Cutting/removing the lens off a HVX ?

[Jim Martin]

Originally Posted by Dennis Hingsberg

Has anyone here tried or have any experience removing the stock lens from a HVX? Anyone gone that far in taking one apart that they could provide testament to it's ease or difficulty? Any manuals or such would be highly appreciated.

I'm looking at designing a 35mm mod for HVX cameras that will yield an overall brighter setup compared to traditional methods.

In theory the mod could provide an effective light "gain" of 2 stops when using 35mm setup.

This would effectively mean you could bet back to the cameras original working ASA. (ie. approx 400 for the HVX) Currently with most 35mm add-on adapters you end up with an approximate working ASA of 100, or between 50 and 60 once you throw a 35mm lens into the equation.

This is impractical for most film style lighting setups, particularly indoor.

If you don't understand the relationship between ASA, f stops and shutter speeds? Here's a really good read: http://www.uscoles.com/fstop.htm

Just so you know, the guys in Crooklyn have been doing it for years on the DVX and I think now on the HVX. You know the Crooklyn gang, Broadway Photo er Royal Camera er Prestige er Digital Liquidators er ETC. When you call them up about there below cost price and they'll tell you "thats the body only...it won't work without a lens but we can sell you the lens"....and a charger..and a battery...and a mic...and a mic holder...and a cable....and the cardboard box.....So I guess its no problem taking the lens off!

Jim Martin

p.s. Chris, hopefully you won't delete these Crooklyn names but if you have to, I understand

[Dennis Hingsberg]

Well I still think for the overall price, ability to shoot 4:2:2, workflow advantage of P2, increased light gain over any other 35mm adapter option on the market and no need to flip images in post or add on TFT/LCD's - the HVX lens & CCD modification would surpass many digital 35mm systems falling only short of RED which is nearly 5 times the price!

[Matt Burton]

I'm with you Dennis, if nothing else it would be nice to get rid of that dam digital focus at least.

[Jeff Kilgroe]

Originally Posted by Dennis Hingsberg

Well I still think for the overall price, ability to shoot 4:2:2, workflow advantage of P2, increased light gain over any other 35mm adapter option on the market and no need to flip images in post or add on TFT/LCD's - the HVX lens & CCD modification would surpass many digital 35mm systems falling only short of RED which is nearly 5 times the price!

I don't know about that... The XLH1 has superior resolution and interchangeable lenses as it is and after P2 cards are considered, is no more expensive. You can also use 35mm lenses with it using an adapter like the Brevis which attaches directly to the camera body. The only thing it lacks is true 24p (uses 24F mode due to the 1080i sensor). But I don't see that as a problem - at times the increased resolution is superior. The real advantage to the HVX is it's color reproduction. Color out of the HVX200 is still plain awesome. As for the 4:2:2 color space, I thought it would be superior for chromakeying, and it is, but with the HVX's pixel-shifting and lower resolution, the end results are no better vs. keying with 1080 24f from the XLH1. As for RED, it's true it's more expensive, but it has a lot more to offer and in some ways less too. It's a different camera, different market, different purpose. It's going to work a lot better for me, but won't be a good fit for many HVX200 users.

[Dennis Hingsberg]

I wouldn't say superior resolution. Both cameras use the same bag of tricks to achieve higher "effective" resolutions. Canon goes with horizontal pixel shift to achieve 1440 pixels horizontally but since it's an interlaced sensor offers 1080 pixels natively for vertical resolution. The HVX being a progressive sensor uses pixel shift to get 1440 pixels horizontally and scans 1080p vertically and converts it to 1080i.

A resolution chart in front of both cams may reveal slightly more sharpness with the XLH1 in 1080i modes but for 720p shooters likely no advantage - especially if you have to deinterlace in post or go with the frame mode of the Canon.

As for interchangeable lenses, is there still only one HD lens available for the XLH1?

Adding P2 cards to the HVX may not make the XLH1 more expensive, but if you want to record 4:2:2 you need to go external with the XLH1 and that's no cheap option either. Let's not forget the advantages of shooting to solid state memory either ie. no mechanical parts, no maintenance, quiet, use in harsh environments, drop it in your fish tank, etc..

