ASC and BSC camera tests ...

[Steve Freebairn]

You have to keep in mind that some of these formats and cameras only look worse in comparison to each other in certain circumstances... when an entire project is shot using one format or camera, you tend to accept that picture quality and look because you don't have a frame of reference. Hence why 16mm may look worse compared directly to some other formats, but an entire feature shot in 16mm can seem quite acceptable in quality, even attractive.

Very true, anyone who saw Dark Knight in Imax can attest to this. The 35 footage looked good up until they cut to Imax source footage, then you could see where it was lacking in the resolution department.

 

[StephenMurphy]

Hi David,

Any truth in the rumor that 24 is now being shot on 35mm film?

Stephen

24 has always shot on 35mm except for some sequences on Genesis for a particular episode a few seasons ago.

Rodney Charters mentioned on Twitter that they had been given permission to shoot the current seasons first 4 episodes on 35mm but after that the format would change.They tested various digital formats and settled on Super16mm instead. At the last minute the decision was made to stick with 35mm.

 

[Craig W. Bickerstaff]

Did they test the f23? The clip that went up didn't show any f23 footage.

 

[Tony Brown]

Did they test the f23? The clip that went up didn't show any f23 footage.

No. F35

 

[Sanjin Jukic]

Sanjin is engaged in the Bush administration technique of discrediting the messenger when he doesn't agree with the message. If the tests had shown that RED was superior to everything else out there, he'd be praising them and endlessly quoting them.... but since that didn't happen, he's engaged in character assassination in order to discredit what they are saying. This is politics and propaganda.

You can disagree with the techniques used for testing, you can dispute the accuracy of the test, but you don't have to start attacking people when they don't get the results you expected.

David,


sorry but me again to say something more here:


Also thank you for your "nice" comparison... Bush & Co... :biggrin:


1. I didn't attack anybody here or there and as explained above more or less it's opposite as each of us RED USERs were on the attack with that sort of tests.

2. About David Mullen ASC wrote: "This is politics and propaganda." :emote_head_explode: LOL!!! And I would reply on that with a quote: "Man is by nature a political animal." as Aristotle (384 BC - 322 BC) said in "Politics".

3. Those tests obviously wanted to drive a certain future "acquisition-decision" policy, we just don't have to be "stupid" enough and to ignore that intention.


4. I don't promote here any political dilettantism or even "camera agnostic" conformism as some of our RED User members like to do.

 

[Laco Zamba]

24 has always shot on 35mm except for some sequences on Genesis for a particular episode a few seasons ago.

Rodney Charters mentioned on Twitter that they had been given permission to shoot the current seasons first 4 episodes on 35mm but after that the format would change.They tested various digital formats and settled on Super16mm instead. At the last minute the decision was made to stick with 35mm.

Few seasons ago and June 2009

http://pa.photoshelter.com/c/onloca...sM1cPX7UAkKiau2e8u6OPj2tQJDZFAGW0HAc-&_bqO=19

 

[Clint Johnson]

I'm not sure what you're talking about -- ECN2 color neg processing is pretty standardized and automated, there's no "tweaking the processing" unless you request something like push or pull processing, or skipping the bleach step. All color negative stocks, all speeds, Fuji and Kodak, go through it, though 5222 wouldn't since that's a b&w negative stock. And if the labs are following Kodak standards, the results should be the same at any lab. The whole color neg processing system is designed to NOT need fine-tuning and tweaking for individual rolls - in fact, it is near impossible to do so.

Why do you think they did something "special" for the film stock other than a standardized D.I. workflow?

--

The ASC tests were timed to match each other as closely as possible and I would assume the BSC tests were too. Now some would object to that, making the images match each other, but if you don't do that, then people get the wrong impression -- which is the problem with sending the footage off to different D.I. facilities that specialize in just Red or SI2K footage, etc. If you do that, then it will not match footage from other cameras and no one watching the results could get past that.

David, I stand corrected. This just goes to show how basing my thought processes on decade old knowledge in a related field just ends up making me look ignorant. I've only messed around with a little Super8 and 16mm film so I dropped back to my more robust knowledge of 35mm still photography and my constant hunt for a developer who knew what the hell they were doing. The results I got were all over the map and that experience lead me to comparing apples to pineapples in this case.

