Missing Nix frames

[David G. Healy]

Chris. [ my take on this : . ] I think that transition does highlight some of the difficulties. If you rotate one of the frames about 0.2 degrees so the line of the hedge lines up and the top of the visor and the rear hand hold lines up, the rotational component of what you see almost disappears. Then it seems there is a uniform displacement in the horizontal. If the lawn is then lined up it seems the shift that remains is due to blur and slight parallax shifts all over (single grass leafs and their background which while one cannot differentiate, the overall pattern shifts) as the camera shifts or pans. So there are some of the components.

Another thing is that if one keeps in mind the curvature of the lens and how this bends things at the periphery, then very localised fine alignments cause large shifts as you look away from this localised alignment. So what one ends up with is lvery much a choice. Different ways to show different things.

With the particular MM frames there is this happening as well as other things, to a much greater degree. As David indicates, alteration is an option. I think if one cannot explain (and that doesn't mean it's not explainable, just that the understanding isn't there yet) a time may come when alteration can be said to be proven. Personally I don't think we're there yet.

Hi John,

A longway off proving alteration, non-alteration of anything. Need the original film or 1st generation 35mm *provable* frames! 6 years ago examples as these led me to believe the background area [Elm St. south curb-south] (and/or the foreground) was enlarged for whatever reason. Later [4 years ago], John Costella did not agree with me. Around this time John Costella did the correction work on the MPI Z-frames. Perhaps Chris can try [in another thread] the Costella distorted corrected Z-frames -same frames as above?

David

[John Dolva]

Some of the issues arising made a revisit to the N-Z sync necessary. A possible match at N76 Z366 appears to confirm previous findngs.

As far as M - N,Z sync goes. It's still problematic. I think the earlier finding at M63 is important as it spans to the other end of the film.

I think anomalous imterpolation within that range should not cause a questioning of the sync points while ignoring the possibility of frame drops/alteration.

ie if M01-Z270,271 is correct and a match at M63 is correct and if that doesn't make M42 the headshot frame, then it's possible it is not, irrespective of its numbering and/or irrespective of what the frame as we know it, and the frames on either side of it, shows.

[Frank Agbat]

Some of the issues arising made a revisit to the N-Z sync necessary. A possible match at N76 Z366 appears to confirm previous findngs.

As far as M - N,Z sync goes. It's still problematic. I think the earlier finding at M63 is important as it spans to the other end of the film.

I think anomalous imterpolation within that range should not cause a questioning of the sync points while ignoring the possibility of frame drops/alteration.

ie if M01-Z270,271 is correct and a match at M63 is correct and if that doesn't make M42 the headshot frame, then it's possible it is not, irrespective of its numbering and/or irrespective of what the frame as we know it, and the frames on either side of it, shows.

My issue with timing is rather simple. A frame showing head damage simply cannot precede one that does not -- either within a single film, or across multiple films. If a timing computation leads to that result then it seems to me that something is out of whack somewhere... Either the timing is wrong (frames per second), the sync point is wrong (alignment), or frames have been messed with. I'm not ignoring alteration, I'm merely removing it as a variable for the moment in an attempt to keep things simple, as it were.

I was thinking about this at great length today, and I think understanding sync and timing is incredibly important. Once understood, it gives us, in effect, a higher-speed film of the events. For example: What happens between Z312 and Z313? N22 followed by M42 happens in that space. Our overall "frame rate" has tripled. What one camera missed, another one may have picked up.

[Frank Agbat]

I don't know, but something about the previous animation struck me as exposing a potential optical illusion. Some 3d effects are/can be created by quick flipping of close-perspective images. Naturally, when you stabilized on the vehicle, the background appeared to move. This one is a quick-and-dirty slower frame flip animation with a general registration on the nearer background objects.

Even still, the background appears to shift a bit.

[John Dolva]

I see your point, Frank. A frame not showing head damage cannot be a frame of the headshot. That's true. Does that mean that if the frame according to computation should show damage that the computation is wrong or the frame is messed with.

