it's not unreasonable to expect resolution similar to the Canon XF100/105 which also uses a single 1920x1080 sensor.
Effective or actual pixels normally refers to the pixels actually used for image capture. It's quite normal to have extra masked pixels around the edge of the light gathering part to measure sensor noise and voltage levels to asset with noise reduction and black balance. These are not normally included in the "effective" pixel count, but often included in the gross pixel count.

Of course resolution isn't everything. Noise, contrast, dynamic range etc all combine with resolution to produce the image and if all the others are very good, then slightly lower resolution may not be that big an issue. In addition there is still a massive amount of confusion over resolution and image sharpness.

The question raised by the use of a new sensor with 1920x1080 pixels is... where is the camera with 3 of them? A full 1920 x 1080 resolution 3 chip camera, maybe EX1 replacement?

[Gabor Heeres]

It looks a lot like the NX70's design. Why didn't Sony add the dust- and rainproof cover as with the NX70? I do still hope Sony comes with a dust- and rainproof EX1-like camera one day. If they than would even add 50 Mbps you have an ideal Camjo cam which will get popular among camera-journalists around the world.

[Ron Evans]

Originally Posted by Alister Chapman

The question raised by the use of a new sensor with 1920x1080 pixels is... where is the camera with 3 of them? A full 1920 x 1080 resolution 3 chip camera, maybe EX1 replacement?

Or maybe a NX5 replacement.

Ron Evans

[Glen Vandermolen]

Originally Posted by Ron Evans

Or maybe a NX5 replacement.

Ron Evans

If it records to SxS cards using the XDCAM 4:2 :2 codec, then it will be in the XDCAM family of cameras.
If it records to AVCHD, then it will be in the NXCAM family.

I think some of you are missing the big selling point of this camera, which is the XDCAM 50mbps, 4:2:2 codec. This is a broadcast quality codec, in a small camcorder body. I would say it's revolutionary, but it's not - Panasonic and Canon have already accomplished this feat. Why Sony is so late to the game is the big mystery. Even their vaunted F3 uses a 35mbps, 4:2:0 internal codec.
I'm hoping Sony will quickly incorporate this better codec into their remaining XDCAM line. There's simply no reason not to. Imagine the EX1/3 line with this codec. It would extend the life of these cameras for years.

[Ron Evans]

The competition is the XF105 so it will be interesting to see the price comparisons when it is available. Unfortunately there isn't a XF100 version though in the product line the NX30 will now compete with the Canon XA10. Though just a little more expensive. Will have to see what the street price is when it arrives.

Ron Evans

[Alister Chapman]

Put the very best codec in the world behind a sub standard sensor and you'll still get a sub standard image. To be fair to the PMW-100 we don't know how good that image is yet, but based on the specs I don't think it will meet broadcast specs. The minimum specs for BBC and Sky broadcast are 50 Mbps with a 1/2" sensor. The only exception to this rule at the moment is the XF305 and even though the sensors are only 1/3" at least they are full resolution. The XF105 is not approved for broadcast despite it's 50Mb/s codec.

I'd rather have a great sensor/lens with AVCHD than a poor sensor/lens with XDCAM 422.

[Glen Vandermolen]

Why can't we have a great sensor/lens AND a great XDCAM 4:2:2 codec? Must they be mutually exclusive?

The 3-1/3" CMOS HPX250 and HPX370 are also approved by the BBC - or so I've heard.
And we don't know the specs on the PMW-100's sensor, and no, a single 1/3" cam probably won't be approved by the BBC, regardless of the codec.

The point is, Sony CAN put a good codec into a small camcorder, something they haven't done yet. Why they chose to put it in this small camera and not the EX1/3 and F3 line is the big question. I'm guessing - hoping? - that they will soon, and this is the first of several new XDCAM, 50mbps, 4:2:2 cameras.

[Stephen Crye]

Allister, your pointing out that the specs state "effective picture elements" has really piqued my curiosity.

I had always assumed that the Exmor chips, even with the diagonal arrangement, still had at least 1920x1080 sensor sites, and used the diagonal thingie with the extra green sensor sites to do even better than a traditional 1920x1080 square array, such as what is used on the XF100 or the Panny AC160.