Just for my own clarification; So a starving indie film maker who managed to gather up a mishmash of short ends from various sources (staying with colour negative for arguments sake)- say Fuji Eterna Vivid 160T and Reala 500D, Kodak Vision2 50D out of some guys fridge and some Vision3 500T... they could shoot with it and then send it out to a half dozen labs randomly selected from the phone book based on whoever had some idle time and could give the best deal. The film could show up in unmarked cans and the lab would just load it up and run it through the ECN-2 process without question and we could expect to get indistinguishable results from the six different labs?

Thinking about it, if the lab did anything different without the express say so of the production company that would be a quick way to get some lawyers involved.

That shows how film is a stable and reliable process and why folks are skittish around something new that hasn't settled in to an automatic process. It doesn't explain to me why they would try and cram that something new into an automatic process.

In a comparison, I'm not interested in matching processes at all, I'm interested in seeing the best that each could do with their own optimized process.

I'd think that the idea would be to create a fixed target for the participants to shoot for and then let them use whatever method works best for the individual cameras. So you create a set with a range of exposures, contrasts and colours that exceeded the best capture device (film for now)- then augment the colour chart with some grid spot metering for light and colour to list what was actually there to be captured. Now tell the teams to get as close to this target as they can - and see you in a couple weeks.

In a perfect test, the results would be viewed on a calibrated monitor with the actual set off to the side for visual comparison to reality as well as each other.

Maybe in my ignorance I am missing something but I personally would be far more interested in the results of this test than either the ASC or BSC tests as I hear they were conducted. It strikes me as being more indicative of what I could get from a selected camera system if I really tried rather than what I would get if I put the lowest common effort into it.

With the plethora of workflows for some of the cameras (Red particularly) it may be best to have a number of alternative capture and process methodologies?

Hey, how about a permanent set that is tested and calibrated for consistency on a regular basis that is open to people for testing purposes? If anyone has a contention with a tests results they could book time on set themselves.

Hey Jim, if you're listening, that might be something that you would want to build at the Red Ranch. It would be a consistent test bed for charting your progress with new sensors and cameras as well as a place to fine tune workflow... and a ring where any other camera system could throw down the gauntlet.

 

[David Mullen ASC]

If you sent a bunch of different motion picture color negative stocks of different ages to a lab, Fuji or Kodak (or really old Agfa) they all get processed the same, ECN2, normal time & temp, unless you request pull or push processing, or skip-bleach processing.

And then you'd see the various inherent differences in the stocks, and how much they've aged.

The labs are all supposed to follow Kodak's LAD system for standardized results and the most modern stocks have been designed to be less sensitive to misprocessing.

That said, the results from each lab can have some minor variations (and major if they screw up) based on various factors (age of chemistry, cleanliness, I suppose even the water purity and circulation can have an effect.) But basically your film gets stapled onto the thousands of feet of other people's film and it gets processed at a very high speed, so it's not designed for individual attention.

Now combine this with the ECP print processing system, which is also standardized but has some mild variations (plus there are two Kodak print stocks, two Fuji, and an Agfa print stock, all using the same process) plus however it gets timed and you may feel that the image looks different at different labs, but it's hard to know if that's happening at the negative or positive side. It's hard to just look at a negative and "see" the characteristics of the image, color and contrast -- it has to be printed, or it has to be transferred to video, or it has to be scanned and switched to a positive image, so that adds more variables.

The trouble with not trying to match color and contrast of different digital cameras to each other, or to film, is that many people can't get beyond the mismatch. For example, if you took the Red test to an entirely different facility and "optimized" the look, but when put up against a shot from an F35, it looked (in comparison) slightly more blue and slightly more pastel and slightly more contrasty... some people would say "hey, the Red camera produces blue, pastel, contrasty images". Then you'd have to explain to them that the shots could have been timed to match better, but weren't. But they'd never get it out of their heads that the Red image was blue, pastel, and contrasty, or that the F35 was the opposite, when neither was true.

 

[Tony Brown]

With the plethora of workflows for some of the cameras (Red particularly) it may be best to have a number of alternative capture and process methodologies?

Hey, how about a permanent set that is tested and calibrated for consistency on a regular basis that is open to people for testing purposes? If anyone has a contention with a tests results they could book time on set themselves.

Clint. The BSC site has been stored and will be used again for subsequent tests

50 people were involved in the BSC tests for one week, I've heard (from the BSC) the ASC tests involved 200 people(?)