Speaking of 4:2:2 it's more than just less jaggies for chroma-keying. DVCPRO HD uses intraframe compression which compared to HDV makes for a lot more ease during the editing and rendering process. It's exactly why programs like Cineform exist - so people can convert their HDV non frame independent compressed footage to something similar to DVCPRO HD (intraframe compression) and be happy editing.

I currently own Canon (XL2) so don't get me wrong - I'm not brand dedicated and could care less what brand I own or use. In the end it has to get the job done, be cost effective, offer some useful features... but at $8500 vs $5300 it's a no brainer for me.

Let's not even begin to started on other features of the HVX like 4 channel 16 bit uncompressed audio, variable frame rates from 12 to 60 fps, not front end heavy........ . ...

[Chris Hurd]

Originally Posted by Jeff Kilgroe

The XLH1... the only thing it lacks is true 24p (uses 24F mode due to the 1080i sensor).

The word "true" in "true 24P" is completely irrelevant anymore. 24F is 24P. 24F is no less "true" than 24P. Nevermind that the CCD block is 1080i... what matters is what's coming from the DSP, and the output is in fact progressive 1080p24 (or 1080p30 if desired). As proof that there's no difference between 24F and 24P, consider that 24F captures *as* 24P in Final Cut Pro, Premiere Pro and other editing applications. The software can't tell the difference, and neither can the eye. The 24F vs. 24P thing is a complete non-issue. For all practical purposes, they're identical.

Originally Posted by Dennis Hingsberg

Canon goes with horizontal pixel shift to achieve 1440 pixels horizontally

Incorrect. It's important to note that each CCD in the Canon XL H1 image sensor block is 1440 x 1080 already (horizontal pixel shift then provides a resolution boost up to 1920).

Originally Posted by Dennis Hingsberg

is there still only one HD lens available for the XLH1?

Currently there are two HD lenses available for the XLH1; the stock 20x HD lens and the wide angle 6x HD lens. Most H1 shooters who own the older 16x manual lens from the XL1 / XL2 line have reported it to be adequate for HD use (it is built to specs that are better than standard definition). And although limited to only certain types of shooting applications, any of the high-end Canon L-series EOS 35mm still photo lenses will provide HD resolution on the XL H1.

Hope this helps,

[Dennis Hingsberg]

Originally Posted by Chris Hurd

Incorrect. It's important to note that each CCD in the Canon XL H1 image sensor block is 1440 x 1080 already (horizontal pixel shift then provides a resolution boost up to 1920).

The specs on the XLH1 say "effective" 1.56 mega pixels in 16:9 HD mode. If 1440 x 1080 = 1,555,200 and Canon uses horizontal pixel shift the "true" horizontal resolution can only be less - can it not?

BTW Chris, I just love this site and all the work you've done here. I spend more and more time here everyday and soon will get fired from my job because of it. Cheers!

[Chris Hurd]

Thanks for your kind remarks, Dennis.

So-called "effective" pixel counts always refer to the actual number of pixels used on the CCD *before* the Pixel Shift process is applied. The differentiation between total pixels vs. effective pixels is due to the fact that there is almost always a certain number of unused pixels along the outer edges of a CCD sensor. The effective count is important because that's what's actually being used. Pixel Shift is never figured into that number though. For example, the specs for the original Canon XL1 state that it had 270,000 total pixels on each CCD, of which 250,000 were effective. Do the math and you'll find that the effective count was less than the native resolution for standard definition DV (345,600 pixels). A combination of horizontal and vertical Pixel Shift processes boosted the resolution of these less-than-SD chips up to 720 x 480 (and more importantly, nobody complained about it back then).

Likewise, each CCD on the Canon XL H1 has a total 1.67 megapixels, of which 1.56 megapixels are effective. They're "native" 1440 x 1080, before Pixel Shift is applied. Hope this helps,

[Jeff Kilgroe]

Originally Posted by Chris Hurd

The word "true" in "true 24P" is completely irrelevant anymore. 24F is 24P. 24F is no less "true" than 24P. Nevermind that the CCD block is 1080i... what matters is what's coming from the DSP, and the output is in fact progressive 1080p24 (or 1080p30 if desired). As proof that there's no difference between 24F and 24P, consider that 24F captures *as* 24P in Final Cut Pro, Premiere Pro and other editing applications. The software can't tell the difference, and neither can the eye. The 24F vs. 24P thing is a complete non-issue. For all practical purposes, they're identical.

