Addressing Motion blur in Progressive modes.
 

Hi,
The motion blur during pans and on moving objects when shooting in Progressive mode is something I don't understand. Is there a way to check that? Shooting interlaced on my HM100 gives a much cleaner footage but it has that 'videoy' look. Progressive addresses that but the picture goes for a toss no thanks to that blur. I was really excited about shooting progressive, but is there a way to fix that since I intend to shoot something in 24p or 25p and transfer to film later? At this rate I'd rather stick with 50i...
Thanks...

24p/25p in all cameras, (film/video) should be 1/48th and 1/50th shutter speed generally. 60i/50i should be 1/60th and 1/50th as well. There is about the same amount of blur to be honest. What is really happening is there are less frames per second in 24/25p compared to 60/50i interlaced fields (notice I didn't say frames.. ) that make up a supposed 30/25 frames that at no time ever really exist. So interlaced video seems smoother, but it also has lower resolution and your HDTV is making the odd or even lines as to how what it thinks the lines would be. A good 24/25p SD looks as good a many 1080i cameras for that reason.

Now if you referring to "Juddering" in panning and zooming, that is something you can deal with while taping. old rule (80+ years?) panning or having subjects cross the film plane should take 5 seconds or longer on typical 4:3 or 16:9 aspect ratio. Longer for wider format cameras. You can pan faster as well, (whip pans).

There is a speed that many of us used to pan and zoom with 60i/50i cameras that don't translate well to 24/25p cameras.. so short story, pan and zoom a little slower or faster.

Thanks Alex..

I'll try with different shutter speeds as you suggested and keep you posted. But as you may know, it's won't be always possible to control the movements within the frames.

So you suggest I keep shooting in progressive if I want to convert my footage to film later?? Even seeing shots frame by frame, I'm worried that blur or the 'juddering' will copy directly on film, and will be even more prominent on the big screen..

For any fast motion I would use 720/60p.

since film is 24p 1/48th second standard, i would advise to experiment a lot with it and get used to it.... I shoot everything from sports and action etc in 24p 1/48th. It forces me to be a little better camera operator to remember what I used to do with 16mm and 35mm film. Have to unlearn a lot of bad habits from video 60i.

If you shoot 60p at 1/120th, if you are NOT planning on footage for slow motion later, then you will have UGLY shutter speed issues if your final 24p etc. If you shoot 60p at 1/60th then you have nearly as much blur as you were saying you didn't like. 1/100th still starts looking like the action scenes of SAVING PRIVATE RYAN... so be careful and use your ND on Max in daylight.

If your friends have kids, go tape a soccer game or softball etc and shoot an entire game and experiment. The reason I suggest it since you really won't have control over your subject so it makes you make decisions on the fly. Some will be good, some will be bad, but you will learn from it. Focusing, zooming, panning, shutter speed, frame rate etc.

Also if you end up distributing via iTunes (largest legal music and video content in the world) the only HD content they do is 720p @ 24fps. 30fps or 60i HDV comes out as 960x540 @ 30fps. not bad really.. but it's closer to European PAL than HD. You might find later that the magical "HD" might sell more content as well. Just a feeling of mine.... since I buy or watch HD before SD.

Alex, first of all a lot depends on your output. If it is broadcast application 30/60 frame rate is a must. I worked on several docs like this: YouTube - Oregon Field Guide Ice Climbing - Part 1 and a lot of them were filmed just in 60i.
I also worked on some ski and mtn bike projects- all of them filmed in the past in 60i, however on the recent Red Bull shoot by Bella Coola B.C. 720/60p was used a lot.

Alex,
your insights are a real eye opener. Something I am beginning to learn to deal with...It's all in the shoot stupid..not just the cam. LOL. Interestingly I dug up another article that mirrors your exact sentiment, maybe of interest to other readers like me out there.

Thanks for your help, and am looking forward to deriving more out of your tos and fros with Robert ;-) Keep it coming guys.

Well not exactly. There are many reasons why low-cost HD cameras shooting at 24, 25, and 30 show far more judder (both "sampling" judder and "2-3 pulldown" judder) than do CineAlta and Varicam cameras.

