A quick look back at the Sony HDW-F900

[Karim D. Ghantous]

In and of itself, the F900, released in 2000, is a very good camera. It is a 3-chip camera for clean colour. It has a CCD which, I presume, is a global shutter by default. Base ISO is around 300. Frame rates are limited to 30fps progressive, and 60 field/sec interlaced. It's only an HD camera, so today you might now want such a large body just for that. Oh, and it records to tape. For those who don't know, it was used to film Star Wars episodes II and III. So no 4K masters exist for those movies, although the VFX could be redone to 4K I am sure.

IMHO, George Lucas went backwards, not forwards, by using these. 35mm was unmatched until the Red Dragon, released in 2012. The old Alexa and the Epic 5K cameras were great but not a match for Vision 3. V-Raptor is another level, but that's another discussion.

https://cinequipt.com/cms-files/sony-hdw-f900-brochure.pdf

 

[Brian Timmons]

In and of itself, the F900, released in 2000, is a very good camera. It is a 3-chip camera for clean colour. It has a CCD which, I presume, is a global shutter by default. Base ISO is around 300. Frame rates are limited to 30fps progressive, and 60 field/sec interlaced. It's only an HD camera, so today you might now want such a large body just for that. Oh, and it records to tape. For those who don't know, it was used to film Star Wars episodes II and III. So no 4K masters exist for those movies, although the VFX could be redone to 4K I am sure.

IMHO, George Lucas went backwards, not forwards, by using these. 35mm was unmatched until the Red Dragon, released in 2012. The old Alexa and the Epic 5K cameras were great but not a match for Vision 3. V-Raptor is another level, but that's another discussion.

https://cinequipt.com/cms-files/sony...0-brochure.pdf

Had a bittersweet encounter with that camera when filming with it for the very 1st time in early days of my filming adventures.
Was told we were filming with the F900 last minute and did not have time for prep with the camera (mistake in its own right).
Filming was with 2 well known artists in their respective genres.
I'll just say the settings went to shit (likely user error) and we had to get the settings back to what we had before lunch.
Going through the rather archaic menu system was a technological existential crisis.

Barely made it through the shoot but soldiered on to the finish.

Director was pissed and I had to do another 12 hours graveyard shift in post production on another gig.
Fond memories indeed.

Months later and with a bout of insomnia, I happen to see the end result in an early morning airing of the project.
Shoot came out well. Another example of the drama audiences don't see behind the scenes and likely don't care to know.

The F900 was from a pretty primordial era in digital cinema but I respect the impact it had on the industry's digital future.

I believe Star Wars Episode III was filmed with the F950 which I think had better color sampling (4:4:4 vs 4:2:2 in the F900).

IIRC, DP M. David Mullen who used to frequent RU back in the day used the F900 for the film "Jackpot".
That film was one of the 1st I remember reading using it.

To this day though the F900 had the most densely layered menu I've ever encountered.
In many ways it was a highly sophisticated ENG camera than what we understand as a digital cinema camera.

This was a period where image acquisition was limited by video tape formats, no RAW, and barely log or LUTs.
Menus full of various types of gamma, knee, pedestals, color matrix among others
but it went much deeper than other cameras of the day (or likely any day for that matter).
Concepts of UI/UX for menus? Yeah whatever.
DIT was a position largely made to tame that camera as you really needed a video engineering background (and perhaps access to Sony engineers)
to fully understand it.

George Lucas had his reasons for using the F900 as it fit much better for his then new digital workflow
marrying a heavily green screened production like Star Wars with CGI (i.e characters, virtual sets).
Remember scanning film was quite expensive and Lucas was possibly not just thinking like a technologist but like a frugal entrepreneur (he paid out of pocket for those films).

Interesting that Sony was supposedly reluctant to make a camera that filmed 24 fps and Lucas had to push a bit to convince them to make it.

I'm sure there's many, many stories out there about that camera.
It was influential and would definitely be named as a major technological step in the evolution of filmmaking.

Brian Timmons
BRITIM/MEDIA

 

[Phil Holland]

Whelp, this is one area I can contribute some information worthwhile.

I turned down working on Episode II, but it was one of the harder career decisions at the time for me.

I won't speak for George and co., but I do know what was going on in a big way. At a time when 300-600 VFX shots were "a lot" for most films, Episode II had somewhere in the realm of 2200 or VFX shots. Production began in 2000. This was a very, very different world compared to now. The digital workflow allowed for a faster filming schedule and easer pathway to seeing what was shot rather quickly to adapt, change, or move on.