Er... Chris, I'm usually in agreement with everything you say, but I don't agree here. The 24F mode in the in the XLH1 still comes from an interlace-scanning CCD imager and the internal DSP must de-interlace to create the progressive frame. On fast motion, I can still get interlace combing or scanline shift in 24F mode. In controlled shooting situations, this isn't a problem when you plan your camera moves... When shooting live, high speed events, it shows. It's still best to shoot in 24F mode if your target is 24p because the CCD can scan at 48i and the de-interlacing is done internally on raw info from the sensor block before correction and compression. It's fine and great that it captures as 24P, because you're right, it is encoded that way, but de-interlacing of an interlaced CCD source sill isn't true 24P and no de-interlacing algorithm can reach back in time 1/48th of a second to make half the image match the previous set of scanlines.

[Jeff Kilgroe]

Originally Posted by Dennis Hingsberg

I wouldn't say superior resolution. Both cameras use the same bag of tricks to achieve higher "effective" resolutions. Canon goes with horizontal pixel shift to achieve 1440 pixels horizontally but since it's an interlaced sensor offers 1080 pixels natively for vertical resolution.

Chris already beat me to it, but there is no pixel shift on the Canon CCD. All elements are native 1440x1080 active pixels.

A resolution chart in front of both cams may reveal slightly more sharpness with the XLH1 in 1080i modes but for 720p shooters likely no advantage - especially if you have to deinterlace in post or go with the frame mode of the Canon.

There is more detail there, but the frame mode does present a problem on fast moving subjects - as I noted above. On a static res chart, the pixel shift of the HVX makes it appear to have more resolution than it does... Various wobulation techniques can break down the pixel shift, as can fast motion once again, and detail is lost. Subjects that are far from the camera such as scenery type shots begin to test the limits of the pixel shift. Look at Robert's report on the H1 vs. the HVX for scenery/landscape shots.

Adding P2 cards to the HVX may not make the XLH1 more expensive, but if you want to record 4:2:2 you need to go external with the XLH1 and that's no cheap option either. Let's not forget the advantages of shooting to solid state memory either ie. no mechanical parts, no maintenance, quiet, use in harsh environments, drop it in your fish tank, etc..

The 4:2:2 isn't all it's cracked up to be... Not by the time the DVCPROHD codec trashes your resolution from a pixel shifted 960x540 (up to 1280x720 or 1920x1080) down to a 960x720 and 1280x1080 respectively (1440x1080 for PAL users). The pixel shift along with Panasonic's color matrix create stunning visuals with the HVX and I do love the camera for this. But sooner or later, depending on the task at hand, the pixel shift and low-res CCD block become a hinderance to the camera.

Speaking of 4:2:2 it's more than just less jaggies for chroma-keying. DVCPRO HD uses intraframe compression which compared to HDV makes for a lot more ease during the editing and rendering process. It's exactly why programs like Cineform exist - so people can convert their HDV non frame independent compressed footage to something similar to DVCPRO HD (intraframe compression) and be happy editing.

True, but the conversion to Cineform RAW is quick and painless and it can be done as the video is captured. There are times that the extra color depth of the DVCPROHD from the HVX gives more range to pull a key, but I've found with a properly lit set and greenscreen I have just as easy of a time with the higher resolution of the HDV originated footage. Honestly, I don't know which one is truly better. I have found both to work just fine and both to be lacking of what I want... Hence why I'm shifting over to RED. The keying tests I've done on RED frames are mind-blowing. It's so clean and has all the color depth I can hope for, images practically key themselves in Shake. Personally, I despise HDV in every way. Especially with all the in-fighting amidst the HDV crowd and how tapes from one brand of camera/deck won't always work in another. Sony cameras/decks play back 720p HDV from JVC cameras/decks just fine, but Sony cripples them and won't let the signal pass out over HDMI or firewire, only component. That's just wrong... I have no love for DVCPROHD either.