If you want to understand why this is the case, click to Broadcast Engineering at:

Broadcast Engineering and Digital Television

Now search for "mullen" and select the "24p judder" story OR simply click to:

24p judder

The advice of "correctly select shutter-speed" and "shoot right" will never enable you to duplicate the judder of 16mm or 35mm film stock or that of an expensive HD camera with an expensive lens. The advice can certainly help you, but only $$$ solves the problem fully. There's a reason why folks rent HD cameras and carefully choose prime lenses when they plan to shoot HD to go to film.

You might want to ask yourself a couple of questions:

1) what is the real chance you'll have the money to finish to film?

2) does the subject matter really lend itself to 24p/25p GIVEN judder will be high. Are you sure 720p50 isn't going to look better in the end?

You can also use panning speed charts if you're using 16/35mm adapters.

I have a German one, but you can google and find items like this: Panaflex users' manual - Google Books

I don't know if anyone has computed a panning speed chart for 1/3 lenses

Most of you panning/motion problems can be solved with careful pre-production (except in surprise documentary situations, e.g.), as is true for many depth-of-field issues.

This wonderful site has some great resources & tips !

oh well.. bust goes the bubble... again.

you are right Steve, no amount of permutaion and combination of shutter speeds or aperture settings etc. improves the judder on my HM100. I guess as you said, it's the 1/4th inch CCDs that are the culprit here.

However, can you please tell me if I shoot in 1920X1080 50i, ( which does not judder btw) will I be able to successfuly transfer my footage to 35mm film? assuming I manage the audio separately. I mean wouldn't an interlaced film transfer from a 1/4th inch camera be better than a progressive transfer?

Thanks.

Amir, I just don't see why do you insist on shooting 1080? 1080 resolution with these cameras (as well as many others) is achieved via pixel shift. However they record 720 native. I filmed ski and mtn bike footage and the best results were achieved (mostly with HVX200, 2 days with HM100) in 720/60p. (not in 1080/30 or 24) I also shot some footage with GY-HM100 and for faster motion I used 480/60p option, as the matter of fact the footage looked better then 720/30 or 720/24p in regard of motion blur.
I don't understand this obsession with "film look". From a historical perspective it doesn't make any sense. Way back when all the news footage was shot on 16mm film. But now nobody expects "film look" in the news, the video look is what everyone expects. I mean if you want that "film look", shoot film. 99% of the content will end up selling a few hunded to a few thousands of copies on a DVD or will be distributed on the web or will never see the world beyond edit bay. So I'd say don't sweat over the technicalities (to a degree, after all it should look decent), and have fun with it.

well, to go with Robert's post, "film look", so to speak, is not just dof and controlled motion blur. what distinguishes a theatrical motion picture from video newsreels or tv documentaires is another approach to lighting, blocking, composing in the frame and movement in and out of it, instinctive or deliberate use of film grammar, sophisticated camera movement, and many more criteria. also, not all stories deserve a 35mm or widescreen treatment.


Furthermore, cult films such as Aronovski's "Pi" were shot in 16mm; more recently, the lovely "Once" by John Carney was shot with a standard HVR-Z1. so - film look has a little to do with the camera and a lot with the director's and DP's ability to stage and direct.

Claude, I don't know if you had a chance to see a documentary "The Bridge" THE BRIDGE ::: The Movie ? This doc was shot with XL2's and for one moment I was not thinking how this movie was shot. I think there are numerous samples of other movies (Slumdog, Jesus Camp, Deliver us from evil, Gabriel and many more), shot on simple equipment. The most watched Discovery program "Deadliest Catch" is shot in HDV with Canon A1's (mostly). I'd say content, content, content. Like you pointed out: lighting, shot selection, captivating story, interesting characters, and not forget edit will make a great documentary or a film.

Robert, this looks like an amazing movie, i MUST get hold of the dvd.

as you say, story, characters, content - an issue many filmmakers struggle with, especially with docus, myself included.

luckily, my present producer pushes me hard on story development, and he's right.

