Uh oh! Jarrah's mad at SG Collins!

[ejstans]

Each method of capture has its own unique artifacts. Combining them produces a whole extra set of artifacts. Maybe there are many people who wouldn't be able to tell, but there are a large number who would.

I'm interested in learning more about how one can tell these apart.

For example, film has grain, which is very different from video noise, but if using eg 70mm film, it would seem to me wholly unfeasible to detect film grain in a (relatively) poor video recording.

And a telecine process of standard 24fps motion picture to video faces the problem of differing frame-rates but in the context discussed here, one would obviously synchronize the frame-rates of the film projector and the video recorder so no artifacts show up. Same with the shutter.

But film also had a lot wider dynamic range than video, and due to the non-linearity of the S-curve, this would be compressed  meaning that a video recording of a projection might look different than had it been shooting the scene directly. But if the linear range of the S-curve is wide enough for the given contrast while the video recorder's poor enough, I think it'd be possible to simulate?

Perhaps the most difficult properties would be the optical systems rather than the storage media. Film's larger aperature makes depth-of-field differ greatly between film (especially 70mm!) and video. Would it be possible to simulate a video camera's DOF with a 35mm/70mm film camera by stopping down the lens? In fact, DOF is only a subset of the different parameters of the optic systems. For example, one would have to use a different focal length to try to simulate the proper field-of-view of a video camera, which effects DOF. Really not sure of the feasibility of all of this, and would be happy if someone knowledgable would give it a good thought.

Finally, now that I think about it, the largest 35mm film magazines could only hold enough film for about 10min at 24fps and the Apollo record contains long uninterrupted video sequences, probably greatly exceeding this capacity (especially if the filming is supposed to have been shot at higher framerates to try simulate lower gravity.) I'm not sure what technical difficulties there would be with a customized camera+transport to allow capture of enough material matching the Apollo videos?

[Echnaton]

I'm not sure what technical difficulties there would be with a customized camera+transport to allow capture of enough material matching the Apollo videos?

Among other thing, it requires the invention of a camera and video transfer system that has never been shown elsewhere to exist.  One that has the exact properties needed to produce the video in a way that leaves no trace of its existence.  In this particular case the invention is wholly in the mind of the hoax proponent who wants to say the EVAs were recorded on film.

[ka9q]

I worked in TV broadcasting in the mid 1970s and I noticed that it was very easy to tell the difference between a video signal from a TV camera and a video signal from a telecine that was playing 16mm movie film. I always wondered just what it was that distinguished them, but I never did figure it out.

It was also very easy to spot TV from the BBC that had been scan converted from 625/50 PAL to 525/60 NTSC.

Monty Python was a good example of both effects. Their indoor scenes were almost always video and outdoor scenes almost always film.

[Sus_pilot]

I worked in TV broadcasting in the mid 1970s and I noticed that it was very easy to tell the difference between a video signal from a TV camera and a video signal from a telecine that was playing 16mm movie film. I always wondered just what it was that distinguished them, but I never did figure it out.

It was also very easy to spot TV from the BBC that had been scan converted from 625/50 PAL to 525/60 NTSC.

Monty Python was a good example of both effects. Their indoor scenes were almost always video and outdoor scenes almost always film.

Hmmm...  Maybe dwight or someone else more knowledgeable about video processing can help, but from a semi-educated layman's POF (film buff), I can think of two reasons.  First, direct video, especially in those days, looked "hotter".  Also, telecine had the photocopy of a photocopy effect - that is the grain of the film was compounded by the "grain" of the video systems.  Also, panning would jump a bit due to the different frame rates.

The conversion of PAL to NTSC was pretty obvious for all of those BBC and other European shows because the raster lines seemed "thicker".  Also, ghosting off of highlights seemed to be a bit more persistent, but that might have had more to do with the videcon tubes used by each country.

Today's digital systems, such as TI's theatrical presentations, blows me away.  I can't tell it's NOT film (film has always had a certain luster that seemed to be lacking in direct video), other than there are no edit marks, reel change markers, or the tiny imperfections even a brand new first-generation print has.  The closest I've seen with film technology is the loving restoration that My Fair Lady got.

[Jason Thompson]

And a telecine process of standard 24fps motion picture to video faces the problem of differing frame-rates but in the context discussed here, one would obviously synchronize the frame-rates of the film projector and the video recorder so no artifacts show up. Same with the shutter.

