Frank Agbat

Why should I change my monitor to 800x600? Also, it might help if you gave what Z-frame this image is from, so I can use one that is a little larger and a little less lossy than the one provided.

Bill, I'm not, by any means, an expert on copying Kodachrome II, so I'm asking this for my edification, not as a challenge to what you've written. If I understand the sequence of events, several copies of the Z-film were made essentially at the same time it was developed in the lab. What was used to make these copies?

Honestly, I don't see anything other than a reasonably good example of MPEG2 macroblocks and JPEG artifacting.

Yep -- hidden, as you say, in plain sight. 4 feet short of the 5+00 station... So, if the third shot occurred at what is now Z343 (I'm going to call it Z"343"), giving only 1.6 seconds between shots (not enough), one would have had to lose only 18 or fewer frames to make the gap between Z313 to Z361 (343+18) shrink from 2.6 to 1.6 seconds (Z313 to Z"343"). Reducing 48 frames down to 30 would be damn near impossible to detect without a comparison point of reference...

I've looked at the various files in question regarding NIX frames. I believe in both cases, the NFV video (Groden) was the source of the clip. I took a look at the telecine/interleaving pattern that was used on this DVD. In general, the pattern used seems to be: AA BB B+C CC DD D+E EE FF F+G That is to say: Frame A (full frame) Frame A (duplicate, full frame) Frame B (full frame) Frame B (duplicate) Interlaced (Frame B interlaced with frame C) etc...

Correct. PAL systems typically do not bother with the telecine process the same way as NTSC has to. They usually just do a 1:1 frame transfer and let the movie run slightly fast (25fps vs. 24fps). PAL also has much more true color reproduction... Here are some sample frames. The source in this case is NTSC. Frame 1: (Non-Interlaced Frame) Frame 2: (Interlaced Frame - contains frame 1 and frame 3 on alternating lines - i.e. odd/even fields) Frame 3: (Non-Interlaced) The interlaced frames, along with duplicate frames "correct" the frame rate. Notice the telltale horizontal lines in the interlaced frame. This is very noticeable when these frames pop up on a progressive device (like a computer monitor).

John, Yes, it sounds like you've got the basic pattern correct. If your source is film shot at a rate BELOW the rate of the target (i.e. 24 fps, 18.5 fps, etc), then something must be done to make the film appear to run at the correct speed when played on a video device. If you merely did a 1:1 frame transfer, the video would appear to run too quickly. (For example, if you watch footage shot in the early days of motion pictures, it frequently appears that everyone is moving a bit too quickly. This is because older cameras operated at a frame rate below 24 fps, but the people who produced the video telecined it as if the source was 24...) The nature of the target video device also has something to say about how these frame-rate corrections must be made. If the target follows the PAL (European) standard, then the target is 25 fps. If NTSC, 29.97 fps (frequently noted as 30fps). So, for example, if you're dealing with an NTSC target, you need 29.97 fps to appear to be running at 24fps. This is the telecining process, and it is accomplished by creating duplicates of each original frame in a 3:2 pattern. So, for movie frames 1-4, you would end up with video frames in the pattern: 111 22 333 44 The important thing to note is that at no point in the telecining process (there the source film is at a frame rate BELOW the target device) are frames lost. Copies are made. The secondary process, which may or may not be necessary depending on the target, is interlacing. In the case of NTSC devices, the spec calls for 59.94 fields per second. This is comprised of "odd" and "even" frames. In essence, the device draws the odd numbered lines in the picture, then backs up and draws the even ones. Each half is drawn at 59.94, but it takes two fields to represent a frame... The effective rate becomes 29.97... These numbers are not coincidence, by the way. The NTSC (~30fps/~60 fields/sec) standards were a by-product of North American power generation standards which have a line frequency of 60hz. The standard designers could use this frequency to help synchronize the field/frame events. Thus, standard TVs in North America draw a new field every (approx) 1/60th of a second - right in sync with the power standard. Also note that the European standard (PAL) is 25fps, which just so happens to be a multiple of their power generation frequency (50Hz)... For those that remember the days of TVs having a Vertical Hold knob, that if misadjusted would cause the picture to "roll", what you were doing when you turned that knob was 'fine tuning' the vertical synchronization of your TV to the broadcast -- necessary if there were slight differences in power frequency/phase. Unlike television, the output from computers is progressive. This means that interleaving is not taking place. Each frame sent to the monitor is a full frame, not half. (Not to confuse matters, but some computer *monitors* actual interleave internally. However, the video card's output is progressive). When we start with a source that is interleaved and want to display it on a computer monitor, then the fun begins... David's article went into this in a fairly superficial way, but did explain the issue. If your computer contains DVD playback software, it contains an algorithm to deal with the ugly interlaced frames so they look decent on a computer monitor. The algorithms vary, but many use a technique known as "bob and weave". The key is to make the video look good at full speed. If you want to take interlaced video and display it as an AVI, for example, your computer will NOT automatically remove the interlacing effect. This requires de-interlacing the video. (This, too, is a subject with many opinions and methods. Everyone seems to have their favorite way to do this, and that's not necessarily a bad thing.) There are numerous tools that will do this to varying degrees of success. Once you've de-interlaced the video, if you want progressive frames, you then must un-do the telecining process. This is called "inverse telecine" (IVTC). If the telecining process has been done correctly (and this is NOT always the case, by the way), one can assume which frames to remove. Many software packages accomplish this automatically, simply by assuming a perfect 3:2 pattern. However, it is NOT always flawless, and sometimes in the IVTC process, a frame will get lost here and there (and/or a duplicate frame might survive the process). In the course of a two-hour movie, or something that isn't being looked-at frame by frame, this is not a big deal. On the other hand, when we're attempting to analyze a film source and are hoping to get progressive frames, an incorrect IVTC process that loses a frame is unacceptable. This is what I *think* has happened in the AVI that you're looking at. For my stuff, I did the IVTC process by hand. This made certain that I ended up with progressive frames. Obviously, for a large number of frames, this is an unworkable proposition. I'll try to post some examples of frames, interlacing, and the rest if time permits...