Brian is correct in saying that digital scanning, part of what I did at the time, would have been an absolute monstrous task. Using our 4K scanner at R&H at the time would have been around 26 or so seconds a frame to scan. That would be a all hands on deck year long journey just on the scanning side at the time. The industry didn't get faster or better scanners for a bit there. We eventually went to the Northlight 6K, which was significantly faster, but came along much later.

This and Episode III were moments in our industry as we were digesting what a digital workflow would be or could be. George paved that road by being an early and energized innovator, which is par for the course with ILM during that era. I can't stress enough that during that time when preproduction, production, and post production were far closer together, the minds and efforts of the top 3 VFX houses provided a lot of invention, ingenuity, and innovation shaping the industry along the way.

And just my general POV from somebody who was in the absolute thick of it when it comes to RED releases. RED One, people laughed at the announcement, but also took extreme notice when cameras shipped and footage was shown. Epic with Mysterium-X turned heads by doing many things that the RED One wasn't capable of or ideal for. It became the 3D darling as well with it's at the time radical modular body design. Dragon sensor technology for me was the most important sensor up until that point as it provided much more juice in the color and texture department. I know first hand by the many film v. digital tests I even shared here, but more importantly what was going down at studios or even in parking lot tests was really pushing many film cinematographers into a new era.

From there, in my mind, all of RED sensors have been noteworthy and building upon what got them here. But those camera designs and sensor technologies were the big ones. I think it's probably worth noting the impact of high volume sales and popularity of Komodo having an impact not only Komodo-X, but Raptor bodies. Though different from DSMC2, the DSMC3 bodies do a lot in a very, very small footprint.

 

[Brian Timmons]

Cool story Phil.
I figured you would have a bit of input on this given your background at R&H.

Saw this BTS which is directly from Lucas and company.

 

[Karim D. Ghantous]

Was told we were filming with the F900 last minute and did not have time for prep with the camera (mistake in its own right).

I made that mistake once, and it was 100% my fault, as it was my first wedding (I don't do a lot of those FWIW). At least you had someone to blame!

That video you posted says a lot. It backs up my opinion, despite what Lucas was saying about progress etc.

W
Brian is correct in saying that digital scanning, part of what I did at the time, would have been an absolute monstrous task. Using our 4K scanner at R&H at the time would have been around 26 or so seconds a frame to scan. That would be a all hands on deck year long journey just on the scanning side at the time.

According to my calculations, if it takes 26 seconds per frame, you are looking at a total scan time, post-editing, for a 3 hour movie, of 78 days. This would no doubt slow down the VFX artists, would would have to wait maybe 2.5 hours to scan one minute of footage. But that's with just one scanner. Add a second scanner and you halve your scan time. Add three extra scanners and your 3 hour movie now takes 20 days to scan.

Hypothetically, it might have been better for Sony to develop a faster film scanner? I'm thinking aloud, from an armchair.

 

[Phil Holland]

But that's with just

I'll just lightly reiterate that this was an entirely different time in digitizing and recording film. There we're a lot of quality digital scanning tools back then, hence the reason the 4K scanner was invented in the mid-90s. I know when we switched over to the Northlight we looked at other options from Imagica and so on. To put it lightly, our Northlight setup was upwards of $600,000 with a a yearly service plan and additional costs surrounding our particular setup, which likely brought it into the $1 million territory. Similar story for our laser recorders and CRT based Solitaires, then eventually switching to the ArriLaser.

During this era there were less than a handful of houses that could do this caliber of work or even attempt it and to do it in this quantity was an additional layer of complexity.

I should also underline that editorial, additional departments, as well as labor represented additional costs.

Which actually was part of the reason you'll see many VFX houses contributing to a single film's VFX during this time. Films with demanding VFX and scheduled release dates requires some creative thinking as creation, iteration, and quality control on a quantity of shots with many teams pulling off what's in frame takes a great deal of time.

Top VFX dailies and review generally consists of watching shots very closely with a team over and over again to change and refine things until it reaches a very high level of finish. Outside of production dailies, VFX dailies didn't need to be filmed out, and digital projection was just happening as well.

Circling back to some interesting relevance, as digital was coming along, there was tremendous pushback as well as pushing forward during this time. Our Scan and Record Department transformed into Digital Imaging Resources. Which took care of digitizing film, exporting to film, digital scanning, color correction, color grading, various VFX element capture and acquisition in stills and motion. The majority of this occurred in about 6 years total and it was actually extremely quick in terms of innovation and deployment. That span of time alone was being involved with about 100 feature films in some way shape or form. A lot was happening, changing, and transforming.

