PRO8 50D, 200T, 250D & 500T CLIPS, PIX & INFO POSTED

[S8 Booster]

Tom, I was merely pointing at the technical max colour depth of film, not "cramming" the film into 8 bit colour depth.

Anyway, there has to be some benfit with film origination when transfered to DV as can be seen on many pure TV production. If the film did not offer soem advantage no one would use it due to the costs.

As of film scanning for DV transfer.
I would like to see the day it is possible to scan/transfer thé K40 (or any other high res reversal) at 2:1 or 4.1 oversampling like analog to digital audio "sampling".

Then I believe we will get the true material digitally available.

It will not be long before we get there.

R

[Tom]

S8 Booster

I understand your point about film's range. Though I would point out that between the extremes the gradation is more or less infinite as there is no digital sampling in film.

35 and 16mm are used to give high percieved production values to drama on television - there are no technical benefits per se. I don't need to tell you about the aesthetic richness and advantage film has over video. It is not a question of resolution - 720 x 576 pixels on Pal digital video will always be so even if you transfer a 70mm print of Star Wars. Hollywood has traditionally been set up for film and in the USA at least they still tend to make the majority of TV dramas there - though this is changing with HDTV. Format and stock costs are less of a concern in this arena.

But Super 8 is not in the same league as these film formats and can never look as good - and the whole point is that if you start with a 35mm source, resolution loss is minimal but with Super 8 it is significant - I have merely been trying to point out an inevitable law of physics - to copy is to degrade, however minimally. This is why Super 8 is only used professionally for effect in say promos or commercials - none has ever used it consistently for tv drama.

I am just speaking technically and if you are happy with the results you get transferring to video then none of this stuff really matters.

Regards.

[mattias]

Though I would point out that between the extremes the gradation is more or less infinite as there is no digital sampling in film.

not true. it's limited by the signal to noise ratio in pretty much exactly the same way as digital sampling is limited by quantization. this is a very basic principle of signal theory, so please don't make such statements if you don't know what you're talking about. we already have way to much misinformation floating around here...

/matt

(for those interested in discussing this issue further, a nice starting point would be that film is actually also digital, since a photon is just a photon and a silver halide crystal is just a silver halide crystal. there's no such thing as a medium grey pixel in the emulsion, so it's really a one bit image with high resolution random dithering, or three such images if we're talking color film)

[Tom]

Mattias

You clearly do not know what you are talking about either. Digital sampling takes the form of steps or slices that are sampled from an analogue image or sound.

If you read my post more clearly instead of responding in a dogmatic manner as I have noticed you do often, you would have noticed that I said gradation. Yes there is a signal to noise ratio, but analogue recording techniques are more or less infinite in their gradation of 'level' if you will, from the definitive slices that are taken in digital sampling.

The random nature of analogue noise has nothing to do with digital sampling or quantization which can be controlled to a much greater degree through cut off filtration.

Film is NOT digital - if you can prove that film can be copied in a lossless process then you will have broken the bounds of science until then respond intelligently.

Regards.

[S8 Booster]

Posted new 10 sec DV clips of 50D, V200T, V250D & V500T.

Hoping they retained the full quality with this setting.

Approx 35MB each.


50D Clip:
ftp://ftp.filmshooting.com/upload/video/E50D.mov

V200T Clip
ftp://ftp.filmshooting.com/upload/video/V200T.mov

V250D Clip:
ftp://ftp.filmshooting.com/upload/video/V250D.mov

V500T Clip:
ftp://ftp.filmshooting.com/upload/video/V500T.mov

R

[S8 Booster]

GENERAL EXPERIENCE WITH THIS TEST BATCH.

The following are my very personal comments and should not by any means be considered as a "Fazit" (Liked that Pedro?). Other cameras and gear may produce different results.

ABOUT PRO8MM:
Excellent service & guidance from Mr. Giles Musitano at Pro8mm in UK.
Transfer quality is good.
Expensive but it is available if you want it.


ABOUT THE FILM STOCK:

E50D:
Good but I do preffer the K40 over this film.
Colours are rich, richer than reality (Common for all neg stocks tested) maybe too rich. The 50D and 250D has richer colours than the "T" stocks. Also the colours appears a little "old fashioned" to me on the 50D, much like I remember colours from an 70´s Instamatic 35mm photo camera although this might be a processing issue.

The 50D was not as sharp as I expected and though it usually is pretty grain free it can be grainy as transferred anyway.
This film had severe jitter and sideways floating and "wobble" on my cam.
If this film "normal" it can not be used with my cam under any circumstances and I do not intend to send it away for adjustment. Rather stay with the k40. Guess the famous pressureplate would stabilize the film around the film port but I am really not sure it would solve it all.