We can arrive at a determination of whether it should. Of that I'm pretty sure. If the frame continues to not show head damage then there is only one simple choice. The frames are messed with. So a determination of frame number for location will in itself answer that question.

This is one reason why side by side with this I've been looking at the film from the point of view that there is something wrong with it. There seems to be or rather what is available cannot rule it out. If computation determines that there is nothing wrong I have no problems with that. It will make the process go ahead again.

______________

That's an interesting viewpoint re continuity. While it would be hard to visually disply it, I look forward to having a 'Franks DP Films Tables of Reference' to refer to. It will be indispensable IMO. So let's keep at it.

What do you get with the M20-M63 span?

[Frank Agbat]

I see your point, Frank. A frame not showing head damage cannot be a frame of the headshot. That's true. Does that mean that if the frame according to computation should show damage that the computation is wrong or the frame is messed with.

We can arrive at a determination of whether it should. Of that I'm pretty sure. If the frame continues to not show head damage then there is only one simple choice. The frames are messed with. So a determination of frame number for location will in itself answer that question.

Well, that really is the crux of the issue in a nutshell, John. If we find a sync point, and run the math of frame timings, and this leads to a frame showing damage preceding one that does not, then there are only three alternatives: sync point is wrong, math is wrong (aka frame rate is wrong), or the frame is wrong (messed with).

This is one reason why side by side with this I've been looking at the film from the point of view that there is something wrong with it. There seems to be or rather what is available cannot rule it out. If computation determines that there is nothing wrong I have no problems with that. It will make the process go ahead again.

______________

That's an interesting viewpoint re continuity. While it would be hard to visually disply it, I look forward to having a 'Franks DP Films Tables of Reference' to refer to. It will be indispensable IMO. So let's keep at it.

What do you get with the M20-M63 span?

A useful display *is* difficult to conceptualize. It would be a minor programming exercise to get the films to run simultaneously in a reasonable approximation of what might be called "true sync" (assuming one can be established). However, the eye and the brain are not well suited for looking at things from different angles simultaneously.

[Ashton Gray]

This may be a sputteringly naive suggestion, but might there be any possible value in searching for two different sync point for Z/MM that would resolve two discrete sections of the films bookending the MM problem area (one from beginning of MM toward the middle, the other from end of MM toward the middle), and thereby possibly be able to isolate more narrowly where/how much won't sync?

Ashton

Edited January 16, 2007 by Ashton Gray

[Frank Agbat]

This may be a sputteringly naive suggestion, but might there be any possible value in searching for two different sync point for Z/MM that would resolve two discrete sections of the films bookending the MM problem area (one from beginning of MM toward the middle, the other from end of MM toward the middle), and thereby possibly be able to isolate more narrowly where/how much won't sync?

Ashton

I think your point is neither sputtering, nor naive... The issue is finding said points. MM/Nix sync has been illusive, but I have several theories as to why:

1) The cameras are running at nearly identical frame rates

2) They are a fractional frame out of sync with one another

3) The clips overlap is relatively small, and the MM clip is short, so drift doesn't noticeably occur.

MM/Zapruder sync is a bit easier, but the characteristic movements that would be useful in establishing a "solid" sync point invariably end up obscured some way or another.

I have not given up! I just have taken a short sanity break (no -- it hasn't helped, but it is fun to say).

[John Dolva]

This may be a sputteringly naive suggestion, but might there be any possible value in searching for two different sync point for Z/MM that would resolve two discrete sections of the films bookending the MM problem area (one from beginning of MM toward the middle, the other from end of MM toward the middle), and thereby possibly be able to isolate more narrowly where/how much won't sync?

Ashton

I think your point is neither sputtering, nor naive... The issue is finding said points. MM/Nix sync has been illusive, but I have several theories as to why:

1) The cameras are running at nearly identical frame rates

2) They are a fractional frame out of sync with one another

3) The clips overlap is relatively small, and the MM clip is short, so drift doesn't noticeably occur.

MM/Zapruder sync is a bit easier, but the characteristic movements that would be useful in establishing a "solid" sync point invariably end up obscured some way or another.

