The JVC HD100 Takes No Prisoners!

[Tim Dashwood]

Originally Posted by Tony Tibbetts

I find this bit pretty suspect myself. First they acknowledge that the JVC and Panasonic are progressive and none of the "faux" camera modes can compete. Yet only the JVC has the true "feel" of real 24fps. Gee the Panasonic doesn't cut it because why? It isn't true progressive like the JVC? Even though it uses progressive chips and has frame discreet compression? I'd argue because of that the HVX is "true" 24fps as opposed to the JVC.

Originally Posted by Tony Tibbetts

The point that was being made in the article related to frame rate cadence as the culprit for the "look" Charters liked. The insinuation being that the 24fps cadence is somehow different than all the other cameras... even the progressive ones. Yet he uses faux progressive modes as the culprit, yet claims that only the JVC progressive 24fps cadence is the true "24p".

Tony,

I try to avoid involving myself in "platform wars" like this because I really don't think they serve any purpose, but Chris asked me to chime in here.

I don't know Rodney Charters that well, but I did meet with him twice in the summer when he was shooting here in Toronto.

I'm the one who gave him some tips and tricks on how to setup the gamma curves in the HD100! He struck me as a straight up honest guy who is very excited by the new technologies available to independent filmmakers.

Like most professional DPs, he and his AC test the cameras extensively before before start of principal photography - regardless of the shooting format. He tested all of the HDV/DVCPro cameras available when that article was written. He is an analytical thinker and looked at all of his options in an unbiased fashion.

Rest assured that Rodney has looked very closely at all of the results and worflow of his tests, and the ProHD acquisition and workflow was determined by him and the producers as the best 'economical' option for shooting BG plates for '24.'

He later agreed to endorse it for JVC.

As far as I know, JVC doesn't pay anyone for endorsements. Rodney is just a nice helpful guy, (born a Kiwi - built his career here in Canada) and I'm sure simply felt that endorsing the product would cause no harm and maybe help his peers who need to make similar choices.

If you read Rodney's quote again, you see that what he says is actually true. Of the Z1, HVX200, XLH1 (which were the only cameras available for comparison at the time) the HD100 is the only camera that can progressively scan the chip at FULL CCD resolution in ONE PASS. This is very important for fast action shooting at 1/48th when you need to match 'real' 35mm footage. The XLH1 was ruled out because it only scans 540 lines when in 24F mode. The HVX200 uses pixel shifting (which is different than interlacing, but still requires multiple scans for every frame) and hence can create a "double image" similar to a bowtie/butterfly shutter instead of the smooth film-like motion blur of a standard 180degree shutter.

This could cause major problems when shooting material from computer screens or electronic signs, which '24' is full of.

Don't believe me? You can see an example of this in a HVX clip Barry Green uploaded back in the spring. Look at post #10 for a frame by frame comparison. http://www.dvinfo.net/conf/showthread.php?p=450390

I'll temporarily re-open this thread and see if we can all be civilized about this.

[John Vincent]

Originally Posted by Tim Dashwood

Of the Z1, HVX200, XLH1 (which were the only cameras available for comparison at the time) the HD100 is the only camera that can progressively scan the chip at FULL CCD resolution in ONE PASS. This is very important for fast action shooting at 1/48th when you need to match 'real' 35mm footage.

The XLH1 was ruled out because it only scans 540 lines when in 24F mode. The HVX200 uses pixel shifting (which is different than interlacing, but still requires multiple scans for every frame) and hence can create a "double image" similar to a bowtie/butterfly shutter instead of the smooth film-like motion blur of a standard 180degree shutter.
This could cause major problems when shooting material from computer screens or electronic signs, which '24' is full of.

Wow - I never knew any of that - despite reading like a million things about all 3 cameras. Does this single pass scan also effect a film transfer (ie - meaning the JVC would be more effective)?

john
evilgeniusentertainment.com

[Barry Green]

Originally Posted by Stephen L. Noe

Although, they are entitled to their opinion the same as you. To them I guess it's not hogwash but a reality. To you it's hogwash.

But ... it's not a matter of opinion. It's either absolutely identical, or it isn't. I'm talking solely about the cadence here, and that's the only thing I was referencing.