That wouldn't eliminate artifacts as you might expect. Now all my knowledge on this is derived from my years as a Doctor Who fan, and the state of Doctor Who episodes from before 1975 in the BBC archives is a matter of great discussion in the fan community. Everything I'm about to say applies to British TV, basically.

Video doesn't have a 'frame rate' as such. On film you take a discrete image every 1/24th of a second. On video you are effectively scanning the scene with a series of horizontal lines. The crucial difference for our purpose is that standard video scans (for example) all the even numbered lines then all the odd numbered lines, then back to the even numbered lines. What's more it does this at about twice the rate of a film recording. The result of this is effectively half the lines update every 1/25th of a second, the other half at the same rate but offset by 1/50th of a second. Visually this provides a much more fluid motion than on film, and is similar to a 50 fps film playback. (Recently there was some complaint about The Hobbit because to get the 3D version they recorded at 50 fps to get the effects they needed, and a lot of people felt it looked more like a TV show than a big budget film, mainly due to this difference in how motion looks on film and video). That effectively eliminates one well-known artifact of film recording: motion blur. Even if you sych up the video and film so that one set of lines updates at exactly the same time as the film frame changes, you will simply transfer the film motion blur to the video.

(To complicate matters even more, there was also the difference between stored and supprssed field film transfer, but we won't go into that. Suffice to say that if they used the wrong one the result was that everything that moved on the video produced a double image on the film.)

On top of that there are things that happen on one medium that just don't happen on the other. Take a look at video and film taken of a burning torch, for example. On film you get a lovely image of a flame flickering. On video you tend to get a bright, formless blob, with colour fringing if you are using colour video, and if you move the flame you get a bright formelss blob witha  big streak behind it. Bright things bloom and streak on video in a way they don't on film. Transferring from one medium to the other won't eliminate that problem. If you have video of a cleanly burning flame you can bet it was shot on film originally. If you have film of a bright blob you can bet it was shot on video.

Also, consider that when you look at British TV shows up until about the late 80s, as ka9q says, it was common to shoot studio scenes on video and location scenes on film (and some of the more difficult effects sequences as well). When you watch the show on a TV you are actually watching a transmission using a videotape of the entire episode. In other words you are watching a film sequence that has been tranferred to video. The difference is marked.

In the case of Doctor Who it gets even more complex. Taking the old black and white episodes as examples, nearly all of them were shot on 2 inch videotape (405 line PAL originally, 625 line PAL from about 1967). Film sequences were also shot on 16 mm film and then transferred to video tape to be spliced into the episode. The whole episode was then transmitted from 2-inch videotape. For overseas sales a film print was made by pointing a 16 mm film camera at a large sreen. These days, these film prints are all that survives in the archives (assuming the episode survives at all: 106 episodes are still missing entirely).

So, with all the various processes involved there is VHS made from 16 mm film of stuff shot on video. There is VHS made from 16 mm film of stuff shot on 16 mm film then tranferred to video. There is some 35 mm film in there too. When we get to colour episodes there is even more variety, with PAL back-conversions of NTSC conversions of colour PAL video, recolourised episodes made from cleaned up black and white 16 mm film prints of PAL colour material which used an NTSC colour signal to put the colour back, or the recent 'chroma recovery' process. There are manually recoloured black and white episodes. There is also a process called VidFIRE that restroes the video look to film prints by interpolating an intermediate field and interlacing them as on video. This process can't deal with motion blur, however, and so you can even tell the difference between an original video shot and one that has been transferred to film and then reconverted back to the video look. Quite often on the film prints you can also discern the pattern of the projection screen overlaying the action!

In the 90s the BBC tried a 'filmising' process to try and make episodes of TV shows shot on video look like film by effectively removing and deinterlacing alternate fields on the video. The result was terrible, since it couldn't compensate for the different lighting requirements of film and video. These days it has improved greatly, with digital techniques for regrading and so on. Today's Doctor Who is shot on digital video that is processed to give it a more 'filmic' look, and it looks superb.

All in all, the lesson here is that the differences between film and video are significant and not easily removed or obscured by recording on one medium and transferring to another. I can watch all those various examples of TV shows and pretty much tell which was used where. The technology to do it really has only just been developed with digital techniques. With analogue technology I doubt it would have been possible at all, at least not without a substantial research and development program that would have left an evidence trail just as large as the one that details the actual techniques used in Apollo to provide live imagery across the Earth from a small, battery-powered, hand-held video camera on the moon.