A number of alternative capture methodologies wasn't required as each camera was set up, not by the BSC but by the manufacturer supplied techs that came with the cameras to optimise the capture process

As for 'process methodologies' it is beyond practicality to do this. Even with the fixed workflow the results have taken months to produce, relying on grading time over a period of time simply due to personnel / hardware availability)

If RED (or Panasonic or .........) feels the need to demonstrate various workflows then I'm sure they'll do so. Or if embarking on a project its surely not beyond a serious productions capabilities to shoot some footage and test test test.... like we've done for years on film. But that wasn't the purpose of these tests. The purpose was to find a common flow and optimise the Macbeth chart within the sphere of that flow. Sue Gibson repeated this again and again. This was a single set of parameters applied, to have done more would have been wonderful. But it would take a year to do and week to watch.

I happen to disagree (as it seems most people do) with the BSC in so far as I would have preferred the grade to optimise the skin tones, not some poxy chart. However thats not what they did, so the audience has to be grown up enough to observe the results and use the information in any way they see fit. If only to conclude, that for them, the tests were flawed.

Personally I've worked with 35mm for 30 years, I've worked with 16mm maybe 10 times (I hate it), I've done one job shooting 2 x RED camera (which I enjoyed and would not have known it wasn't film.....after a lot of work) and I've shot with a Panny 2700 and an Iconix on a viral a couple of weeks ago. So for me and I suspect the vast majority this was extremely informative.

The audience contained people who refused to see the potential of anything other than film and the film is dead brigade who tried to defend the digital capture.

Didn't care. I was there for me. I learned a lot.

For posters to blindly criticise these tests on the grounds of hearsay and that link to misleading footage of the tests shot from the auditorium merely weakens any argument by showing some blind desperation to discredit results that happened not to go a certain way. Which ironically is exactly what the same people seem to be accusing ASC/BSC of.

 

[Tony Brown]

David,



Even BSC test directly "told" me that my shooting talents or actress/actors would get a 'true pay-tan' in skin tones if I shoot with RED ONE.

So I tried to "defend" my purchase decision and to "roll" on here some true facts.

The test “told” you nothing of the sort, you are being deliberately belligerent. The test showed that the chosen workflow did not work for RED. Period. It also didn’t work for SI 2K or the Canon.

That taught me something. That in itself was good information.

You have nothing to “defend”, but your defensiveness does you no favours. I doubt there was anyone in the audience that wouldn’t recognise the incredible potential of RED (especially with the latest builds) for the price point. You have to accept that intelligent people would pick the bones out of the tests and be inspired to learn more.

The Canon actually looked less like film than RED, a different workflow would OF COURSE improve that aspect of it, but the results I saw looked FANTASTIC (to me). It had a high contrast 3 strip look that I’d love to apply to a commercial. So even in the ‘flawed’ results I saw something good.

This was not a competition Sanjin. RED was not ‘attacked’

I strongly suggest you see the tests before you comment futher

 

[Sanjin Jukic]

The test “told” you nothing of the sort, you are being deliberately belligerent. The test showed that the chosen workflow did not work for RED. Period. It also didn’t work for SI 2K or the Canon.

That taught me something. That in itself was good information.

You have nothing to “defend”, but your defensiveness does you no favours. I doubt there was anyone in the audience that wouldn’t recognise the incredible potential of RED (especially with the latest builds) for the price point. You have to accept that intelligent people would pick the bones out of the tests and be inspired to learn more.

The Canon actually looked less like film than RED, a different workflow would OF COURSE improve that aspect of it, but the results I saw looked FANTASTIC (to me). It had a high contrast 3 strip look that I’d love to apply to a commercial. So even in the ‘flawed’ results I saw something good.

This was not a competition Sanjin. RED was not ‘attacked’

I strongly suggest you see the tests before you comment futher

Tony,

thanks for your kind explanation.

At the end of this "useless" discussion I would like to repeat once more what I wrote before in this tread:

ASC and BSC have to precisely elaborate all the tests with written and publicly open "Technical papers" and then we could see where the thing have got wrong with RED1.

Also I do not make any further comment either.

Just it's not make sense.

This summer I don't have plan to visit UK (I often meet Brits in Ibiza) and maybe Blu-ray disk

about the test could be pretty good option for all of us to have it in archive.

 

[Tony Brown]

ASC and BSC have to precisely elaborate all the tests with written and publicly open "Technical papers" and then we could see where the thing have got wrong with RED1.

