Keep video and film seperate please!

Greg Lowry said:
I was, in fact, making a bit of joke for the benefit of the readership here. Geez. Depth of field is the correct term, not depth of focus. Deep focus would have been a better layman's term. It's also an industry term. Sometimes movie reviewers and industry writers throw in terms they think they understand when they don't.

Yeah, I got your joke immediately - and it was a good one! Obviously depth of field is the correct term. Deep focus may have been a better laymman's term, but depth of focus would mean about the same thing in the brain of most layman readers of the NYT. They wouldn't know or care that deep focus was an official term in our industry. Scott writes a fair amount on the media industry. I suspect he knows that depth of field is the correct term, but was simply trying to clarify to media-ignorant readers that focus depth was what Mann and Beebe were after.

Gibby
RED #8
www.cut4.tv
www.4umat.com
 
OK, now I've seen Miami Vice. It certainly isn't a deep focus movie, that's for sure. Selective focus is used in a fairly conventional way. Composition, on the other hand, is decidedly ... odd. They use the increased depth of field of the 2/3" format to shoot in very low light (sometimes very, very, VERY low light) while still having enough depth to see backgrounds. Sometimes the images look like grainy, underlit S16, sometimes like slick 35mm. Whatever serves the story is fine by me.
 
Last edited:
I just came back from a screening of it too....not with Greg, although he is a nice fellow...
I thought it was strange how there was almost no set rules in the context of scenes....there is all types of differing noise levels and seeming shutter speeds and even the way some of the video is rendered - almost 30p in some spots. Some of the stuff looks available light....with nothing used as a key...some of the open ocean stuff in particular is extremely noisy, but I must say that overall the imagery works in the vein of the story.
I wouldn't say the movie is necessarily beautiful, but the dirty, grungy and artificial images closely relate to the thematic elements of Mann's story. It certainly never looked, or even tried to look, filmlike...some of it even started to pull away from being cinematic in my opinion, but I think this goes towards what a lot of people have been talking about - the medium is cinema, but the format is digital...and this is definitely an emphatic embrace of that type of visual storytelling.
 
Last edited:
It's not as visually stylized as, say, Tony Scott's most recent movies have become, but as you point out there are no discernible rules from scene to scene or even shot to shot. It's always interesting to see a major director taking chances, even if he's not always successful.
 
Last edited:
Agreed - certiainly not as pyrotechnic in terms of slow motion, layered dissolves, and jump cutting. I would say though that I think it takes balls to make a picture that looks so unconventional in terms of a hollywood action picture of a known property. At least Collateral was almost a character piece with a limited cast. Miami Vice was "MTV Cops" for the time, and I think it is a risk to turn out something so rough in its execution that has as big a scope as it does...has Miami ever looked less like a tourist spot?
 
It's funny you mentioned Miami vice Greg. I was going to start a thread on it , the Viper and RED.

So, I went to Vice to see how Mann handled the Viper, and how its images came out on the silver screen. Now, to preface all my remarks I know that Mann's production had the BEST equipment, lenses and post production workflow All that being said, they did have to use a camera that is comparable (spec-wise) to RED. Hence, the question that I had throughout the movie, mainly, Will RED perform at the same level or better than the Viper ?

I've looked over the specs on both cameras (that we know) and unless I am missing a tech point (which is possible), I think it will. RED has a S/N ratio of 66db compared to Viper's 54db has over 11.4 million pixels, versus Viper's 9.2 million. And although I don't know how the 3 2/3" Viper's CCD's compare to the one Super35mm CMOS of RED, I can't imagine a huge advantage. Now, as to the differences in RAW output, and/or RED code, it seems a little early to do comparisons, but all else being equal I don't see an advantage that the Viper has over RED.

Thus, to answer that long comparison, "yes" RED could achieve a similar picture as Miami Vice (given 35mm primes, good DP etc...). In fact, i have a Miami script (written awhile ago) that i had stuck in a drawer due to financial considerations. However, if RED delivers on the dynamic range, color fidelity, and amount of useable data, that they are claiming, then it probably will be a different story. Particularly when you don't need 3 gennies and three trucks worth of lights to film in low light scenarios (not saying that you "don't" have to light, just light "differently" and less expensively)...