Actual 24p rendering is absolutely identical between all the 24p camcorders out there. There are differences in gamma and color and sharpness and all that -- but it is completely scientifically observable and provable that the motion rendition is the same.

So saying that it isn't, is easily disprovable, and thus becomes hogwash.

I'm not arguing that someone isn't entitled to prefer the look of one camera over the other, for whatever reason; obviously that's all completely subjective. But the assertion that there's any difference in the way they handle motion is, at best, mistaken. The only one that handles 24p-ish motion differently is CineFrame 24, the others are absolutely completely thoroughly identical.

[Barry Green]

Originally Posted by Tim Dashwood

The HVX200 uses pixel shifting (which is different than interlacing, but still requires multiple scans for every frame) and hence can create a "double image" similar to a bowtie/butterfly shutter instead of the smooth film-like motion blur of a standard 180degree shutter.

Not in any way accurate. The HVX scans a full progressive frame in one pass. "pixel shift" is not a digital uprezzing technique, and it doesn't scan in multiple passes, and it renders motion identical to how a film camera would.

Don't believe me? You can see an example of this in a HVX clip Barry Green uploaded back in the spring. Look at post #10 for a frame by frame comparison. http://www.dvinfo.net/conf/showthread.php?p=450390

I hadn't seen your post on that before. I would suspect that whatever double image is there happened during the conversion to Quicktime. I don't have that source footage but I'll guarantee there is no such double image in the source footage.

[Barry Green]

similar to a bowtie/butterfly shutter instead of the smooth film-like motion blur of a standard 180degree shutter.

Also, I'm not sure what you're getting at here. A butterfly (or "bowtie") shutter exposes exactly the same as a half-moon shutter. The butterfly shutter just spins half as fast.

Butterfly shutters don't expose the film frame twice. You get exactly the same type of motion, and the same type of motion blur, and the same type of exposure, from a butterfly shutter as you do from a half-moon shutter.

[John Vincent]

Now I'm really confussed.

It's like watching my parents fight :)

Seriously, all good information and interesting stuff.

john
evilgeniusentertainment.com

[Tim Dashwood]

Originally Posted by Barry Green

I'm not sure what you're getting at here. A butterfly (or "bowtie") shutter exposes exactly the same as a half-moon shutter. The butterfly shutter just spins half as fast.

Butterfly shutters don't expose the film frame twice. You get exactly the same type of motion, and the same type of motion blur, and the same type of exposure, from a butterfly shutter as you do from a half-moon shutter.

That's correct. I probably should have been more specific in that I was referring to the CP16, or at least one I used to use in film school (many years ago) that would give us weird "double-exposures" in some cases - especially with TV monitors. We determined at the time that this was due to the butterfly shutter, even though it was supposedly 144° (2 x 72°). It is entirely possible at some point the wrong shutter had been assembled with a body intended for a half-moon shutter, or maybe the CP16 movement just works differently than Arri or Panavision?

Anyway, I apologize. I should have used a better analogy that more of us could relate to.
Maybe software based motion blur would make a better comparison?

Originally Posted by Barry Green

Not in any way accurate. The HVX scans a full progressive frame in one pass. "pixel shift" is not a digital uprezzing technique, and it doesn't scan in multiple passes, and it renders motion identical to how a film camera would.

OK. I'll admit that after re-reading my statement it may have fallen into the "truthiness"(®Stephen Colbert) category, disguised as accepted fact. Considering there is no definitive literature from Panasonic on the topic of scanning rate, I should have been clear that this is my own assertion drawn from limited personal testing and viewing of others' clips.

I also understand that the official line published by Panasonic is that simple pixel-shifting of the green CCD is used to create an effective resolution of 1440 x 810, and that the DSP takes care of the up-conversion to 1080P.

The point I was trying to make (and I think Rodney was alluding to) is that in the case of the HVX200 the motion blur has just never looked to me like one single progressive scan of one moment in time - no matter how long that moment of time determined by the electronic shutter.

I've downloaded and analyzed a lot of clips supplied by forum members.

The motion blur in the images always seem to have a hard edge to hard edge (instead of the expected smooth falloff), even with exposures shorter than 1/24th. In some cases there are obvious "hot spots" of exposure in the middle of the blur.

The most noticeable phenomenon is the "pre-emptive" motion blurring that seems to happen a frame before the camera even starts to move from a fixed position. It just doesn't feel natural.