I have upmost respect for Lucas committing so hard and it's tremendous that this occurred so many times for the Star Wars franchise in particular. Though those films couldn't take all the credit, our whole industry has changed due to much of the creative and technical innovations required to bring Star Wars to life. And thankfully they were successful enough to benefit subsequent films and filmmakers desires as much of the technology was used or built on to go further.

 

[Phil Holland]

Cool story Phil.
I figured you would have a bit of input on this given your background at R&H.

Heh. That BTS gives me some PTSD. But it certainly outlines how different shooting digital was in comparison to the film workflow. It's actually amazing what they achieved visually in terms of the quality of everything.

It is nice the nods to the evolution of digital. I mean look where we are now with camera hardware.

 

[Marc Wielage]

George Lucas had his reasons for using the F900 as it fit much better for his then new digital workflow marrying a heavily green screened production like Star Wars with CGI (i.e characters, virtual sets). Remember scanning film was quite expensive and Lucas was possibly not just thinking like a technologist but like a frugal entrepreneur (he paid out of pocket for those films).

George was extremely angry at Kodak because the original version of 5247 negative started fading and falling apart, which effectively ruined Star Wars, The Empire Strikes Back, and Return of the Jedi. We spent months restoring them in 2004, and the feeling then was the negative was almost on its last legs at that point. Lowry Digital did the lion's share of the restoration work in 2003-2005, and we did the color up at ILM at San Rafael (right before the move to San Francisco).

Lucas told me he felt "betrayed" by Kodak, who had promised him that 5247 was going to be every bit as reiiable as the previous emulsion, 5254... but a lot of cinematographers preferred the older stock. Eventually, Kodak got a handle on it and the film improved, but the damage had already been done. I don't think the Sony F900 or the F950 were good enough, but there weren't a lot of camera choices in 1999-2005.

I don't think Lucasfilm could have waited much longer to shoot the Star Wars prequels, because he was already in his late 50s/early 60s by the time those films were made. That was one of the factors for handing off Lucasfilm to Disney in 2012 -- he didn't want to devote ten more years of his life to working on the sequels.

 

[Brian Timmons]

George was extremely angry at Kodak because the original version of 5247 negative started fading and falling apart, which effectively ruined Star Wars, The Empire Strikes Back, and Return of the Jedi. We spent months restoring them in 2004, and the feeling then was the negative was almost on its last legs at that point. Lowry Digital did the lion's share of the restoration work in 2003-2005, and we did the color up at ILM at San Rafael (right before the move to San Francisco).

Lucas told me he felt "betrayed" by Kodak, who had promised him that 5247 was going to be every bit as reiiable as the previous emulsion, 5254... but a lot of cinematographers preferred the older stock. Eventually, Kodak got a handle on it and the film improved, but the damage had already been done. I don't think the Sony F900 or the F950 were good enough, but there weren't a lot of camera choices in 1999-2005.

I don't think Lucasfilm could have waited much longer to shoot the Star Wars prequels, because he was already in his late 50s/early 60s by the time those films were made. That was one of the factors for handing off Lucasfilm to Disney in 2012 -- he didn't want to devote ten more years of his life to working on the sequels.

Quite an interesting story Marc.
I heard the Star Wars negative was in bad shape but didn’t realize it was that extensive.

I think at a certain point Lucas was willing to compromise
on image quality for the other advantages digital would give him (i.e compatibility with workflow, perhaps longevity, among others).

I think Lucasfilm’s greater attitude toward image acquisition was that it ultimately would be digital data they could mold to what they wanted in post/VFX.

Would have been interesting to see what Lucas would have accomplished if he did the latter 3 episodes of the trilogy.
Disney really….. nevermind.

Brian Timmons
BRITIM/MEDIA

 

[Karim D. Ghantous]

Would have been interesting to see what Lucas would have accomplished if he did the latter 3 episodes of the trilogy.
Disney really….. nevermind.

They were very expensive fan films.

 

[Brian Timmons]

They were very expensive fan films.

You said it.
I was done after Ep. 7.
This was a film series that truly influenced me getting into filmmaking. Was disappointed to see it reduced to a
corporate product line and a formula.

Even post apocalyptic psychopaths have occasional words of wisdom.

Brian Timmons
BRITIM/MEDIA

 

[Marc Wielage]

Quite an interesting story Marc. I heard the Star Wars negative was in bad shape but didn’t realize it was that extensive.