[ejstans]

And a telecine process of standard 24fps motion picture to video faces the problem of differing frame-rates but in the context discussed here, one would obviously synchronize the frame-rates of the film projector and the video recorder so no artifacts show up. Same with the shutter.

That wouldn't eliminate artifacts as you might expect. Now all my knowledge on this is derived from my years as a Doctor Who fan, and the state of Doctor Who episodes from before 1975 in the BBC archives is a matter of great discussion in the fan community. Everything I'm about to say applies to British TV, basically.

Quick note: In case it wasn't bloody obvious, I'm a happy amateur in this area as well, but I have had an interest in cinematography (until video started taking over...)

Video doesn't have a 'frame rate' as such. On film you take a discrete image every 1/24th of a second.

Just to nitpick here, but in my opinion video has very well defined frame rates (or field rates in case of interlace): 25FPS PAL, ~30FPS NTSC.  And while film has a world-wide standard projection frame rate of 24FPS, it is very common to shoot at different frame rates for various effects such as slow-motion. Until very recently such effects were impossible with video where recording and playback frame rates necessarily matched. I relied very much on this fact in the thread about the Apollo 11 cloud passage to conclude that the CBS video is the best available indicator of time.

On video you are effectively scanning the scene with a series of horizontal lines. The crucial difference for our purpose is that standard video scans (for example) all the even numbered lines then all the odd numbered lines, then back to the even numbered lines. What's more it does this at about twice the rate of a film recording. The result of this is effectively half the lines update every 1/25th of a second, the other half at the same rate but offset by 1/50th of a second. Visually this provides a much more fluid motion than on film, and is similar to a 50 fps film playback. (Recently there was some complaint about The Hobbit because to get the 3D version they recorded at 50 fps to get the effects they needed, and a lot of people felt it looked more like a TV show than a big budget film, mainly due to this difference in how motion looks on film and video). That effectively eliminates one well-known artifact of film recording: motion blur. Even if you sych up the video and film so that one set of lines updates at exactly the same time as the film frame changes, you will simply transfer the film motion blur to the video.

Yes, very true, although motion blur results from the chosen shutter speed rather than the frame rate. And I guess video cameras don't really have a shutter in the sense of a film camera. At high shutter speeds, film cameras can be made to sort of mimic the scanning action of video in that the whole frame is not exposed at the same time.

But anyway, I did say that frame rate and shutter speed of the film should match the video parameters to eliminate these kind of artifacts, and as you note, films shot and projected at higher frame rate look a lot more like video albeit usually not intentionally! We have to keep in mind that for my scenario, we are operating kind of opposite of the normal modus operandi where the goal of a telecine is to preserve the quality and "film look" as much as possible whereas we are here trying to remove as much of the "film look" to have the illusion that the video camera is recording the scene directly without an intermediate.

However, there is a problem that I didn't quite consider (I was writing on my lunch break): assuming the Apollo video is standard ~30FPS, as you have noted above that would be 60 fields-per-second, so to get the proper effect the film must actually be projected at 60FPS. But on top of that, to get the "low-gravity" effect, it had to be filmed at something like 150FPS, which I don't even know if film cameras of the time could do! And as mentioned, this would result in a lot of film to cover the long uninterrupted takes, which also requires very strict synchronization to avoid slippage between projection and recording as well as a fantastic custom transport and storage systsem, particularly as the camera is sometimes even operated hand-held if I recall correctly!! Um, it's starting to seem more feasible to just "shoot on location"

(To complicate matters even more, there was also the difference between stored and supprssed field film transfer, but we won't go into that. Suffice to say that if they used the wrong one the result was that everything that moved on the video produced a double image on the film.)

On top of that there are things that happen on one medium that just don't happen on the other. Take a look at video and film taken of a burning torch, for example. On film you get a lovely image of a flame flickering. On video you tend to get a bright, formless blob, with colour fringing if you are using colour video, and if you move the flame you get a bright formelss blob witha  big streak behind it. Bright things bloom and streak on video in a way they don't on film. Transferring from one medium to the other won't eliminate that problem. If you have video of a cleanly burning flame you can bet it was shot on film originally. If you have film of a bright blob you can bet it was shot on video.