So, I have been looking closely at the specs of various Sony's. There seem to be two general categories -
1.] cams like the CX550V/MC50 and the NX70 and VG10/VG20 that have multi megapixel counts and take high-res still photos, but do some kind of down-sampling to produce the 1920x1080 video image. Uniformly on the web, and in my tests, these cams have much sharper video images than the NX5, and *seem* to look sharper than the XF100, but look identically sharp to the AC160 (which I have not tested)
2.] cams like the EX1 which use three 1/2" sensors and a splitter to direct color-specific light to the individual sensors. I must confess that the limited PWM-EX1 footage I have seen seems to have that smeary-look that I associate with the XF100 or the NX5. Perhaps I'm not looking at the right vids ...

From the official Sony specs for the EX1: "Effective picture elements 1920 (H) x 1080 (V)"

I then looked on Slashcam, and they list the pixel count for the EX1 as : " image sensor effective pixel count 16:9 2073600 Pixel ", which equals 1920 * 1080. On Slashcam, the NX5 does not have the same spec - that makes me suspicious, but in the "still" section of the review, it states: image sensor effective pixel count 16:9 1037000 Pixel" . I'm not sure where that number comes from! It seems close to 960 * 1080, which is 1036800.

For comparison, on the Panny site for the AC160, they say this: "Pick-up Device: 1/3 Type Progressive, 2.2Mil Pixels, 3MOS Number of Effective Pixels (HxV): 1920 x 1080"
That 2.2 Mil Pixels number appears for a lot of cameras. It is bigger than 1920 * 1080 . Not sure if there is an outer edge of pixels, or if they are just rounding?

I am confused on so many levels here. One of them has to do with the fact that a 3-chipper would not seem to need any special sensors to differentiate different colors - that was already done by the splitter.

Can you penetrate my confused haze and perhaps help me understand all of this? I'm trying to understand "De-Bayerd", what I'm finding on Wikipedia is not helping much ...

I did find this:
Sony Global - Technology - CMOS Sensor "Exmor"

Thanks!

Steve

[Ron Evans]

The single sensor small Sony's have this sort of sensor Sony Global - Technology - ClearVid CMOS Sensor and HowStuffWorks Videos "Sony ClearVid CMOS Sensor Technology for Handycam Camcorders". Cannot find now any info on the NX5 but these from the Z7 I think are the same or at least they are the same in concept. Digital HD Video Camera Recorder - HVR-Z7U/HVR-S270U. This at least shows the relationship between the sensors and the pixels which are interpolated from the sensor outputs for each of the 3 chips. The sensors are twice the area of the interpolated pixels. 960 sensor sites on each row and alternating in the vertical alignment on each of the 1080 rows. As you can see from the diagram the interpolator creates a pixel between 4 surrounding sensors and uses the center of the sensor as a pixel too. Resulting in 1920x1080 interpolated pixels. The goal being low light performance and noise reduction. For the single chip sensors this is not the case and with the "R" sensor has achieved both maybe. The video shows difference between Bayer and ClearVid array.

Now I have got used to setting up my NX5U I can make it look just as sharp as my CX700 or XR500 for instance. However in stock setup the NX5U does not look as good as the consumer cameras.

I am sure Alister has a better explanation and references.


Ron Evans

[Anthony Ching]

Originally Posted by Emmanuel Plakiotis

Any information on the price?

About 400,000 JPY

[Stephen Crye]

Originally Posted by Ron Evans

Now I have got used to setting up my NX5U I can make it look just as sharp as my CX700 or XR500 for instance. However in stock setup the NX5U does not look as good as the consumer cameras.

Ron Evans

Thanks, Ron. I had seen those Sony pages a while back but it was good to review them.

Aside from disabling Macro, what specific settings do you use on the NX5?

So ... if the EX1 uses the same Clear VId with diagonal pixels, why the difference in the SlashCam pixel count numbers? Wild speculation - perhaps Sony used the same sensor element size and orientation, but just put more of them on the 1/2 inch chips?

One thing I did noticed in my week of testing the NX5 - it had very good low light performance, much better than the CX550V. For the same gain and shutter, it was many stops brighter.

Steve

[Ron Evans]

Yes for the same iris and gain the NX5U is brighter than the CX700. However it shows lots of grain at 12db whereas the CX700 still has less at 21db so the end result is the CX700 has a more useable picture when it gets really dark and with LOW LIGHT setting on its even better.

As to settings I have macro and anti flicker OFF. I use a modified PP3 setting with ITU 709 gamma, master black -3, color level 2, detail +3. I have various settings around these main differences.

Ron Evans