Hey Robert..
Guys shooting FICTION will always be obsessed with the film look. You can't help that. You just get conditioned that way having grown up admiring, learning and seeing movies - movies that were shot on film. They'd rather die than see their film looking like an afternoon soap.. YUK! (which btw are shot on high end cameras too) Haha.

But the harsh check on reality is that film is expensive to shoot on for Indiemakers. So it's no wonder our excitement knew no bounds when we came to know there are affordable cameras out there that can emulate the film look without us having to shell out big bucks for stock, processing, developing, telecine etc. - not to mention the long and cumbersome process involved. And a film judder in the footage is the easiest way to achieve that ;-)

Shooting on non film stock cameras is the future, sure.. but the filmmakers are taking that route only because they are convinced that the new cameras give them that look. Even if Lucas, shooting on HD, was intending a video only release , he still would have perhaps converted it to film and then back to video before releasing it on DVD.

And you are right.. Sure you dont' need film cameras or film look for shooting mountain bikes or nature. Ditto for newsreels and documentaries. But dramatic film making does require those..even if going by the fact that audience doesn't really care what format it's shot on as long as the content is captivating enough for them.. But as a filmmaker I'd still want to give it to them with a film look and not video - even if it's high end. I can't say why? it's just built in the psyche. or like you said obsession. haha.

Regards..
btw the hot pixel on my HM100 still needs to be fixed :-(

I think you might be surprised what "soaps" look like when shot in HD in the USA. Likewise, watch some Japanese and Korean "costume" drama shot in 1080i60. I've watched Indian soaps and the cameras may have been expensive but the production values were just plain horrible. Which comes back to lighting and shooting style.

You can, of course shoot 720p60 and drop into a 720p24 Timeline and get "real" 24p to go to DVD/BD. The shutter speed will be too slightly too fast, but 24fps sampling judder will be there -- plus the 2-3 motion judder needed for 720p24 to go to a 60Hz flat-screen. You can even get fancy and, in post, add a bit of motion blur if you want.

This is how 720p60 is being used with the Panasonic GH1.

Hey Steve,
that tip for the timeline fix sounds great.. can't wait to try it and will let you know how it goes. Thanks.

And Indian soaps were nowhere in the reckoning in my post. These guys still shoot on Digibeta and production values as you said, do leave a lot to be desired. However have not really been able to dig any HD samples of current U.S. soaps on the net, nor the costume dramas you mentioned. The only ones I managed to view were current clips of some American soaps that had video suds all over them ;-)

Though will keep looking and get back to you. (a link or two might be useful, since you've really aroused my curiosity about soaps shot on HD)

Thanks and cheers

In the USA some cable networks have Korean TV which includes Japan and India programming. The HD stuff I saw when I lived in Asia. HD has been in Tokyo since about `86. All 1080i60. SD was all DigiBeta. No film except for stuff shot before the `80's. No interest in 24fps for anything.

Not sure if any programming is 720p60 or there is an HD service that is anything other than 1080i60.

Interestingly, much SD drama shot on-location with a very economical style. One camera with one run-through on person A and another run-through on person B. Then simple cuts between the two takes. Zoom used between takes on a camera to get CU or 2-shot or "group" shot.

HD used differently. Shoot one run-though of "group" and then occasional CU. Much like American TV in the 50's. Use HD like a "scope" movie.

Steve, returning to issue of motion blur, when I was shooting 720/60p I liked sometimes more the effect of shutter speed 1/100 vs. 1/120. At least to my eye the ski sequences looked more pleasing. Any comments? (btw we used HVX200, not HM100 then)

Have you watched your 24p footage on a 120hz HDTV?
 

Just to stir the pot a little......