Hi David, Yeah, I know all about telecining, 3:2 pulldown, interlacing/de-interlacing and the like... (A few years ago, in the early days of burnable DVDs, I produced a crazy "torture test" dvd that included such nifty things as MP2 audio tracks, incorrectly telecined sequences, incorrect 3:2 pulldown flag, non-standard GOPs, wrong DVD book-type and the like. It was really fun to watch poorly designed DVD players try to play some of the stuff. It was even more fun to watch the befuddled salesmen...) I was reasonably certain that my source was purely progressive last night, but I was so darn tired I really couldn't think straight. This morning, I noticed my computer wasn't playing *anything* at the correct speed... One reboot later and all is good. Unfortunately, immediately thereafter I had to head to work. I suspect the problem is a memory leak in a program I've been writing.

Well, I have to admit, I'm a bit puzzled about a couple of things. I ran the same process against the Nix avi from the site you indicated. Here are the results: I also re-did my previous one, in an attempt to improve the registration: On your source AVI, I was never able to achieve the level of clarity you achieved, but there are some irregularities in the frame spacing. On my source, I think the results are slightly improved. I blame some of the fuzziness on MPEG-2 block distortion, as the DVD "chops" a bit. The frame spacing seems quite good, with no indication of missing frames. But there is another problem... When I play these frames at 18.5fps (the speed of the Nix camera) the film seems to be in slow motion. I'm at a bit of a loss to explain this at the moment. I'm going to have to go back and look at the source again... When something looks to be in slow-mo, there are usually duplicate frames present, but there are clearly no dupes in the film. There also don't appear to be any missing... (there is probably a simple explanation that I'm missing... it has been a LONG day).

Hi John, My alignment isn't as crisp as yours. I certainly can probably crisp it up a bit with some fine tuning and a more relaxed hand on the mouse (I've had a rather, um, interesting and nerve-rattling day). I also might have used a slightly different procedure than you, upon further review. I attempted to optimize the alignment of each frame with its predecessor, not the first frame. That might lead to some progressive error creeping in to the process. Another possible issue is that I might have screwed up on reversing the telecine process. I would expect mucking up in this area, though, would have yielded some duplicate frames *or* some skipped frames. That did not appear to be the case. Also in the mix was the disappointingly poor quality of the source, even though it was a DVD. All DVDs are not created equally, and this one stinks! I'm going to download the avi file that you used and play with it for a while...

John, I repeated your experiment, albeit using different software, but attempted to match your registration/stabilization technique (which is quite slick, by the way). The only problem I ran into was a minor procedural error regarding how I marked the actual wheel-center location on each frame. Anyway -- here are the results. The small vertical markers (which due to my procedural error got a bit too light in the final version) track the center of the wheel. As you can see by some minor blurring in the background, my registration was not 100% perfect, but it isn't bad for a quick evaluation. My version of Nix doesn't seem to show missing frames. I'm going to use the same AVI that you used, reduce it to individual frames and see if, perhaps, the source is the problem.

Hello John, Yes. It answers my question quite nicely. You are, in effect, stabilizing the frames -- you do this by keeping the background stuff in the same alignment. This takes camera panning out of the equation. (p.s. I like your security system idea. Your idea has given me another idea, but I need to understand how it could be used to measure relative "rate" of motion. Might have applications in a traffic monitoring system.)

John - Two questions: 1) What software are you using to make the panorama-style combined view? 2) How are you eliminating pan-tracking problems from skewing the results?

Agreed. As Ed O'Hagan said in another thread, either the film is altered or it isn't. I didn't agree with some of his other points, but I certainly agreed on this one. One possible complication: Using the front fender may bring too much perspective into play. That is to say, the fender starts to "vanish" as the limo moves out of his panning area toward the end of the sequence. I'm going to attempt to use both the front fender *and* the centerpoint of the front wheel as a positional guide. For those who know: In addition to the framerate of the Nix camera, what is the proper parlance for numbering the Nix frames? Is the very first frame (shot on Houston) N1, or is "N1" the Elm street sequence?