I can't scientifically test this phenomenon because I don't have 24hour access to a HVX200, but I can speculate that the HVX200 might be 2X overclocking (or always sampling 60Hz) and combining frames in a similar fashion to JVC's "Motion Smooth" or maybe there is some other timing discrepancy between the green CCD and the red/blue CCDs. I don't know, but I am curious, because there is obviously some sort of digital 'magic' going on.

Originally Posted by Barry Green

I hadn't seen your post on that before. I would suspect that whatever double image is there happened during the conversion to Quicktime. I don't have that source footage but I'll guarantee there is no such double image in the source footage.

Luckily, Jarred still has the footage posted on dvxuser.
http://www.dvxuser.com/V6/showthread.php?t=41211
The file is "720-DVCPRO-ball."

It is native DVCProHD at 100Mbps 960x720, and the double image is definitely there.
Considering this particular clip was shot at 59.94fps and Evin said he thought the shutter was set to 180°, each frame should have been exposed for 1/120th of a second. These motion blurs certainly look longer than 1/120.

The scoreboard and aisle lights could create this phenomenon if they were cycling faster than 60hz or the exposure was 1/30th or longer, but wouldn't the double/triple images be less intense than the stationary frames?

[Stephen L. Noe]

Originally Posted by Barry Green

But ... it's not a matter of opinion. It's either absolutely identical, or it isn't. I'm talking solely about the cadence here, and that's the only thing I was referencing.

Actual 24p rendering is absolutely identical between all the 24p camcorders out there. There are differences in gamma and color and sharpness and all that -- but it is completely scientifically observable and provable that the motion rendition is the same.

So saying that it isn't, is easily disprovable, and thus becomes hogwash.

I'm not arguing that someone isn't entitled to prefer the look of one camera over the other, for whatever reason; obviously that's all completely subjective. But the assertion that there's any difference in the way they handle motion is, at best, mistaken. The only one that handles 24p-ish motion differently is CineFrame 24, the others are absolutely completely thoroughly identical.

Although, for some reason, XL-H1 24F footage is reported as interlaced by my system. It's odd but true.

Anyway....

[Barry Green]

Originally Posted by Tim Dashwood

That's correct. I probably should have been more specific in that I was referring to the CP16, or at least one I used to use in film school (many years ago) that would give us weird "double-exposures" in some cases - especially with TV monitors. We determined at the time that this was due to the butterfly shutter, even though it was supposedly 144° (2 x 72°). It is entirely possible at some point the wrong shutter had been assembled with a body intended for a half-moon shutter, or maybe the CP16 movement just works differently than Arri or Panavision?

CP16s had three different shutter systems. Four actually, depending on whether you're talking about the CP16A or the CP16/R. The very first CP16A systems were, I believe, always 180 degree halfmoon. The first CP16R had a 144-degree bowtie, but not all that many of them were made. The later CP16R had a 156-degree, most of them out there are this type (which would cause interference with the TV). The last few (and most sought-after) CP16/R models had a 170-degree halfmoon shutter.

So I would expect that your issue came about because you were using a 156-degree when you thought it was 144-degree?

I also understand that the official line published by Panasonic is that simple pixel-shifting of the green CCD is used to create an effective resolution of 1440 x 810, and that the DSP takes care of the up-conversion to 1080P.

There's nothing hidden about how it works. It's progressive scan, and only progressive scan. It scans the entire CCD at one sweep, using spatial offset to generate more sampling sites than the native chip count, resulting in a 1440x810 matrix, which is carried within a 1920x1080 frame. The DSP then downrezzes to whatever the particular recording format is.

Point is, it's straight progressive scan, all the time.

is that in the case of the HVX200 the motion blur has just never looked to me like one single progressive scan of one moment in time - no matter how long that moment of time determined by the electronic shutter.

But -- it has to, because that's what it is. That's all it is. There's no oversampling of time, no interlacing, nothing like that. Just a reclocked CCD that runs at exactly the frame rate you specify.

I can get ahold of an HD100. Tell me what test you want run side by side and I'll try to work up a comparison so you can see how they respond.

but I can speculate that the HVX200 might be 2X overclocking (or always sampling 60Hz) and combining frames in a similar fashion to JVC's "Motion Smooth" or maybe there is some other timing discrepancy between the green CCD and the red/blue CCDs.