V200T:
For all round purposes the best film I tested. I like the somewhat cooler colour balance over the other stocks, particulary the 50 & 250D. On the DV copy there was slightly more grain (or uneven grain) structure than the 250D which to me was noticably better. I know the camera sets the film speed exact for the 250 film and as far as I have seen the 200ASA should be correctly set too but it might have been set to 160 ASA which might have had an influence on grain.

I guess this film has a big potential for S8 if one get to know it better than I do. I mean it can be real good and if one know how to avoid circumstances where it performs less good it can be real nice.

Concerning the jitter this one is the best and no worse than the standard Agfa Moviechrome A40 reversal of the past. I would say acceptable but not jitter free. Apparently no sideways wobble. I am pretty sure that the pressureplate might take this one up to K40 low jitter and focus standard .

It was somewhat grainy in Tungsten light, more than I expected but it looks better on a 29" Sony TV than on the computer screen. Might have been better off with some lower exposure like the 500 seemed to like but it was not tested.

V250D:
Very (too?) rich colours and the finest and most even grain structure of the 2-500 film range tested. Very good colours in low and high light condition. Pretty light sensitive. Works fine with 220° shutter in low light too.
Quite jittery but better than the 50D. Not usable for any serious stuff as it worked out on my gear.

V500T:
Possibly the biggest surprise. Not as grainy as expected. In daylight shot at 320? ASA it could produce quite low grain but the daylight images were less sharp or more diffuse than the other stocks. Extremely interesting latitude range and very fun to experiment with. It really has a big potential.

In late evening takes (less than f 1.4) with the 85 filter out it still produces excellent colours and light balance from the darkest to the brightest part. Check the first pictures posted. These takes were done in very tight darkness and it appears to be much brighter on the film than it really was. Even the Video camera started to be noisy here even though it had more to go.

I think that the grain produced under these circumstances were an acceptable price to pay for that excellent performance. No problem to me anyway.

It was (as I feel it) sensitive for the settings under tungsten shoot. Small adjustmens made big differences and the grain structure was extremely changing and on my DV copies there was not a lot to go on from acceptable grain to too grainy by slight over exposure.

Unfortunately this film also was too jittery and wobbly to be used for any serious stuff.

Seemed to work very fine with the 220° shutter in very low light conditions.


GENERAL IMPRESSION:

Using these fast films might be an alternative for those who can afford it and want the end result on video.

Last night I had a screening on an electronic projector and the stuff looks incredibly good when electronically projected, noticably better than the MiniDV projected side by side.

However, the colour reproduction drops on both formats and due to the rich colours of the neg film they win easily.

Due to the lower resolution of a electronic projector the image sharpness drops too which seem to influence more on the MiniDV than the S8DV. The funny thing is that the grain totally disappears when the film is projected, even the 500 at a "normal" viewing distance. The images looks so good that it has to be seen to believe it when electronically projected. The images looks brighter on the projector than on 2 different TVs and much brighter than on my MAC Plasma screen making it very difficult to decide which system is more correct. I give this up.


So can these results be used for anything commercial?
For me the jitter issue is deciding as experienced: NO
Is the imagery (Quality) good enough? Can be for some purposes like a presentation via an LCD projector.
Would I think of using it for anything commercial? NO. I would go 16mm for jitter, stability in general and price/quality ratio. (Everyone mentioning this before were right as I see it).
If you are 4 times better than the 16mm guys and 16 times better than the 35mm guys S8 neg might be good for something.


Where do this lead me? Super8 is K40 in the first place. No doubt. I will continue S8 using K40 primarily and fill up with some VNF7240 and possibly some Provia 100/400 when it is available.

I am a projector fan and thus the choice of K 40 is obvious for me.

It is funny and interesting to exploit the potential of the fast neg stocks but it seems to me that using the K40 with the Workprinter for transfer to video would be the obvious choice except for the film speed. If Andreas gets his WP system setup for neg films too that might be an option. Would be interesting to see what he can do. Might turn out very good.

I belive that a combination of the V200T and the new V2/500T might be very good if one gets control over the jitter problem. The film speed certainly takes the S8 into a new dimention.

R

[mattias]

analogue recording techniques are more or less infinite in their gradation of 'level' if you will, from the definitive slices that are taken in digital sampling.

i'm not saying that digital sampling doesn't have problems, far from it. i'm saying that analog "sampling" also has (basically the same) problems. and i think you're confusing sampling with reproduction, and reproducing digitally sampled material with an "infinite gradation" is trivial. oversampling and dithering will take care of it just fine, and you will almost never see the "steps" or "levels" from the sampling when a digital signal is properly converted back to analog.

as for the dogmatic reply, i often use the style from the post i reply to when i reply. imho your statement was even more dogmatic than my reply. i'm perfectly capable of keeping discussion on other levels as well, but when we're talking about scientific issues i like to stick to some kind of scientific stringency. otherwise it's just quasi science and politics, which i'm not interested in...

