Panasonic Chip De-Mystified?

[David S.]

Foolish question by me then:

If the picture quality is acceptable to you or me or someone else, why are we discussing it?

To me, it seems analogous to what is in Coca Cola or Pepsi or cheetos.

I’ve seen some pretty striking footage. Whether the chip is 1080i native seems academic if it produces footage like I’ve seen and at its price point.

MTC

David S.

[George]

because it is a primary performance indicator and measure of an expensive product.
because every pro camera on the market provides this data to evaluate their product.
because panasonic provides this as a major technical measure on every other pro camera product.

[David S.]

George, without getting into a battle here, what do you say to my point that if the camera produces acceptable footage, who cares?

Do you think specs are more important that actual footage?

Again, I’m not defending anybody, but it’s the quality of what the footage is that is crucial for my work.

Help me out.

David S.

[Gunleik Groven]

All good, BUT

you sort of state stuff without giving your sources. If it gives 720p “true” resolution, I’d be _VERY_ happy. -;)

If it could shoot 720p 50 in PAL-land and there would be a FCP update to allow this format to be edited in some reasonable way, I’d love it.

So where did you hear this?

I wouldn’t mind 1080p 50 either as some native format, but I sorta don not expect this today from a camera which ha a 1/3″ CCD and costs a lot less than just a basic “HD-lens” for the VariCam.

Maybe one day, but I cannot see it just soon.

So:
Who are your sources?
Panny, on an off the record secret phonetap?
(BTW: I sorta agree that it’s not well thought through _not_ to release the data, simply because it generates fuzz around a product that otherwise seems very well done and priced. I don’t mind, though.)

Gunleik

[Gunleik Groven]

eh… and one more thing.

To me the reasoning is sorta like this:

1. Does the phootage look good?
2. What is the compression rate?
3. Does the phootage look good?
4. Is it a sturdy little cam? (The 150/170 are notorious for their ability to allow physical abuse and still work well…)
5. Does the phootage look good?

I know that on tech-sheets a recent Sony Cybershot grossly outperforms the Nikon D-100.
I also know pretty well which camera I’d prefer to use.

Gunleik

[George]

I think the technical data is important. If theres a project where I need to deliver 720p/60 then I can be pretty comfortable with using varicam or hvx (again knowing what the native resolution is) . Kind of like using a video monitor thats 200 lines of resolution to color correct and edit 1080psf. You would not do it. You would likely demand an lcd with over 800 lines or a high end BVM 900-1000 lines of resolution for the job.

If I have a film out project with lots of compositing and blue screen I’d like a camera with 1080p or higher CMOS (dalsa, viper etc) for the job and to ensure I had enough resolution, probably want to make sure the color space spec for the camera was also at least 4:2:2 or 4:4:4.

If the project is a 1080i/60 broadcast, generally the tool of choice may be a camera with 1080 x 1440 CCD like an ENG type camera. again confident that the right tool for the job is being selected.

For a direct to SD-DVD project. Yeah I know to use something like the SDX-800 with 2/3 inch chips and at least DVCPRO 50 codec (520k pixels).

If none of the vendors tell me what the camera specs are then I probably need to test every camera in my budget to see which one works in the above scenarios. a huge waste of time and money.

From what I know now. Its safe to say that the HVX-200 is not a good choice for any project requiring a 1080i/p deliverable. simply because it cannot natively capture at the target resolution. I’d get better results from the Sony FX/Z1 cameras for that type of work. But for fast action sports or stuff like that, sure the HVX would be a better choice as long as the deliverable is not 1080.

Pixel count does not tell the whole story either. Both the varicam and hvx are comparable from a pixel count. but the 2/3 ccd on the varicam and better glass translate into an image without chromatic abberation and without the noise introduced by the 1/3 ccd.

[Karl Holt]

George

“Well I heard today that the camera chip on the HVX 200 is not 1080 native. So that marketing
spoof of 1080 scanning is misleading at best”

If you search this forum you’ll see that Panasonic have not hidden this fact. There have been discussions about the HVX and pixel shift for months and months. HVX developer herself Jan adding to the discussion many times. True, they will not officially release the pixel count of the CCD’s. For good or bad – but this isnt a recent development.

In fairness to Panasonic they have stated that 1080p 1/3 sensor is too noisy and will also decrease the sensitivity and lattitude of the camera. Pixel shift is a technology that allows real world pixels to be generated – they are not interpolated or scaled up. Pixel shift works in gaining extra resolution. Search the forums for a more technical explaination of how it works. If you were taking a 4:4:4 output from the cam then pixel shift may affect the chroma res – but as the best DVCPROHD can do is 4:2:2, then pixel shift does not compromise the colour signal at all, and does gain more resolution.

As for the 1080p capture. Jan stated that the chip was not 1080p, but the image is captured at 1080p after pixel shift and DSP kick in. Then if you’re shooting SD or 720p the image is downsized from here. Maybe this is why there isnt much difference in 1080p to 720p mode. For one, trying to get light to fall onto a 1/3 sensor can only resolve so many lines – as can the lens. So from what I hear, its practically impossible to resolve 1080 lines on a 1/3 chip, even if the chip has 1080 lines. Its to do with the amount of light falling on that area and the lens. So again a 1080p chip at this size would be overkill. In addition to this, the DVCPROHD codec only stores 1280×1080 (not the full 1920×1080 raster) so at best they would need a chip they could pixel shift to get that amount of res and no more.

The initial reports are say 1080p is maybe 20-30% better than 720, but it certainly wont be the full deal.

Finally the camera does shoot true progressive but since DVCPROHD codec is used to store the information, you have to store the progressive information in an interlaced stream. This is just a way to make 24p fit into the DVCPROHD codec, its not any electronic processing altering the image at all. If you shoot 720p onto P2 you can store the native progressive frames (720p native mode, i think) this just results in a smaller file size as it only stores the recorded frame while the other mode flags the active frames for use (storing redundant frames also).

By the way Im not saying the HVX is great or that its a good that they are hiding the CCD specs, I’m just saying that it has been discussed for a long time and that practically every other piece of info about the camera Jan has been willing to discuss on here.

I dont expect this camera to perform like a varicam, just a HD version of my DVX which seems to be looking promising.

[Toke]

[Karl] “The initial reports are say 1080p is maybe 20-30% better than 720”

Can you point us, where has been resolution comparison between 1080 and 720 with HVX?

[Mike Schrengohst]

“To me, it seems analogous to what is in Coca Cola or Pepsi or cheetos.”

I do some work for Frito-Lay and I KNOW what is in Cheetos,
but I still eat them.

[Karl Holt]

toke

over at dvxuser.com there is currently extensive testing going on with the cams. Barry green and 5 other people are doing a 4 way shoot out between the HVX,Canon, JVC and Sony HD cams.

The “20-30%” was from a question I asked barry green after he worked with a pre-production model. So of course, you cant judge that conclusively. However Barry did say that it simply could not record the full 1080p, and it ‘looked’ slightly higher in 1080p mode, but no scientific test until this week.