If you really don't like the 24p judder.. get a better monitor or HDTV.. No I'm not being a jerk, I'm just acting like one on the internet. :)

I watched some footage on the newer Sony 32" XBR9 1080p 120 hz LCD HDTV, Freaking AWSOME! The 24p came out MUCH smother (all motion smoothing was off by the way) and was very film-like (as in the movie theater experience showing 24p film) In another couple years all the mid priced HDTV's will be 120hz. It has to do with that 24 doesn't play nice with the standard 60 hz HDTV's. Why is for someone else to chime in. I am told it's because of cycle rate. The 100 hz HDTV's often have motion smoothing systems than the standard 60 hz HDTV to give DVD's and broadcast 24p source material the film motion again, but I'm not impressed.. spend $200 more and get a 120hz name brand HDTV. The better HDTV's are now claiming 240 hz... Does it look any better? Maybe on a bigger HDTV... I didn't see a world of difference in the 40" range.

side note about DirecTV and Dish. Apparently DirecTV and Dish have been dropping their broadcast resolutions to combat macro blocking. Both are compressed to near DVD bandwidth and there was too much objectionable artifacts. Reports of DirecTV being dropping stations to 1280x1080i resolution and Dish down to 1440x1080i. The 720p stations (ABC & Fox) don't seem to be doing that. NBC & CBS stations are 1080i.

Here is an interesting link if you want to read more about multicasting and further squeezing of signal. Could explain why I'm not impressed with my DirecTV, and prefer to get DVD's.

Broadcasters' HD Squeeze Play - 2009-07-02 15:00:00 EDT | Broadcasting & Cable

Because even with pixel shift, 1080 still looks BETTER in most cases. I have the JVC GY-HM100 and have done extensive comparison shots between the exact same scene shot in both 1080 and 720. In shots that involve a lot of fine detail (e.g. leaves of trees in the distance), the 1080 footage (either progressive or interlaced) is definitely sharper and clearer, providing that the footage is filmed under bright light (and, I assume, providing there is a reasonable amount of each of the primary colours (red, green and blue) in the shot). Also the overall look of the shot is less "soft" in 1080 than in 720.

There is nothing "evil" about pixel shifting if the end result looks good. I can honestly say that this camera comes close to Blu-ray quality for many scenes.

I will only ever be using 720p60 (or 720p50) if I have slow-motion scenes that I require.

Actually it's more accurate to say that 1080 still looks better *because* of pixel shift, not despite it. Pixel shift is a good thing; it's something you want, not something to be worried about. With only a couple of notable exceptions, just about every three-chip sensor block ever designed has had some form of pixel offset technology built into it.

Pixel shifting has always been a trade off... lower inherit resolution but gaining light sensitivity and often dynamic range. Sony DV-Cams for instance where always very pixel shifted, but they also where very sensitive in low light even before running gain up. The problem is with more agressive pixel shifting it creates more and more weird artifacts, especially in a progressive format. So it's a balancing act. Zero pixel shift means a better image especially with progressive when recording subjects with straight lines, but less sensitive in low light. Pixel shift in various amounts to increase light sensitivity withoug going to a larger chip (getting away with a 1/4 or 1/3 chip instead of 1/2 or larger) at the expense of various degrees of artifacts depending upon the factory settings.

I personally prefer zero pixel shifting. My JVC HD110 is more than fine with zero gain for daylight situations, and for indoor i'm using soft boxes and fresnel lights so i don't see a problem. You can get the uprezzing of 720p to 1080p in post as well, but you are in control of the scene by scene uprezeing with keyframes. More of a hassel, but I'm fairly certain the results will be better. However these new JVC HM class 1080p does look good, though I haven't used one yet. So I expect JVC has hit a decent balance with their 1080p pixel shifting.

Similar note: Exceptions to what I said.. The Panasonic HVX200 DOES looke appreciable better at 1080p/i than their 720p settings. Apparently when shooting 1080p the chips are 540x960, pixel shifted, then uprezed.... the 720p uses even less pixels and are just uprezed. Not sure on that, but that's what I hear and I tend to believe it from the video testing I saw. Still has great color and pretty good low light capabilites... Again, it's all about trade offs.

This has been the trade-off with CCDs. With CMOS there's no problem to getting chips that have the same number of elements as the recorded resolution AND are very sensitive. A 2M pixel CMOS is trivial while a 2M pixel CCD is very hard as the sampling rate drives-up power consumption and HEAT. And, the smaller the chip the greater the heat issue becomes. With CCDs, 960x540 to 1280x720 is, for a small chip, just about the limit.