This is due to the limited dynamic range of video, and I guess non-linear effects when operating at the limits. The question is, can a projected film have enough dynamic range to cause the same effects to appear in a video recording of it? Usually this is of course something that is to be desperatly avoided, but in our scenario it would be intentionally strived for in the name of authenticity. Of course, I have already conceded the improbability of recording video of a film projection above, but, for the sake of the argument

[Jason Thompson]

Just to nitpick here, but in my opinion video has very well defined frame rates (or field rates in case of interlace): 25FPS PAL, ~30FPS NTSC.

I guess it depends on how you want to define a frame rate. As far as i know video is (or was until the advent of digital) always interlaced fields, so there is never a discrete 'frame'. That's why i said you can't match up the frame rates due to their overlap with the interlaced fields. That was related to my comment earlier about stored and suppressed field telerecordings. In once case even though the rates are aligned the difference between video and film imaging results in a double image when things move because one film frame captures both fields, and the image on the fields doesn't align due to motion. In the other case it still looks like film even though its on video now.

Until very recently such effects were impossible with video where recording and playback frame rates necessarily matched.

Not entirely true, since slow motion video could be achieved using some very expensive and bulky equipment. However, it could only do a very short section of video at a time. It was often used for action replays of sporting events in the 1970s and 80s, at least on the BBC.

We have to keep in mind that for my scenario, we are operating kind of opposite of the normal modus operandi where the goal of a telecine is to preserve the quality and "film look" as much as possible whereas we are here trying to remove as much of the "film look" to have the illusion that the video camera is recording the scene directly without an intermediate.

I am keeping that in mind, but with all the difficulties listed in the normal process and the variations you'd have to try in order to reduce them (though I doubt they could be eliminated), it certainly seems a lot easier to just send a TV camera to the Moon! It also seems that if such techniques were developed in the 1960s they'd have had far wider application in the following years. The more so considering that NASA's way of working was to contract the actual technical expertise from comapnies who did it themselves, rather than have NASA itself as one all-powerful entity that develops things under a cloak of secrecy.

[nomuse]

I worked in TV broadcasting in the mid 1970s and I noticed that it was very easy to tell the difference between a video signal from a TV camera and a video signal from a telecine that was playing 16mm movie film. I always wondered just what it was that distinguished them, but I never did figure it out.

It was also very easy to spot TV from the BBC that had been scan converted from 625/50 PAL to 525/60 NTSC.

Monty Python was a good example of both effects. Their indoor scenes were almost always video and outdoor scenes almost always film.

According to commentaries I've read about Doctor Who, the BBC had at the time some peculiar regulations on when you could use film and when you could use video.  In many episodes the difference between studio and location is marked.  In others, due to special circumstances, special pleading, or attempts to get around yet another of the many strikes, they'd shoot location using studio methods or vice-versa.  It made for a complex story.

[ejstans]

Just to nitpick here, but in my opinion video has very well defined frame rates (or field rates in case of interlace): 25FPS PAL, ~30FPS NTSC.

I guess it depends on how you want to define a frame rate. As far as i know video is (or was until the advent of digital) always interlaced fields, so there is never a discrete 'frame'. That's why i said you can't match up the frame rates due to their overlap with the interlaced fields. That was related to my comment earlier about stored and suppressed field telerecordings. In once case even though the rates are aligned the difference between video and film imaging results in a double image when things move because one film frame captures both fields, and the image on the fields doesn't align due to motion. In the other case it still looks like film even though its on video now.

Well, a field might well be considered a 'frame' at least for the purposes I was suggesting. Providing synchronization holds there wouldn't be any artifacts related to fields or frames if recording a 60FPS projection.
Edit: Hmmm, I'm not even sure of this now. Each film frame must be fully projected exactly long enough for each field to record it. I'm too sleepy to even work out of this is technically possible with proper shutter speeds...

Until very recently such effects were impossible with video where recording and playback frame rates necessarily matched.

Not entirely true, since slow motion video could be achieved using some very expensive and bulky equipment. However, it could only do a very short section of video at a time. It was often used for action replays of sporting events in the 1970s and 80s, at least on the BBC.

Interesting, I didn't know that!

We have to keep in mind that for my scenario, we are operating kind of opposite of the normal modus operandi where the goal of a telecine is to preserve the quality and "film look" as much as possible whereas we are here trying to remove as much of the "film look" to have the illusion that the video camera is recording the scene directly without an intermediate.