Good catch, John. While I admit that my "alteration jury" is still out, the prospect of removing frames is one that seems *very* plausible and could certainly have been done. I have individual frames of the Nix film. I'm going to attempt to repeat your test and see if I get the same result. For those that care, my source is 29.97 NTSC interleaved (DVD source). I have, by hand, removed the interlaced frames as well as duplicate (i.e. telecined) frames necessary to match frame rates. What remains are individual, full frames (aka Progressive video) that I can play at any framerate I choose. What was the framerate of Nix's original camera? 16fps?

Richard, Thomas has recently outlined his scenario and shot timings in a thread entitled "Conclusion" http://educationforum.ipbhost.com/index.php?showtopic=8560

Ashton, I'm not trying to be a jerk or a naysayer, but I do have one observation about this model... JFK was listing notable to the left at the time of the headshot *and* was facing more to the left. This might change perspectives a bit on feasible vs. non-feasible shot locations. In the case of the postulated location, it might make a shot that strikes the crown of the head more feasible. It is, however, quite useful to see different angles and locations other than the same-old, same-old.

Thomas, I appreciate your reply. It helped clear up a LOT of things -- and I apologize for messing up my station numbers (4+96). In this scenario, is it possible to accomplish the shots *without* z-film (or any other film) alteration? that is to say: a shot at Z205-ish, followed by the headshot at Z313, and then the final shot at Z355 (that is right in front of the yellow marker by Altgens *and* is 2.3 seconds after Z313)... The only part of this scenario that I'm having any trouble with from a technical standpoint is the alteration component. My personal jury is still out on the entire alteration claim, to be honest. I'm more inclined, however, to accept alteration through selective removal of frames than via other more elaborate methods, as this is well within 1963 capability. It also might explain why Dan Rather claimed to see JFK's head driven dramatically forward by the shot...

Thomas, Let me see if I have the theory correct: 1) Three shots from an MC rifle from the TSBD 2) Three hits. a. First one hits the tree, then JFK, and then creates some of the wounds in JBC. b. Second one is at Z313 and hits Kennedy's head high, perhaps tangentially (?) causing massive damage c. Final is after Z313 (station 4+86), when the limo is at the yellow marker near Altgens' position, hits JFK and does the wrist damage to JBC. 3) The time between each shot is longer than normally considered, plenty of time to cycle the MC and re-acquire the target 4) Selective frames were removed from the Z-film to obfuscate the speed of the limo, and eliminate the stop/near-stop reported by many witnesses. 5) Also removed was any frame(s) around Altgens' position showing another bullet striking JFK. 6) Survey data was fudged/ignored to convince the public that Z313 is the final shot. 7) A missed first shot was concocted to move the timetable of the entire shooting forward, to add to the plausibility of Z313 being the last shot. 8) LHO did the shooting, acted alone. He was more than capable of making the shots, especially considering the added time. But... The piece I simply don't understand is: If the actual shooting can be linked to LHO acting alone, but merely with a different sequence of shots than we normally consider, why would the WC, FBI, etc, concoct an alternate story that points to the same conclusion? Isn't that a bit like incorrectly reporting the sequence of scoring in a game, yet correctly reporting the final score? Doesn't this theory hinge heavily on media alteration (z-film, all other films, all other stills)?

Jack, I see where you're coming from -- thanks for the clarification. A better initial phrase from me would have been "many technologies included in the computer you are using now were simply not possible in 1963." Now, back to the discussion of alteration...

Jack, Of course computers existed in 1963... I've been involved in computers professionally since the 1970's. On this subject, I know precisely what I'm talking about. And yes, some of the projects with which I have been involved have been *very* much at the leading edge of the technology. I believe the example is completely and perfectly apt for Mr. Black's question. I can state with 100% confidence that, even including DoD and "leading edge not-known to the public" technology, with all of the DoD, NASA, MIT, and other brainiacs involved, that there are innumerable things that simply were impossible with computer technology in 1963. What does this have to do with alteration directly? Nothing. I'm merely pointing out that at any given time in human history, there *are* things that are impossible to do.

The WC print doesn't look any clearer to me than the same frame from either MPI or Costella.

Charlie, I could list numerous things that could simply have not been achieved in 1963, regardless of how much time and money was invested or how many smart people were involved. The computer you are using right now was simply not possible in 1963... and the list goes on and on. Yes, there are times in human history when things are impossible... Also, even if the ability to do a given thing DID exist at a given time, that doesn't mean that it was done...

If there is a strategy behind the allegations, we're possibly falling for it... Since this morning, we've racked up three pages of posts (perhaps 4 when I add this one), and nearly 700 views. Not only that, but we've got people very close to "calling each other out". If the tactic is diversionary and "divide and conquer" it appears to be working...

Well, John, if you were an agent of the CIA, you could be running this forum under the philosophy of, "keep your friends close, but keep your enemies closer." You must be getting too close to *something* -- whether it is JFK directly or something of nearly equal significance that is in the surroundings.