I have not given up! I just have taken a short sanity break (no -- it hasn't helped, but it is fun to say).

Yeah, me too. There's also the added complication of frame rate reduction as the coil unwinds but I do not think it's imposssible and I do think (in credit of your great work here, I mean it) there should indeed be an index of Franks Synch Tables appendix in any serious book about these issues.

(edit:typo)

Edited January 25, 2009 by John Dolva

[Frank Agbat]

I certainly appreciate the kind words, John.

Admittedly, my tables do not make any unwind assumptions with respect to the frame rates of the respective cameras. I've simply linearized the rates based on the tested and published rates of the actual cameras. While this is probably inaccurate, the results *do* seem to make sense and are observable.

At the same time, it could be said that we don't truly know how far a given camera may have *truly* unwound. So we are back to assumptions and estimates... However, I do agree that it would be possible to include several columns of time adjustments based on several assumed unwind rates.

[John Dolva]

I certainly appreciate the kind words, John.

Admittedly, my tables do not make any unwind assumptions with respect to the frame rates of the respective cameras. I've simply linearized the rates based on the tested and published rates of the actual cameras. While this is probably inaccurate, the results *do* seem to make sense and are observable.

At the same time, it could be said that we don't truly know how far a given camera may have *truly* unwound. So we are back to assumptions and estimates... However, I do agree that it would be possible to include several columns of time adjustments based on several assumed unwind rates.

Like I said, Frank, I mean it. And you're most welcome.

So, while firstly, the idea itself is excellent. Secondly, the aqcuisition of the skills and the process of doing it sets the stage for another round, which may not be right now, but it's there.

Thirdly, I hypothesise that a great deal of accuracy can be achieved as far as frame rate reduction goes by delving deeper into some matters. Apart from what you correctly (IMO) say re. unwind rates, there may be other approaches as well, for example you may remember my odd postings re deducing the speed of an object in relation to camera movement, static background, object movement, from an analysis of the various blurs and knowing the exposure time, form a SINGLE frame or photo. Also, the wheels of the Limo has a specific dameter, and if one can identify any point on this wheel and deducing how far it has moved in an arc during the frame exosure then one may have a very accurate speed. I wonder to what extent, if any, the windown, framerate in anyway changes the exposure rate. I imagine that it is a good idea to have it built so that the frame rate mechanism trips a frame shutter opening time that remains to all intents and purposes constant.

Edited January 26, 2009 by John Dolva

[Frank Agbat]

Thirdly, I hypothesise that a great deal of accuracy can be achieved as far as frame rate reduction goes by delving deeper into some matters. Apart from what you correctly (IMO) say re. unwind rates, there may be other approaches as well, for example you may remember my odd postings re deducing the speed of an object in relation to camera movement, static background, object movement, from an analysis of the various blurs and knowing the exposure time, form a SINGLE frame or photo. Also, the wheels of the Limo has a specific dameter, and if one can identify any point on this wheel and deducing how far it has moved in an arc during the frame exosure then one may have a very accurate speed. I wonder to what extent, if any, the windown, framerate in anyway changes the exposure rate. I imagine that it is a good idea to have it built so that the frame rate mechanism trips a frame shutter opening time that remains to all intents and purposes constant.

You may be on to something here... I need to ponder a bit on what you've written, but I find the ideas intriguing.

Thinking out loud for a moment:

Motion blur is tricky, because we don't have all three axis of motion available to us. Rotation might prove to be more revealing. Also -- we need to look for variances in the shutter speed as well as the frame rate. Frame rate has been used to "timestamp" the events. Time also passed while the frame was held open. I'm sure there is some variability to this -- after all, it is a mechanical contraption at work. It may, however, be more regular than frame rate. I don't know if anyone has researched this aspect of the film cameras involved...