Thus, to record HD, pixel-shifting of some type is necessary if you use small CCDs. (And, "yes" CMOS chips have their own problems but multi-channel read-out is helping.)

Here's the rub.

Chris is correct that pixel-shifting has almost always been used. But, for a VERY different reason than it is used for now. To record SD, a chip ideally should be 720x486 for NTSC or 720x576 for PAL. These are EZ chips to build.

Let's look at a PAL chip. At 576-rows it means these CCDs already generate 540-lines of video. To obtain a wide-screen picture WITH THE SAME NUMBER OF HORIZONTAL PIXELS PER INCH on the chip, the chip needs 960 pixels. (Other cameras simply squeezed the wide image into the 720-pixels.)

So companies have been making 960x540 chips for years for their SD cameras that could switch between 4:3 and 16:9. This is why I call them SD chips.

When in 4:3 mode, the design could simply use 720 of the 960, or by off-setting the green CCD, it could SUPER-SAMPLE the image which adds about 15% (nowhere near 50%) more resolution. It's like whip-cream on a Sundae. Nice, but NOT necessary! (My memory is hazy on whether this was really done. I do know that even non-switching camcorders used pixel-shift to SUPER-SAMPLE 720x480 chips to add a bit of "extra" detail.

When you know who wanted to build a low-cost HD camera it realized it could use the "PAL" CCD design and not have to invest in a new HD chip -- as did Sony and JVC. It tried to market this chip without specifying its pixel count. When caught, by me, they switched to the advantages of light sensitivity and dynamic range. Which, to some degree is true.

Of course, SD is a long way from HD. So they used both H and V offset. Since many reviewers were content to use the "increases resolution by up to 50%" concept/claim they thought the camera could capture 1440x810 which is all that's needed for a solid 1280x720 recording. (I guess some thought pixel-shifting could somehow get 200% more resolution so as to record 1920x1080. Even though that's not how the camera worked. If I remember, the image was up-scaled.)

Of course, measurements -- especially NON-static measues -- showed the real increase was about 15% on each axis placing the recorded resolution for 720p far lower than needed to be a true HD camera. Worse, the combination of fundamental and aliasing made it's actual recorded resolution, according to the BBC, impossible to measure.

Of course, the eyes of those who bought these cameras saw none of this. Finally, when the Sony EX shipped and folks saw what 1920x1080 chips could yield, many began to admit that "yes, indeed, the 960x540 generated recordings were a bit soft. Of course they were! You cannot generate REAL information much beyond that which you capture.

PS 1: I'm not saying that the 960x540 generated recordings were not pleasing to many.

PS 2: The ACTIVE process used by JVC on the HM700 is NOT the same as the passive process used in the past. An active process as used by JVC, and Sony on the FX1, really can obtain a cleaner (and with the FX1) a more detailed image. It's takes a DSP chip to do this.

If anyone's interested I have JVC's slides on how their active process eliminates artifacts.

Sure, it is pretty interesting.

I second that. Also I'm very interested in what JVC means when they say their HM cameras have a new Dynamic Digital Signal Processor in which "processing is performed on the full progressive 1920 x 1080 signal, regardless of the camcorder’s settings, ensuring the highest picture quality in any shooting mode."

The chips of the HM cameras are supposed to be 720p native, so I can't quite get my head around the 1080p DDSP thing.

While I look for my slides: It seems then the 1280x720 chips are always pixel-shifted to obtain 1920x1080. JVC is assuming that a 50% increase in resolution is possible. I would expect dynamic (moving) measurement to show more like a 15% increase on both axes.

All signal processing is performed on the 1920x1080 signal. Then the signal is sent to the encoder as 1920x1080 OR converted to 1280x720.

Here's where it gets really interesting.

For a 720p recording, pixel-shifting SUPER-SAMPLES the image to obtain "extra" information. In other words, more detail than can be obtained were pixel-shifting not used. The information each axis is a "bonus."