I am keeping that in mind, but with all the difficulties listed in the normal process and the variations you'd have to try in order to reduce them (though I doubt they could be eliminated), it certainly seems a lot easier to just send a TV camera to the Moon! It also seems that if such techniques were developed in the 1960s they'd have had far wider application in the following years. The more so considering that NASA's way of working was to contract the actual technical expertise from comapnies who did it themselves, rather than have NASA itself as one all-powerful entity that develops things under a cloak of secrecy.

Yeah, I agree. Actually, I now kind of wish I hadn't even raised the point because I didn't really think it through very much and in my haste I underestimated the difficulties a lot, but it seemed an interesting point to consider, using film as an intermediate rather than doing the "fakery manipulations" directly with video.

[Jason Thompson]

According to commentaries I've read about Doctor Who, the BBC had at the time some peculiar regulations on when you could use film and when you could use video.

It was all to do with how much it cost to do each. To go on location with film required a film camera or two and plenty of film, and that's about it. Outside broadcast video was rather more complex, requiring the larger video cameras and a few support vehicles. If you've ever seen The Claws of Axos, the big van in the convoy in the first episode which is presented as UNIT mobile HQ is actually an OB video support van. As long as the film cameras had power to drive the mechanism they'd work. With video the whole process was electronic and prone to faults and interruptions once outside the studio.

Exceptions were made. Several interior sequences were shot on film if they needed certain visual effects, mainly because the Ealing film studio was better able to handle those effects than the confines of Lime Grove or BBC Television Centre. Occasionally OB video would be used. Tom Baker's first story, for example, called for a giant robot walking around fields and streets. This could not be achieved on film so OB video was used so they could use the CSO effects to put the giant robot on the scene. Up until 1985 it remained standard practice to record exteriors on film, until an incident during the location shoot in Spain for The Two Doctors. Producer John Nathan-Turner heard from someone back at the Beeb that one set of film prints that had been sent back to be processed had been scratched, so he had to arrange to re-shoot several scenes. On returning to England he found that the original scratched film was actually perfectly usable. To avoid such issues in future he insisted that OB video was used from season 23 onwards. It does make the whole episode look visually more consistent.

[ka9q]

When film was shot for TV in 50 Hz countries, was it run at 25 fps or 24?

Being an American engineer I'm quite familar with 3:2 pulldown, which provides an exact ratio between 24 and 30 fps (the .03 Hz slowdown for NTSC color is small enough for most people to ignore). I've heard that in 50 Hz countries it is (or was) common to just show 24 fps film at 25 fps and hope people wouldn't notice. But what about film shot specifically for TV?

[Jason Thompson]

Providing synchronization holds there wouldn't be any artifacts related to fields or frames if recording a 60FPS projection.

And there's the big problem, I think. Holding synchronisation between an electronic and a mechanical device would be a huge technical challenge. Even the slightest deviation would add up over time to become a noticeable visual anomaly.

Edit: Hmmm, I'm not even sure of this now. Each film frame must be fully projected exactly long enough for each field to record it. I'm too sleepy to even work out of this is technically possible with proper shutter speeds...

Yes, and the transition from one frame to the next, which involves the shutter closing, the film advancing one frame and the shutter re-opening, has to take place in the very short interval between the end of one field scan and the start of the next, or you end up with a blank section on part of your image.

Interesting, I didn't know that!

Apparently there was quite a bit of competition to get time on that machine, and they were booking its use between different departments in allocated minutes. This was from a 'making of' feature on a recent Doctor Who DVD purchase.

Actually, I now kind of wish I hadn't even raised the point

I'm glad you did. Made for an interesting discussion and we learned things in the process.

[dwight]

Trying to explain in words how I can tell sd video from film is difficult. Similar to how I can tell immediatedly whether Hal Blaine or Jim Gordon was playing on a particular song, just by listening to the drumming. Lay the two separate images sources in front of my and my professional expertise will tell withou fail whether its video, film, film of video, or video of film. Thats why i get paid good money for what i do in TV. Also because i am so great and moderately handsome.

[ka9q]

I also have trouble explaining what distinguishes pure video from film through a telecine. Even though I'm pretty familiar with the details of the technologies, I can't quite put my finger on it.

[AtomicDog]

On a similar subject, in the '60s, I used to be able to tell by the audio alone whether a TV broadcast was from ABC, NBC or CBS. I cannot explain it,  but each network had a unique quality to its audio that was instantly identifiable. This changed in the 1970s, and all of the USA broadcasts sound alike to me now.