[John Dolva]

Thirdly, I hypothesise that a great deal of accuracy can be achieved as far as frame rate reduction goes by delving deeper into some matters. Apart from what you correctly (IMO) say re. unwind rates, there may be other approaches as well, for example you may remember my odd postings re deducing the speed of an object in relation to camera movement, static background, object movement, from an analysis of the various blurs and knowing the exposure time, form a SINGLE frame or photo. Also, the wheels of the Limo has a specific dameter, and if one can identify any point on this wheel and deducing how far it has moved in an arc during the frame exosure then one may have a very accurate speed. I wonder to what extent, if any, the windown, framerate in anyway changes the exposure rate. I imagine that it is a good idea to have it built so that the frame rate mechanism trips a frame shutter opening time that remains to all intents and purposes constant.

You may be on to something here... I need to ponder a bit on what you've written, but I find the ideas intriguing.

Thinking out loud for a moment:

Motion blur is tricky, because we don't have all three axis of motion available to us. Rotation might prove to be more revealing. Also -- we need to look for variances in the shutter speed as well as the frame rate. Frame rate has been used to "timestamp" the events. Time also passed while the frame was held open. I'm sure there is some variability to this -- after all, it is a mechanical contraption at work. It may, however, be more regular than frame rate. I don't know if anyone has researched this aspect of the film cameras involved... (my italics, let's see)

The three axis issue may not be as elusive as you reasonably suggest. Back to it later.

_

What's your thoughts on this (musings derived from a tangent of the convo Craig and I were having yesterday) : In the Towner film there is an excised frame.

If one looks at the preceding and following frames one sees a reflection that increases in luminosity and then decreases and it seems reasonable to me to assume that this frame was excised because at that moment in time the sun (which is in a known, particular position at any time, anywhere) overexposed the frame. Therefore using standard astronomical, survey, navigational techniques one may be able to say exactly, perhaps even to the second (give or take a reasonable error margin) when that frame was exposed. Then the implications are enormous. Just think about it. Exactly what time lets say 12.33 is. 12.33? is it 12.33.01 or 12.33.59?. Step that through the timing of events such as the time it takes to get to the lower floor issue, or in this case use it as a bencmark to step through all synched films???

The factors to consider is trangulations to determine the exact position of the film surface at that moment, the exact position of the Limo, and then working out where the sun must have been. Then a reference to the appropriate sun tables...

Could this be the primera?

Edited January 27, 2009 by John Dolva

[Russ Connelly]

Thirdly, I hypothesise that a great deal of accuracy can be achieved as far as frame rate reduction goes by delving deeper into some matters. Apart from what you correctly (IMO) say re. unwind rates, there may be other approaches as well, for example you may remember my odd postings re deducing the speed of an object in relation to camera movement, static background, object movement, from an analysis of the various blurs and knowing the exposure time, form a SINGLE frame or photo. Also, the wheels of the Limo has a specific dameter, and if one can identify any point on this wheel and deducing how far it has moved in an arc during the frame exosure then one may have a very accurate speed. I wonder to what extent, if any, the windown, framerate in anyway changes the exposure rate. I imagine that it is a good idea to have it built so that the frame rate mechanism trips a frame shutter opening time that remains to all intents and purposes constant.

You may be on to something here... I need to ponder a bit on what you've written, but I find the ideas intriguing.

Thinking out loud for a moment:

Motion blur is tricky, because we don't have all three axis of motion available to us. Rotation might prove to be more revealing. Also -- we need to look for variances in the shutter speed as well as the frame rate. Frame rate has been used to "timestamp" the events. Time also passed while the frame was held open. I'm sure there is some variability to this -- after all, it is a mechanical contraption at work. It may, however, be more regular than frame rate. I don't know if anyone has researched this aspect of the film cameras involved... (my italics, let's see)

The three axis issue may not be as elusive as you reasonably suggest. Back to it later.

_

What's your thoughts on this (musings derived from a tangent of the convo Craig and I were having yesterday) : In the Towner film there is an excised frame.