Is this bonus detail be recorded? No. But, if we compare the recorded signal from two designs: the one that uses SUPER-SAMPLING will have stronger high-frequency information (fine detail) than one that doesn't. So pixel shifting makes a "better" recording at 720p from 720p chips.

By a BETTER I mean more fine detail.

However, pixel-shifting introduces artifacts. It obtains extra information but it introduces errors. Something not ever talked about. :)

JVC's design on the HM700 reduces these errors making the extra detail have less errors. I suspect it does not actually increase measured rez. But, to the eye its pix will be cleaner which is more import than rez. itself.

Now, if you record shots with minimum movement 1080 looks "better" or cleaner, but I think that's exactly where things fall apart with fast moving action? So my question is: in this specific type of shooting is 720/60p better for smoothest possible video?

I wouldn't say 1080 falls apart, since that sounds like a codec problem.

1080i is interlaced which means either your eye or the deinterlacer in your HDTV must deal with 2 half vertical rez images to get one full vertical rez images. The fields arrive at 60Hz and the frames at 30Hz. Although we live with this, it is not ideal. Moreover, with CRTs the interlace didn't need to be DEinterlaced. The CRT was an interlaced display -- out eyes did the integration.

With flat-panels -- electronics must convert 60i to 60p. With motion this can NEVER be done correctly.

With 60p, we get wonderful smooth motion with no interlace artifacts. HDTVs do not have to deinterlace. Everything looks great.

PS: However, the vast majority of LCD monitors only show 330- to about 660-lines of vertical rez when there is motion. So unless you buy a Samsung LED LCD monitor you will have low rez images even with 60p. Only plasma and DLP deliver, with 60p, a full 1000-lines of vertical rez.

So when people talk about camcorder video quality -- one needs to remember that since most folks buy an LCD HDTV, so what they "see" may be funk from their monitor. Unless one has a Kuro or Panasonic plasma, or a DLP, I don't trust what they see.

This becomes even more of a factor with 24p. Here, even the best Panasonic plasmas screw-up. Only the Kuro -- no longer being made -- can correctly show 24p carried in 60i OR 24p from a BD.

I can vouch for that. My tests have shown that while there is greater resolution in 1080 mode, it is certainly not 50% greater than in 720 mode. 15% would be about right.

1080 vs 720 and i60 vs p30
 

OK, my two cents here.

Before I made a final decision about HM100 I did a lot of test and look it on my Panasonic plasma HDTV (full set up to native 1920x1080p60). I made video of some water sport with intensive ball moving under different shot modes and compared picture from camcoder->HDMI->Panasonic. I decided that 1920x1080i60 is the best for sport mode because it shows a very smooth ball movement and it also shows a pretty good overall picture, resolution and color.

Later I established a framefork with NLE on my notebook and I unexpectedly discovered that 1920x1080i60 is not the best! I turned back and did a couple of additional comparisons to solve a puzzle.

Skipping details I want to say that comparison should be done carefully.

1) Don't use LCD for comparison of fast moving picture (just remind) - many of them have a long pixel light time and it produces a "traceable" picture of moving objects. In this case p30 looks better just because it lights little less pixels. It even may look better then p60!

2) Don't compare interleaved shots with progressive in computer monitor. Computer monitors have progressive output today and that it especially true for LCD. The typical video player in computer doesn't do a good de-interlacing job and picture looks with interlacing jadder amid with poor moving object boundaries. After I applied de-interlacing filter in Vegas Pro with _interpolate_fields_ de-interlace method the result looks much much better and reminds me a video on my Panasonic. Before de-interlacing progressive shots looked better than any interlaced. The same is actually for any HDMI output from computer to HDTV if computer graphic board supply progressive 60 - the interlaced picture looks much worse that the same shot from HM100 directly.

3) High resolution large plasma HDTV is able to cope i60 much better but I don't know how - does it use a field type of screen like a regular SD TV or has a good real-time interpolation filter.

Finally I stay on my initial decision - if you want a shot for HDTV (not film and not BlueRay disk) then the best is 1920x1080i60 for high lights.