We were told that Technicolor/Tokyo had scratched the living hell out of the original Star Wars negative in the 1970s during a re-release. It's hard to believe that Fox was so cavalier with the original negative back then. It is fixable, but it's by no means in as good a shape as Empire or Jedi.

I think at a certain point Lucas was willing to compromise on image quality for the other advantages digital would give him (i.e compatibility with workflow, perhaps longevity, among others). I think Lucasfilm’s greater attitude toward image acquisition was that it ultimately would be digital data they could mold to what they wanted in post/VFX.

Yeah, what Phil says above is correct: the time involved to scan 2000+ film VFX shots would've been crazy. I think their thinking about using (relatively) low-res HD cameras for capture was that only part of the frame was going to be live action, and the rest was going to be digital paintings and composites and so on. They shot those pictures as large as they could in the frame to maximize available resolution, and David Tattersall did a stellar job lighting it. Years later, Cinesite reprocessed these early digital films, fixed some registration problems, a lot of noise problems, and sharpened the images a bit; even though it's still 2K, it's very good 2K.

Note that Lucas' prequels were all "relatively" low-budget ($115M or so), since he self-financed them and basically loaned them out to Fox for distribution. It's interesting to reflect that the later sequels done by J.J. Abrams and Rian Johnson were far more expensive (and two of them were shot on film). The workflow on movies at this level is staggeringly complicated.

 

[Brian Timmons]

Heh. That BTS gives me some PTSD. But it certainly outlines how different shooting digital was in comparison to the film workflow. It's actually amazing what they achieved visually in terms of the quality of everything.

It is nice the nods to the evolution of digital. I mean look where we are now with camera hardware.

I think this is the part that many in the now established digital cinema world don't get.

I know there are many on RU that get this but this is for those that don't.


Around the same time George Lucas was filming the prequel Star Wars trilogy I was on the other end of the indy filmmaking spectrum as a DP on 3 different features shooting on 16mm and Super 16mm.

It took an average of about 3 years to get the films shot not because of anything to do with high end FX but because the producers could only
afford to shoot the film piecemeal due to limited funds.

I would remark that I had monthlies and yearlies instead of dailies because it sometimes took that long for the producers to pay for the processing and tape transfer
before we could see what was shot.
Had a few sleepless nights wondering if the footage shot was even ok as locations were lost, actor's change their look etc.

We probably had no business making a film with those limitations but it was the only true option.
Digital video was not even 24p yet and the transfer techniques for field interpolation of 60i to 24P was not there yet.

10 minutes of 16mm film raw stock, development, and Beta SP tape transfer (Standard Def) could cost somewhere around $300 back then and that was just a basic one light.
Of course you can now buy a 1TB CF Express card for that money that can give over 2 hours of 6K material and the card is reusable.

This is not even talking about the advancement in post production with being able to own the edit system (Avid systems were still $50k
and Final Cut Pro and Premiere were still getting their feet wet).

Negative cut and film prints could easily run over $50k for one finished copy of the film. Some festivals and many distributors required at least a print.

All of this is to say that Indy filmmaking was a bit of a steep mountain to climb then but the technology advancements really made a huge difference
and a good part of that was through the efforts spearheaded by Lucasfilm, Sony, Panavision, Kodak, among others.

It has to be emphasized just how great the filmmaking options have expanded in the last 2 decades.

It also needs to be emphasized that (to me at least) RED specifically pushed acquisition in a direction that I do not think the rest of the industry was going in.

In the mid to later 2000s most digital cinema cameras seemed to be settling into 2K with 4:4:4 color space. Raw was not in most people's (including manufacturer's) conscious.
I really believed RED's actions with sensor tech and REDCODE is what pushed RAW and 4K into the film acquisition space.
Yes- Dalsa did exist and so did Arri's D20 (which could at least do RAW IIRC) but neither had anywhere near the ease of workflow that REDCode had and they were $$$$$$$$
that you couldn't even buy them.

It's the reason why I think RED should rightfully own the patent on motion picture RAW (compressed at least). They were the only company to really make it happen particularly at the unheard of price range.
In fact, I would go as far to say that the filmmaking world would take a considerable step back in innovation and price to performance if RED did not exist.
Not fanboying, just being real about what the company has historically offered in it's product lineup.


By 2010 to be able to shoot, edit, color grade a film using an affordable desktop computer on an relatively affordable camera (RED ONE) and play it back it from a Blu-Ray (which looked incredibly good even compared to pro format like HDCAM) was quite a big revelation yet it didn't happen overnight.

Brian Timmons
BRITIM/MEDIA