Convergent Design Nano Flash & Z7 ?

[Dan Keaton]

Dear David,

Yes, I did not mean to imply that 1440 x 1080 was equal to 1920 x 1080.

I just wanted to explain the all HD-SDI, in 1080 mode is always 1920 x 1080, but sometimes created from a sensor that is only 1440 x 1080.

[Stephen Armour]

Truth is, shooting outdoors in good light...with original footage, the average person would probably be very hard pressed to see much difference in output, with or without the Nano.

With most of these cams, it seems the issues are more with heavy post work or poor lighting and/or fast movement. CF helps with some things, but can't do the impossible.

DOF and shadow detail are often what cause most people to note "something different" with our final output, so unless we're using 2/3" sensors with some good capture codec, it just doesn't seem things are a whole lot different.

Now, if only we could afford a couple of those PMW-350's...

[Alexander Timanov]

That's what I was talking about. Stephen, you've cought me well.

Alex

[Michael Galvan]

Originally Posted by Uli Mors

Its simple.

The Clearvid Cmos uses a non Full HD grid , RGB shifted to each other (pixelshift), interpolated and recorded to 1440x1080.
This grid again is stretched when output via HDMI or HD-SDI or HDanalog out (like HDCAM does too).

The bargain of the nanoflash is that you dont record the 1440x1080 with only 25mbit/s but the NONCOMPRESSED signal (pixelshifted CCD -> 1080) with much higher bitrate, leaving compression artefacts behind.

I am sure the results are better than the HDV recording, but you dont get higher resolution that what is generally capable by the sensor.

BTW: Clearvid is a great technology. Its not full HD but each pixel is bigger and therefore 1/3" can be very lightsensitive (-> Z5, Z7).

ULI

Interestingly, I was reading Alan Roberts white paper on the Canon XH G1 for the BBC. He measured the resolution of the HD-SDI image at around 1015 lines vertically. It is slightly lower at around 940 in the F modes). He mentions that the sensor chipset of the camera clearly delivers more resolution than its pixel count due to its spatial offset. This resolution is not recorded fully onto the HDV tapes, but can be recorded from the HD-SDI using a high-quality full raster codec. I can confirm this from my own recordings from my Canon XL H1S's HD-SDI recorded to ProRes HQ using my IO HD. There definitely is a resolution increase with this camera.

I think the Canon XL H1/G1 cameras are the only sub "full raster" HD cameras capable of showing a resolution increase when recording the HD-SDI to a full raster codec. Because the native pixel count of the camera matches 1:1 with HDV, the increased resolution from the spatial offset is not seen in the HDV tape recording. It is definitely noticeable in the HD-SDI recording. But most other HDV cameras, as stated after going through their spatial offset process, create enough resolution to fill or closely fill the 1440x1080 resolution of HDV.

But of course, there are many more advantages to recording the HD-SDI, so a Nanoflash would be a great upgrade for these Sony cameras!

[Uli Mors]

Of course, pixel shift has a lot of disadvantages too.
For example, when you shoot a live concert and the light is not setup for videorecordings, it can happen that the light engineer jumps into red (full red lit stage), bad enough but:

In that case you clearly get lost resolution - since the red cmos chip has not full HD res.

2nd you get more aliasing and resolution/sampling artefacts, that may be not that important for daily shooting, but its there and will have influence on your pictures.

Anyway: If a Z5/Z7 meets your expectations (like it does for me for certain projects), a nanoflash will even improve it. Simple.

ULi

[David Heath]

Originally Posted by Michael Galvan

Interestingly, I was reading Alan Roberts white paper on the Canon XH G1 for the BBC. He measured the resolution of the HD-SDI image at around 1015 lines vertically. ........He mentions that the sensor chipset of the camera clearly delivers more resolution than its pixel count due to its spatial offset.

Yes, but it's important to note that the XH G1 has sensors of dimensions 1440x1080 (so approx 1.5 megapixel) and unlike the Panasonic cameras pixel shifting is used in the horizontal axis *ONLY*. (The Panasonics use both H & V.) Consequently, the vertical figure you quote above is due purely to 1080 photosites vertically.

Horizontal pixel shift will improve the luminance resolution above the 1440 that the sensor figures may suggest. The downside is that pixel shift comes with side effects, as Uli says. As Alan says in the report you mention:

However, and as expected, this process also generates some horizontal aliasing, and it is at about 1600 pixels that the wanted and aliased frequencies are equal in amplitude.

Also note that the Sony sensors *DO NOT* employ any pixel shifting. Corresponding R,G,B photosites are all optically in the same place. There is no right and wrong to either method, they just emphasise different compromises.

But of course, there are many more advantages to recording the HD-SDI, so a Nanoflash would be a great upgrade for these Sony cameras!

It would be an upgrade, but I'd still maintain that for the combined cost of this camera and nanoFlash, the money would be better spent on buying a better (more expensive) camera in the first place.

It's a bit like buying a cheap car and spending a fortune on upgrading the performance. You may be far better off just spending the same money on a better spec car in the first place.

[Stephen Armour]

Now if we could afford the PWM-350 AND a Nanoflash...that would be something! (Sorry, just dreaming...)

[Alexander Timanov]

It would be an upgrade, but I'd still maintain that for the combined cost of this camera and nanoFlash, the money would be better spent on buying a better (more expensive) camera in the first place.It's a bit like buying a cheap car and spending a fortune on upgrading the performance. You may be far better off just spending the same money on a better spec car in the first place.

I agree, but sometimes a better camera lead to the same fortune to be spent. For example EX3 -pair of SxS pro cards (or even SxS1), spare large capacity battery and last but not the least a nanoflash. I would keep silent on P2 cards cost. Of course it is better to have more advanced camera, but it may happen, just like Stephen says: Now if we could afford the PWM-350 AND a Nanoflash...that would be something! (Sorry, just dreaming...)

Regards,
Alex