Would 1920x1080p60 be better than i60? It is a tricky question - p60 leaves less exposition time for CCD pixels and the final picture would be more noisy. I think doing i60 CCD is able to provide much more efficient pixel shift process and increase resolution on the same CCD size too.

Does this explain the flicker I have seen on my HDTV with 1080i footage such as blades of grass that exhibit a lot of highlights and lowlights? I noticed while watching scenes that have a lot of fine detail that there is this flicker that often occurs. It usually only occurs when the camera is handheld and being held reasonably stationary but involving a small amount of movement; when the camera is panned the flicker goes away. I have noticed this in the 1080i mode. I haven't done tests in the 1080p mode. The TV I was watching it on is a Sony Bravia, 46". Whether I shoot in 1080i or 1080p, it only ever sees an interlaced signal from the camera. This is not the case with 720p footage.

Leonid, thanks for your input. However you talk about 1080p60, which does not exist with this camera. Did you mean 1080p30?

No, Actually The last sentence was a generic, what could happen if JVC supports p60 on the same CCD chip size and how could it compare with another shot modes. Just note about engineering difficulties, nothing more.

I would also back this up on the HM100.

If you see it only with interlace you are seeing ""interlace twitter."." When handheld, a camera moves up and down over fine detail so it appears in one field but not another. When a detail only exists in ONE field, it is displayed every other field time, so it flickers at 30Hz.

When you pan up-down fine detail (brick) the horizontal detail becomes alive with "interlace twitter" which is caused by the same alternating field problem.

Frankly, I am always puzzled by why folks praise the higher resolution of 1080i and yet seem not to be bothered by "interlace flicker/twitter." Perhaps this is one reason way Varicam is so loved by filmmakers. Film doesn't have this type of artifact.

At NAB 2008 Sony showed me 1080p60 on their OLED display. WOW! This is the future! And, given the sensor/DSP already works with 1920x1080 at 60p -- I wonder what it would take to record it. It's not impossible. Sanyo already sells a $700 1080p60 camcorder!

Actually, I am surprised that you can see any flicker on TV. TV - and HDTV too - was designed to keep the upper field on screen sometime during low field is displayed. That was because of CRT pixel light latency - the CRT phosphorous had some light time after electron beam hits it. And the time was designed to keep upper field pixels lighting during update of low field pixels and vice versa. That helps increasing screen brightness. And that can be true for many plazma HDTV for the same reason.

Again, I guess that flicker may happen because player-HDTV pair shows the progressive picture based on cheap reconstruction of two fields (upper and low) of an interleaved shot. In this case you may see bright pixel from upper field which comes dark during next frame reconstructed from low field. That was a primary reason why I recommend to avoid looking i60 in progressive mode. Only after rendering interleaved shot to progressive it has sense. It is actually inexpensive in CPU time, at least in Vegas Pro (unfortunately, I have not a serious experience with other NLE).

A stringer nearby shoots football on his $10K HD, and the vision looks CARP. The images are so juddery its not funny. Its a visual staccatto that is quite nauseating to watch and grossly out of place.

My gut feeling is he cant switch the camera to shoot interlaced... Would that fix the problem?

Ben

Of course. As would shooting at 720p60.

But, what some call horrible judder, other call the look of film. Moreover, the flight of a ball wouldn't have judder since the shot would likely be wide and so the motion vector of the ball would be small.

Because filmakers take care about shooting motion they minimize judder. I really don't want to leave the impression that low frame-rates can't be used. Of course they can. But, one needs to be aware.

In preparation for shooting a project involving small waterfalls which will go out to SD-DVD I have been experimenting with various camera settings (HPX170). My objective was to achieve maximum detail and realistic motion. Here were my test results outputing to DVD with playback on 32 and 40 inch HDTVs.

720/30pn, 1/60 - poor detail in water flow
720/60p, 1/60 - better
1080i, 1/60 - best with very smooth yet sharp motion

I had originally wanted to shoot progessive so as to utilize the HPX170 slow - motion capability. Anyone have any suggestions to improve progressive look with this subject matter.