Anyway, I'm going to get a huge dose of HDV stupidity this weekend and for about 10 days of editing after that. Hopefully it will be the last time... And to make matters worse, it's coming as 720p30 from a JVC JY-HD10U camcorder... w/Letus35 adpert and Nikkor lenses. Should be interesting... I get to cut that with stuff I shoot on the HVX.. All the HVX work is mostly going to be 720p60 and/or greenscreen. At least the project involves hot chicks with guns, so I do at least have that to look forward to.

Let's not even begin to started on other features of the HVX like 4 channel 16 bit uncompressed audio, variable frame rates from 12 to 60 fps, not front end heavy........ . ...

Of course... You just mentioned two of the biggest reasons why I bought the HVX200 in the first place. In fact Panasonic had me at 60fps, 720p. Tapeless workflow was the icing on the cake, even though now I'm a bit grumpy over the [much] slower than initially promised evolution of the P2 format.

[Chris Hurd]

Originally Posted by Jeff Kilgroe

On fast motion, I can still get interlace combing or scanline shift in 24F mode. In controlled shooting situations, this isn't a problem when you plan your camera moves... When shooting live, high speed events, it shows.

I can't help but ask, why choose 24fps for live, high speed events anyway? Personally I've always equated that frame rate with the cinematic style, which is all about controlled shooting situations and planned camera moves (or at least, that's what it should be about, to an old-school geezer like me).

[Chris Hurd]

Originally Posted by Jeff Kilgroe

...there is no pixel shift on the Canon CCD.

Oh yes there is! While all elements on a Canon XL H1 (and XH G1 & A1) are indeed native 1440 x 1080 active pixels, there's also a Pixel Shift process in the horizontal axis added to that. The boost in horizontal resolution provided by Pixel Shift gives the DSP the requisite number of sampling points to create an image "significantly greater than 1440 pixels wide," which of course is then recorded to an HDV cassette as 1440 x 1080 anamorphic (just like HDCAM)... and then automatically scaled to 1920 x 1080 again by any HDTV display... etc.

[Dennis Hingsberg]

If appears I was quite incorrect to state the XLH1 offers 1440 horizontal pixels. It is in fact 1920x1080 (screen resolution) while the block would in fact appear to be 1440x1080. That is quite a bit more horizontal than the HVX in fact.

Vertically however, 1080i and 720p perceptually are suppose to be the same but given Panasonic says their HVX sensor is 1080p (even though you can't get that out of the camera) I wonder how the HVX 1080i stands up to the XLH1 1080i mode - anyone care to comment?

Anyway I think the point being made was why go through the trouble of building a direct add-on 35mm adapter to a modified HXV200 when you can use the XLH1 or HD100 to do the same thing.

My answer to that again is given the low cost of the HVX with all the functionality it packs it becomes very clear why - well at least to me. I'd be shooting 35mm, DVCPRO HD on solid state memory with a few extra in camera features like gamma, etc.. all for less than $6k. Think of it as RED's little brother.

Jeff - by the way that's great you're getting into bed with RED. It's clearly the next step once you factor in the price of a good HD camera and mini35 adapter. I haven't decided on a path for myself yet. I'm working with an XL2 and mini35 and recently added a HVX to the collection. We'll see what's next.

[Jeff Kilgroe]

Originally Posted by Chris Hurd

I can't help but ask, why choose 24fps for live, high speed events anyway? Personally I've always equated that frame rate with the cinematic style, which is all about controlled shooting situations and planned camera moves (or at least, that's what it should be about, to an old-school geezer like me).

...To put a "live event" into the context of 24p cinema. I guess that's the best way to put it. It's not something I've done much of, but it's happened once or twice.

[Jon Fairhurst]

I saw a great boxing video a while back on ESPN2 that used 24p and other cinematic tricks, like wild color correction and overcranking. Rather than giving the feel of a live fight, these guys looked like gladiators.

24p definitely has its uses in the real world - specifically to help make it seem beyond reaity.