Film is NOT digital - if you can prove that film can be copied in a lossless process then you will have broken the bounds of science

well, a grain is a discreet unit, on or off, just as a one or a zero. the reason it can't be copied losslessly is that the grain isn't evenly distributed, which actually isn't necessary but just an artifact of the manufacturing process. if somebody would make a film stock with the grain layed out in some kind of fixed pattern it would look just as good and, at least theoretically, would be perfectly possible to copy without loss...

/matt

[Tom]

Mattias

"i'm perfectly capable of keeping discussion on other levels as well, but when we're talking about scientific issues i like to stick to some kind of scientific stringency. otherwise it's just quasi science and politics, which i'm not interested in... "

What are you talking about - where is YOUR science? Take the following:

"well, a grain is a discreet unit, on or off, just as a one or a zero. the reason it can't be copied losslessly is that the grain isn't evenly distributed, which actually isn't necessary but just an artifact of the manufacturing process. if somebody would make a film stock with the grain layed out in some kind of fixed pattern it would look just as good and, at least theoretically, would be perfectly possible to copy without loss... "

Analogue is analogue - it is NOT sampling - it is the real world reaction of recording media to physical stimulus. Digital is digital - it is the electronic sampling of analogue signals into computer bits or data streams. If someone can prove the exsistence of a Higher being then I might be religious, if if if. You are talking palpable nonsense merely to argue. Digital computer bits are just a series of 0 and 1s, and if the medium can copy the data stream without dropping any bits then it is in a very real sense a perfect copy. In practice however with DV for instance which is 5:1 compression we have the added complication of compression and decompression and furthermore the imperfections of tape and hard disks - potential areas for dropped bits - which means we never truly have lossless copies - but once the intital compression has taken place and as long as no further compression occurs (e.g. fx and titles in NLE systems) we will have something pretty close which will copy without perceptible degradation up to about 4 or 5 times. Film contains silver halide grains (which are not discreet in a sense of uniform shape or size) or whatever modern equvilent there is now, which are sensitized at different rates depending on exposure - they are NOT digital bits and if they were merely on or off, in a system as you suggest could be invented, you would have an image of black and white without shading. Instead they are random, they react to physical light in a linear manner and they are incapable of producing the same result twice - every copy is significantly removed from the original!

I really have nothing more to say to you on this subject and do not wish to respond to rubbish like this any more and so I will leave it here - all the best.

Regards.

[S8 Booster]

Not intenting to interfere in this dicussion. My point was only that digital technology gets better every next day. My expression of capturing the very last grain was only a way to say that if you have the capacity to capture the smallest grain it is easy to set the "best" (somewhat lower) resolution available to match that particular film. Not meaning that every grain was to be scanned or used.

Just to visualise Mattias analog description to audio. In the late 80´s the worlds biggest company by then NTT (would probably still have been if it was not split) in Japan invented the MASH ("Multi stAge noise SHaping" with 32 times oversampling) 1 bit D/A converter, (Not Technichs or Matsushita or whoever claims it - you will see NTT is printed on the MASHes) . This converter samples the 16 bit level digital signal by 32 times and by using a feed back / feed forward signal correction system that creates near perfect sinewave signals of the raw 16 bit sample.even at -90 dB levels.

The funny thing is that this unit (improved version) now finds its way into modern DVD players to improve the playback sound. Dual or quadruple DA units totally corrects the phase error too.

Here is an image to see the difference just as a matter of technical detail.

Left is a conventional 44.1 kHz sampler - right is a MASH D/A unit.



New MASH for DVD players: http://www.techtronics.com/files/technics-new-da.html


R

[S8 Booster]

Posting a detail to show how bad the DVDV colour reproduction is compared to the S8DV. This pict is details taken from the first 2 posted.
The reference is the yellow house in the background.

If you go back to the pictures first posted it is easier to see. I have altered the DVDV picture to resemble the same focal lengths between the two cams.
http://www.8mm.filmshooting.com/scripts ... .php?t=885



I would say the film (right) is a perfect match while the DVDV Video (left) is a perfect joke.

My humble opinion.

R

[mattias]

What are you talking about - where is YOUR science?

well, i have no interest in simplifying the already simple to cater for those who can't take a minute ot their time to actually try to follow. i'm sorry i sort of atacked you in my first post. that made you defensive which was obviously counter productive in this case.