I can guarantee you none of that is happening. But again, tell me the circumstances you want to see in order to do a comparison and I'll try to make it happen.

[Tim Dashwood]

Originally Posted by Stephen L. Noe

Although, for some reason, XL-H1 24F footage is reported as interlaced by my system. It's odd but true.

That makes sense. "1080P" isn't actually a HDTV format. It is actually "1080PsF" (Progressive within a segmented frame,) and also the reason why the highest progressive frame rate possible with 1080i HDTV is 30P.
With HDCAM cameras like the F900, the tape runs a little slower in 1080/24PsF mode and the system clocks at 48Hz, allowing 50 minutes per tape. In 59.94i, 60i, 29.97P, 30P mode, the system clocks at 60Hz (or 59.94Hz) and you can only fit 40 minutes on the tape.

I believe in the case of the XLH1 24F mode, the CCD clocks at 48Hz similar to a HDCAM camera, but then adds 3:2 pulldown and records as standard 1080i59.94 HDV. So regardless of the acquisition format, a HDV stream from a XLH1 should always read as 1080i59.94 (or 1080i50 in Europe.)

[Tim Dashwood]

Originally Posted by Barry Green

I can get ahold of an HD100. Tell me what test you want run side by side and I'll try to work up a comparison so you can see how they respond.

If you can get your hands on a HD200 or 250, I'd be interested to see the two strapped together, various matched shutter speeds and frame rates, and lots of whip panning. Just make sure the "motion smooth" is turned off on the JVC.
Las Vegas at night would probably make a great test!

[Barry Green]

Originally Posted by Tim Dashwood

If you can get your hands on a HD200 or 250, I'd be interested to see the two strapped together, various matched shutter speeds and frame rates, and lots of whip panning. Just make sure the "motion smooth" is turned off on the JVC.
Las Vegas at night would probably make a great test!

I know where I can get a 100, don't know anyone with a 200 or 250. I'm not in Vegas anymore so that would be tough to get, but I'll try to get some stuff that should simply demonstrate how they work. But it's gonna be identical, that much I can assure you already. That's why I'm asking if there's anything in particular you're looking for, that you think might illustrate some difference.

[Antony Michael Wilson]

Tests and science aside, I have always found the Panasonic (DVX and HVX) disappointing in terms of build and image quality. Maybe it's down to the lens. Frankly, the technical details are of little concern to me personally beyond the perceptibly lower native resolution of the HVX. Technically, HDV1 is a pretty weak format while DVCProHD had certain obvious advantages. All the same, to my eyes the JVC HD100 nearly always produces better pictures than the HVX (please note I am not talking about high-end DVCProHD cameras) including progressive scan motion rendering. I have just a small suspicion that most of those on this board also favour the JVC and no amount of technical hair-splitting is going to convince anyone that Mr. Charters was wrong to prefer the HD100 for aesthetic reasons. Probably, to most of us, that 24p motion cadence on low-end HD cameras is identical on all but the Canon is moot.

[Barry Green]

Originally Posted by Antony Michael Wilson

no amount of technical hair-splitting is going to convince anyone that Mr. Charters was wrong to prefer the HD100 for aesthetic reasons.

I don't think anyone should try to convince anyone of that. Everyone is entitled to prefer the aesthetics of whatever they prefer, regardless of whatever the technical specs are. And, regardless of what their criteria are. There are people that prefer the 60i look of the FX1/Z1 over all the others, and they're not wrong. That aspect is completely subjective. I doubt they'd find much agreement on this board, but hey, if they like it better, they like it better.

Probably, to most of us, that 24p motion cadence on low-end HD cameras is identical on all but the Canon is moot.

Maybe so, but an assertion was made that the cadence/rendering is different, and that's incorrect, so I would figure that there are (at least some) who would care to know what the actual story is. I'm not trying to "convince" anyone of anything, just demonstrate what is and what isn't. I'm sure there are plenty of people out there who want to insist that the sun rises in the west and sets in the east. But it doesn't.

(and it's identical on the Canon too; it's the Sony Z1/FX1 that is different).

[Antony Michael Wilson]

Originally Posted by Barry Green

I'm sure there are plenty of people out there who want to insist that the sun rises in the west and sets in the east. But it doesn't.

That's true.