If one looks at the preceding and following frames one sees a reflection that increases in luminosity and then decreases and it seems reasonable to me to assume that this frame was excised because at that moment in time the sun (which is in a known, particular position at any time, anywhere) overexposed the frame. Therefore using standard astronomical, survey, navigational techniques one may be able to say exactly, perhaps even to the second (give or take a reasonable error margin) when that frame was exposed. Then the implications are enormous. Just think about it. Exactly what time lets say 12.33 is. 12.33? is it 12.33.01 or 12.33.59?. Step that through the timing of events such as the time it takes to get to the lower floor issue, or in this case use it as a bencmark to step through all synched films???

The factors to consider is trangulations to determine the exact position of the film surface at that moment, the exact position of the Limo, and then working out where the sun must have been. Then a reference to the appropriate sun tables...

Could this be the primera?

This is an interesting idea John, I have read about computer models which can calculate the suns position in the sky relative to an observer, for any location and date (including the creation of sunlight and shadow patterns). This sort of thing is often used in automobile accident investigations when a driver claims that he was blinded by the sun or that a pedestrian stepped out from a area of deep shadow... It is also used in building design to show the impact a building will have on neighbouring buildings in terms of shadows cast, etc...

The unknown for me would be the relative degree of accuracy...will this be accurate within 1 second or 10 seconds, and if so, can we then rely upon the benchmark used for the recorded times for the Hertz sign, the timings of the descent to the 1st floor, etc, etc...in relation to this calculation of 'sun' time as it were.

Edited January 27, 2009 by Russ Connelly

[Craig Lamson]

Thirdly, I hypothesise that a great deal of accuracy can be achieved as far as frame rate reduction goes by delving deeper into some matters. Apart from what you correctly (IMO) say re. unwind rates, there may be other approaches as well, for example you may remember my odd postings re deducing the speed of an object in relation to camera movement, static background, object movement, from an analysis of the various blurs and knowing the exposure time, form a SINGLE frame or photo. Also, the wheels of the Limo has a specific dameter, and if one can identify any point on this wheel and deducing how far it has moved in an arc during the frame exosure then one may have a very accurate speed. I wonder to what extent, if any, the windown, framerate in anyway changes the exposure rate. I imagine that it is a good idea to have it built so that the frame rate mechanism trips a frame shutter opening time that remains to all intents and purposes constant.

You may be on to something here... I need to ponder a bit on what you've written, but I find the ideas intriguing.

Thinking out loud for a moment:

Motion blur is tricky, because we don't have all three axis of motion available to us. Rotation might prove to be more revealing. Also -- we need to look for variances in the shutter speed as well as the frame rate. Frame rate has been used to "timestamp" the events. Time also passed while the frame was held open. I'm sure there is some variability to this -- after all, it is a mechanical contraption at work. It may, however, be more regular than frame rate. I don't know if anyone has researched this aspect of the film cameras involved... (my italics, let's see)

The three axis issue may not be as elusive as you reasonably suggest. Back to it later.

_

What's your thoughts on this (musings derived from a tangent of the convo Craig and I were having yesterday) : In the Towner film there is an excised frame.

If one looks at the preceding and following frames one sees a reflection that increases in luminosity and then decreases and it seems reasonable to me to assume that this frame was excised because at that moment in time the sun (which is in a known, particular position at any time, anywhere) overexposed the frame. Therefore using standard astronomical, survey, navigational techniques one may be able to say exactly, perhaps even to the second (give or take a reasonable error margin) when that frame was exposed. Then the implications are enormous. Just think about it. Exactly what time lets say 12.33 is. 12.33? is it 12.33.01 or 12.33.59?. Step that through the timing of events such as the time it takes to get to the lower floor issue, or in this case use it as a bencmark to step through all synched films???

The factors to consider is trangulations to determine the exact position of the film surface at that moment, the exact position of the Limo, and then working out where the sun must have been. Then a reference to the appropriate sun tables...

Could this be the primera?

I think the biggest problem, John, is that reflections are tricky things. They change rapidly, even with a constant light source in both position and intensity, with even the slightest movement of camera or subject positon. As a studio photographer for decades, tasked with creating perhaps millions of lighting schemes, I have seen this effect countless times. Its that old angle of incidence thing.

You have an interesting concept, but perhaps with many perils.