IMHO, there were some gorgeous shots in Miami Vice that were equal to or better of 35mm. Granted, the night exteriors were gritty, but as aforementioned, they suited the story and thhe aesthetic Mann was going for.

For the rest of the movie, i'll head to the cafe'. But if anyone thinks i've missed something in my estimation i'd be glad for you to chime in :beer:

RED # wish I was higher
 
Last edited:
As I wrote in another thread, the Viper uses a system of smaller sub-pixels that account for the 9.2 million pixels you refer to. This is a misleading number. The actual spatial resolution of the Viper sensors is just 1920x1080 at full resolution. As you know, RED's maximum spatial resolution is 4520X2540 on an S35-sized sensor. It would have been impossible to achieve the same result on Miami Vice from a larger sensor using S35 format lenses.

Here's an excerpt about the Viper from the other thread:

It should be noted that the Viper sensors do not, strictly speaking, each have 9.2 million pixels. The Viper uses an ingenious sensor design with 4320 small vertical "sub-pixels" and 1920 standard sized horizontal pixels which allows the sensor to be electronically configured for optimum 720p and 1080p in 16x9 and widescreen mode (2.37:1) without cropping the sensor, thus maximizing spatial resolution for all formats. Not all of these sub-pixels are used for every format. The maximum number of full-sized pixels is 1920x1080. The 9.2 milllion number is actually: 1920 pixels x 4320 subpixels = 8,294,400 effective pixels (really a combination of sub-pixels and full-sized pixels). The 9.2 million number is really a marketing (or engineering) number more than a practical one. I don't know why Thomson Grass Valley uses it because it can result in confusion.
 
Last edited:
Hmm, bare with me Greg, but does that mean that RED will outperform the Viper (since RED uses 4520x2540 of its active pixel array) ?

And doesn't Vipers 3 chip design give it an advantage since it can capture RGB separately?
Perhaps, they are on an equal footing if we consider only onboard recording 2K scaled ?

edit:
It would have been impossible to achieve the same result on Miami Vice from a larger sensor using S35 format lenses.
I don't quite get this. What about using 2/3" lenses or 16mm on RED?

RED # wish I was higher
 
Last edited:
I think it's better to think of the Viper and RED as different tools with different strengths.

RED should be capable of higher image quality due to the increased spatial resolution of the sensor. We can probably also expect that the RED will boast better performance with respect to sensor speed (light sensitivity) and dynamic range, as the Viper uses a five year old sensor technology. As for 3 2/3" CCD v. single single CMOS with Bayer pattern, based on experiences with the ARRI D20, it appears a single sensor outperforms 3 x CCDs when the full sensor is used but the difference isn't as great as some might expect. Will a 2/3" or S16 windowed RED sensor with Bayer perform as well or better than the best 2/3" 3 CCD cameras? I don't know. Nor does anyone else. It'll be interesting to find out. Until we see the actual performance of the RED camera, it's all opnion, guesswork and speculation.

For Miami Vice, the 2/3" format and lenses allowed for a style that wouldn't have been possible with 35mm film or large sensor digital. The technology appears to have very directly impacted the style, just as it did for Mann's Collateral.
 
Last edited:
Greg Lowry said:
Until we see the actual performance of the RED camera, it's all opnion, guesswork and speculation.

For Miami Vice, the 2/3" format and lenses allowed for a style that wouldn't have been possible with 35mm film or large sensor digital. The technology appears to have very directly impacted the style, just as it did for Mann's Collateral.
Lol. That's part of the fun. :)

Seriously though, these discussions have tangible results for me, regarding allocation of budget, possible pre-production on projects etc ... but I do understand that there is only so much comparison that can be done without real world testing and verification.

I don't quite understand why RED wouldn't perform similarly to the Viper if you put high quality 2/3" lenses on it. Perhaps the fact that it would only be 4:2:2 1080p.

I'll tell ya though, the HVX vs. HDV cameras shootout was technically comfortable to understand, but the RED vs. Dalsa vs. Genesis vs. Viper one is going to be a trip. I guess some of these suppositions will be run up the flag pole (as it were).
IMO, a definite test should be the Viper in videostream mode vs. RED 2K scaled w/35mm lenses and RED 1080p w/2/3" lenses. I know it's only 4:2:2, but i'd still like to see a comparison.