They did. everybody who attended was given a pack containing more information than I could absorb on the day. I have it here. If there is anything specific you'd like to ask I'd be glad to try and answer it from the papers

 

[Antoine Fabi]

This test just reveals what we already know.

Film has a long history and can look superb just out of the box.
RED is very very young, and it takes more work and "expert" knowledge to look superb. But yes RED can look superb.

No surprise here i think...


Antoine

 

[Hrvoje Simic]

Measurements were taken on a scope from REC 709 colour space proxies

I don't understand this part...why that route ?

Was the point in establishing proxy quality, 709 space conversion or RAW camera footage ? What am I missing here?

Prior color science had issues with 709 and that was widely known.

 

[Kevin Liu]

Thanks to Neil and Tony for answering my question. I am someone hoping to expand film from a minor hobby to something that is either major or a career (I'm currently a college student) and I appreciate you guys taking time to reply. Unfortunately, since I live in the US and the UK is not generally accessible to me, I guess I won't be able to see that footage.

Regardless, I'm curious to know whether to S35 Scarlet will have the capabilities of getting a strong "filmic" image (albeit with its own personality, of course - I wouldn't expect it to perform in the same manner as 35mm), so I figured the Red ONE would be a good place to start - and after several months of browsing these forums and looking at test footage and sample footage, I'm fairly convinced that Red is difficult to beat in terms of price/performance ratio.

Right now I'm trying to absorb as much information as I can and hopefully get some experience, although it's difficult without a camera. Can anyone recommend cinematography/color grading books geared specifically towards digital video? I'm already looking into the Steve Hullfish book. Sorry for the long post!

 

[David Rasberry]

Thanks to Neil and Tony for answering my question. I am someone hoping to expand film from a minor hobby to something that is either major or a career (I'm currently a college student) and I appreciate you guys taking time to reply. Unfortunately, since I live in the US and the UK is not generally accessible to me, I guess I won't be able to see that footage.

Regardless, I'm curious to know whether to S35 Scarlet will have the capabilities of getting a strong "filmic" image (albeit with its own personality, of course - I wouldn't expect it to perform in the same manner as 35mm), so I figured the Red ONE would be a good place to start - and after several months of browsing these forums and looking at test footage and sample footage, I'm fairly convinced that Red is difficult to beat in terms of price/performance ratio.

Right now I'm trying to absorb as much information as I can and hopefully get some experience, although it's difficult without a camera. Can anyone recommend cinematography/color grading books geared specifically towards digital video? I'm already looking into the Steve Hullfish book. Sorry for the long post!

Although it is offline for the time being, picking some .r3d test clips off of Redrelay.net and playing with them in RedAlert or Redcine software, available for free download from Red.com is a great way to begin to learn about the grading capabilities inherent in red code raw files. Watch for the relaunch of redrelay.

 

[Kevin Liu]

Although it is offline for the time being, picking some .r3d test clips off of Redrelay.net and playing with them in RedAlert or Redcine software, available for free download from Red.com is a great way to begin to learn about the grading capabilities inherent in red code raw files. Watch for the relaunch of redrelay.

Thanks! I knew about the programs but was having a bit of difficulty finding them (although I'll admit I wasn't looking too hard). I'll see if I can pull some footage from the Shot on Red section and try my hand at some basic color correcting.

 

[David Rasberry]

Thanks! I knew about the programs but was having a bit of difficulty finding them (although I'll admit I wasn't looking too hard). I'll see if I can pull some footage from the Shot on Red section and try my hand at some basic color correcting.

They are in the support section.

 

[Gavin Greenwalt]

There's also some footage on the scarlet user forums.

 

[Michel Hafner]

I have seen both the ASC and BSC test. I wonder about some things.
- ASC test: Bulb shot. Would the clipping highlights on the face be less clipped when the footage had been graded from RAW directly? If the exposure had been chosen to avoid all clipping would the background be very noisy? Would Firmware 20 give better results? Does the dynamic range of the shot exceed Red's capability for now and you will have either clipped highlights or noisy shadows?
- ASC test: Jaws shot (2K). Why did the Red have less spatial detail and lack of spatial resolution compared to film and some 1080p cameras? Recode compression? Bad downsampling from 4K? Suboptimal focus?
- ASC test: Fire place shot: Same problem. Where did the detail go? And the color problems in that shot, are they gone with Firmware 20 and grading from RAW?