Analogue is analogue - it is NOT sampling[/wuote]

yeah, great. do you notice how all you achieve is a perfect demonstration of how you just don't get it? i'm talking about signal theory, how accurately you can reproduce a signal, and not about what physical and logical entities are used to represent it, which is completely irrelevant in a discussion about the "quality". the important thing about the "digitalness" of film is that there are only two "levels" in the film emulsion, which is on and off, black or white, grain or no grain, while you claim that film has an infinite gradation.

Film contains silver halide grains (which are not discreet in a sense of uniform shape or size)

they are discrete as far as levels go. there's no infinite gradation in the grains*. what makes up the grey level of the film is how densely those particles are placed, which is *very* similar to how dithering and oversampling is used in digital recordings. try scaling a 8 bit black and white digital image up to at least ten times its size, reduce it to one bit depth, then back to full depth, and finally back to normal size. you'll notice how all the gradation was preserved in the process, even though you threw it all away. magic? i think not. ;-)


/matt

* by grain i mean a single grain particle, and not the "grain islands" that humans percieve as grain when viewing the film

[S8 Booster]

On these last pictures, what would the difference be in f-stop? How much did you underexpose?

Nag

The brightest picture - the most grainy one was shot at f2,8.
The dark one (lower grain) was set at f5,6.

Cam film speed was set at 400 ASA. Film stock was 500 ASA.

Basically the film wasn´t actually underxposed as Mattias pointed out.

At the above picture (bright) I simply pointed the camera at the dark part of the room where no direct light was visible and locked the aperture at f2.8. Next I pointed it at the bright part of the room and shot the clip.

At the "dark" picture I pointed the camera at the bright part of the room and locked the aperture at f5,6 then turning the camera to the "dark" or normal part of the room and shot the next frames which is the "dark" picture, maybe normal.

Possibly a setting of f4,0 -1/3 would be correct to compensate for the incorrect film speed setting.

Hope this helps.

R

[Guest]

S8 Booster... I transfered your dv clips to digibeta in order to see the results on a pro video monitor, and see what our Da Vinci could do...
I showed them to both our colorist and a cameraman and they thought it was not so bad (which is quite a relief since I'm trying to talk these guys into working with Super8 on some project I have...) The colorist found that the grain was odd in some parts and wondered if there wasn't some noise in the picture actually. Overall your shots lack sharpness, and trying to grade them on a video transfer base is not as easy as it could've been, had we been working directly on the negative. Still, some shots like the kids bathing in the sunset had pretty much latitude and could be pushed within in a reasonable range. What's strange is there is no big grain difference between the 50D and the 500T. But the grain on 50D looks way stronger than on regular K40. So I guess I'll stick with Kodak for my work... I'll try and post the results when we do some tests, if anyone's interested. But that'll probably be with regular K40 and 200T.

[Cheezy]

Sorry guys... that "guest" post was from me...

[S8 Booster]

Hello Cheezy. Thanks for the feedback.

All the clips you worked with seemed to be the ones with the 50% resolution only as Mattias pointed out earlier. I am now posting the 100% res files one by one so if you try to download one of them to do the same process in which it may improve. At least the grainage factor improves vastly from my tests.

In parallell to the 50D clips you tested I shot K40 of some of the same scenery to compare and it looks frightingly good compared to the 50D when projected. Sharpness is incredibly much better and colours are well on par from initial impression.

When Andreas, the master of this forum returns to his service I will send these K40 takes to him for WP transfer to MiniDV to compare with the 50D. I will be very surprised if the K40 do not beat the 50D.


As of the contrast issue (if not ransfer related) I have been wondering if the Canon lens with its "spectral coating" is tuned too much towards optimizing the reversal films in general and K40 in particular for beeing optimal with this neg stock.

It seems to me that this lens is designed to "break up" the hard narrow contrast of the K40 and sort of "widening" the latitude of the K40 which it does miracles with on both sharpness AND colour and exposure balance. Might add "tweak" the latidtude as it seems.

Mated with the 50D in particular this lens may double up the already wide latitude/lower contrast of the 50D and it may prove too much.

Unfortunately I did not test this with my 3 other cams with pure glass lenses with a narrow/hard contrast ratio that could have revealed this.

However, I have a V200T in the can that I will test on the Canon 1014 XL-S, The Nikon R10, the Elmo 1012 S-XL and the Beaulieu 5008/Schenei to compare.

The pure glass lenses might work better with the neg films for sharpness we´ll see.

In general I too thought the higher speed films 200 250 and 500 worked surprisingly good compared to the 50.

New "raw" clips from the cam posted. Totally un-edited and not accurately cut but they are exact copies of the original.

V200T: raw files from the cam 6.8 mb/s playback data rate.
ftp://ftp.filmshooting.com/upload/video/V200T1.dv (16 mb about 2.5s)
ftp://ftp.filmshooting.com/upload/video/V200T2.dv (45 mb about 6s)

R