PS - I don't need the incredible depth of field Mann was trying to achieve (which he did marvelously btw, ex's. clouds at night, city lights atmospheric bounce), just the low light capabilities. :beer:

:dankk2: for the help Greg. Also, do you think it's worthwhile to read some of the workflow papers on the Viper, in preparation for RED ?



Cheers.

RED # wish I was higher
 
Last edited:
D_and_G said:
I don't quite understand why RED wouldn't perform similarly to the Viper if you put high quality 2/3" lenses on it. Perhaps the fact that it would only be 4:2:2 1080p.
My doubts (such as they are) have to do with the quality that's achievable from a 2/3" format section of the RED sensor with Bayer filter v. a 2/3" 3 x CCD camera. I think the only camera using a single 2/3" sensor with Bayer filter is the Silicon Imaging SI-1920HDVR camera, but I haven't seen the results. The reports I've read are favorable, but I'll need to see to believe.

D_and_G said:
PS - I don't need the incredible depth of field Mann was trying to achieve (which he did marvelously btw, ex's. clouds at night, city lights atmospheric bounce), just the low light capabilities.
With some fast lenses on the RED you should be able to turn the lights off and shoot. :)
 
Last edited:
D_and_G said:
I don't quite understand why RED wouldn't perform similarly to the Viper if you put high quality 2/3" lenses on it. Perhaps the fact that it would only be 4:2:2 1080p...I know it's only 4:2:2, but i'd still like to see a comparison.


4:2:2 colorspace isn't the only proposed spec for RED in 1080p. As verified by Stuart English and Graeme Nattress, Red may be able to record 1080p 4:4:4 RGB internally, and also may be able to output 1080p 4:4:4 RGB via HD-SDI. If those projected specs stand, a 4:2:2 colorspace will not be your only option in 2k, 1080p, and 720p. You could also shoot 2k and either crop it down to 1080p 4:4:4 RGB in camera before you send it out via HD-SDI, or crop it in your NLE. Bottom line: if those specs are finalized, 1080p 4:4:4 RGB via RED with quality 2/3" lenses, may compare nicely with the results from the Viper - if not surpass it. It will be interesting to see...

Gibby
RED #8
www.cut4.tv
www.4umat.com
 
Last edited:
Regarding using 2/3" lens and center cropping to 1080p. Yes this will be very like shooting the Silicon Imaging camera which I have shoot with (read about that and see the HD results here : http://cineform.blogspot.com/2006/07/silicon-imaging-in-48-hours-success.html.) But technically bayer center cropping is not quite 4:4:4, as bayer image doesn't have every primary per photo-site. Although you can extract a surprising amount of detail, but it will not quite reach the level of a Viper.
 
That's good to know Gibby.

I know all these specs aren't finalized, so anything is possible. I guess, I just wanted to reiterate how excited I was seeing "Miami Vice" on the big screen, knowing that RED (if it delivers) plays in the same "ballpark" as these other heavy hitters.

This was the first movie for me, that married the artistic + technolgical in the digital realm, with an acceptable film out image. (for others, it may be an earlier digi movie) The fact, we as consumers have the chance to own this gear is something i've been waiting for.

Now for Ibc :beer:

PS - Even the preliminary specs are revolutionary. Before one change. Indeed, even if Jim delivers a Ferrari Spyder instead of an Enzo, i'll be chuffed :beer: :beer:


REd # wish I was higher
 
Thanks for that insight David...

D & G: I haven't seen Miami Vice yet, but I'll check it out in the next few days. My wife always ribs me about not just enjoying a movie, rather than analyzing the production values of it. Like the rest of you, I can't shut off the dual-track brain system of simultaneously enjoying the story and evaluating the tech/artistry values. Watching more footage is like a postman taking a walk on his day off, but we all love what we do, so it's a pleasure to check out someone elses work!

No doubt...we'll all be stoked to get our hands on that new Ferrari, whatever the model, and take it for a roadtest...

Gibby
RED #8
www.cut4.tv
www.4umat.com
 
Last edited:
The camera of choice for Michael Mann was the Sony F900 ,when he made Collateral. Don't know about Miami Vice. I,m hoping one could pull off the same low light capability with RED! I would be very pleased
 
Back
Top