AVID Workflow vs TVC reality?

[conrad gaunt]

They have a valid point about long term digital storage, but they`ll come around when the "nano" drives arrive, and Discovery can store its entire collection on a dice sized disk.
As the saying goes "Only mad dogs and englishmen sit in the midday sun!", er, no, thats not the one, I mean "You can take a horse to a cafe, but you can`t make it drink a nice cup of tea".

 

[Lachlan Ward]

I stopped the Mac hate a little while ago. Although it must be sad, Apple are the most evil IT corporation on the planet. They're like Nazi Germany with less violence, but still similar! What I love about the workflow solution for Red is that for under half a mill you can get a huge massive big oversized killer Japanese robot. It wasn't that long ago that half a mill only got you a tamagotchi that you couldn't even color grade with.

Go mighty killer workflow robot go!

__________________
Don't you worry about workflow, let me worry about blank.

 

[I Bloom]

Bear in mind agencies (fair enough) care about archives and backups, they like the stack of film cans or shelves of digi betas in the vault mentality so they can re edit projects later, they don't like the idea of hard drives and memory cards.

jas

Ok,

While I agree that FCP is the way to go and well... I was walking away from the Apple store on Fifth Avenue, next to central park, with my Final Cut 6 Upgrade box in my hands a few months ago and a commercial editor literally got off his bike to stop and tell me I was making a mistake and that he cuts for such and such huge account and no real editor uses FCP.... True story.

I understand... Avid has a different kind of interface that leads to a different style of editing, just like shooting with primes instead of zooms changes how you place the camera. Fine. Don't switch from Avid and don't listen to anyone who tells you that Final Cut is better. That is none of their business. You're not a fuddy duddy because you cut on Avid.

But at the same time, adapt to tapeless workflow and Red workflow or perish like the Diplodocus.

Now get ready for a huge revelation. Nobody cuts for real today, on Redcode proxies in FCP. We downconvert them to ProRes Proxies and cut them on ProRes timelines. Downconverting proxies is easy and can be done on set faster than you can shoot, you can even incorporate onelight style CC if you are smart with RedAlert or diligent about your in camera metadata.

Cutting Redcode on a Prores timeline, or Redcode on a Redcode timeline is well... a world of hurt... today.

So.... find a way to downconvert your Redcode Proxies to DNxHD or find another quicktime exportable format that Avid can play nice with, bring that proxie footage into Avid and cut away... and then... export an EDL from AVID and conform your Redcode in Scratch. It's beautiful don't let anyone tell you that Scratch can't play with the big boys. (I know a guy in New York that can make that happen for you. Hint: his first name is Off).

Or if you can't do Scratch wait and there are some tools coming that will let you online from AVID through REDCINE and back or into another high end grading system. They are right around the corner.

MichaelP from Avid is on here every day, (check his awesome Avatar

In the mean time, don't wait to start climbing this learning curve. Oh and it's called LTO-3 and people who fear hard drives pull it out of the cartridges and wrap it around themselves for protection.

IBloom

 

[Dylan Reeve]

The post company we're dealing with is quoting 24hrs for every hr of footage to “process” a “best light” grade, create an archival backup and provide standard def rushes on a hard drive prepped into avid bins etc and an SR tape. added to that we're shooting out of town where traditionally you simply Telecine the next morning at sparrows fart and go straight to the airport with your processed neg and a digi beta ready to cut at home. With the Red workflow getting on the plane with rushes and neg under you arm the next day knowing all is well seems to be out of the question.

I know that Images Post in Auckland has a fairly solid workflow to Avid for clients, and I'm pretty certain they can do it in much better time than that. You could probably pop someone of a flight to Auckland with the media, and a harddrive and have them back sooner and cheaper than what you're talking about there...

The theory is fairly simple, but the main thing is practice. As I haven't had any, I can't really offer anything useful in terms of real workflow for you - but I understand your situation...

Now as for the FCP thing... I certainly won't deny that FCP has some great features and some good grunt, but it has some big failings in some pretty major areas... Metadata and Media Management, Ingest and Output functions, Trimming... There are quite a few things that I consider the 'serious' stuff that I find FCP really fails on. Avid isn't going anywhere anytime too soon, hopefully they'l pull their head out soon, and give their product line a bit of a kick, but in general I think in most cases Avid is a faster and more efficient editing tool. When it comes to finishing, there are some serious arguments in favour of FCP over Avid's offerings, but for serious cutting I still think Avid holds the lead, and will continue to do so.

In NZ (and Australia too I assume) Avid is still the standard, FCP suites have found their way into some facilities, but they tend to live in Cinetape and serve a very utilitarian purpose. That may change, but for the moment here we need to be able to cut on Avid if we want to cut something serious.

 

[Hans von Sonntag]

In NZ (and Australia too I assume) Avid is still the standard, FCP suites have found their way into some facilities, but they tend to live in Cinetape and serve a very utilitarian purpose. That may change, but for the moment here we need to be able to cut on Avid if we want to cut something serious.

Same in Germany. The classic TVC post boutiques in Munich, Berlin, Hamburg etc. are based on Avid and FFF / Quantel. FCP is a niche editor for some funky HD stuff nobody really cares about. ALL TVC stuff is PAL, NONE is HD. Shooting with F 900R / F 23 is coming but traditionally hated because it's video. Donwconverting 1080p to 576p is often done with the in-build hardware converter of the tape machine. Quality gets very poor with this kind on down convert but people say: Your bad, shoot 35mm.

Actually they are right. If you consider all the downsides and costs that arise due to a bad workflow shooting 35mm is not more expensive and much safer.

Implementing RED in such a conservative environment will be a challenge.

Hans

PS: And if any one suggests: Make it by your self with FCP (a bargain, have it since V1.1) please remind that achival, versioning and deliveries is key in this industry and the main asset of the obove mentioned post boutiques.

 

[Jason Wingrove]

Hans, the F23, F900, genesis are all catching on here quite strongly, although most stuff is still 35, its really starting to get a lot of use and many DP's here (all with significant film feature credits) are happy to dabble and enjoy the challenge, my experiences with all of the above have been generally good, with well chosen locations etc you can get some great images. I generally choose the right tool for the job, shooting digital only where theres a reason for it but its slowly becoming mainstream. There's at least 2 or 3 features shooting here and in NZ on the red as we speak and Superman returns was shot here on genesis.

Dylan
images are the ones looking at the job for us and at present looking at improving turnaround times. They seem to have a great handle on workflow for getting drives and bins prepped for Avid or FCP until Avid release whatever update they're going to do later this year.

I know this is really the bleeding edge so i should be patient, but just thought there might be a better way than a full day prep per hr of ftg but there's a lot of work going on there to prep things really. Attached is their simple workflow diagram, which was emailed to us so hopefully they wont mind as im sure theyre a regular here :sarcasm:

 

[M Most]

The post company we're dealing with is quoting 24hrs for every hr of footage to “process” a “best light” grade, create an archival backup and provide standard def rushes on a hard drive prepped into avid bins etc and an SR

Yes, that's about right. However, you don't necessarily need a full quality debayer and an SR master to begin offline editing - unless you're creating a new master with new time codes on tape. Since the original time code and original reel number are embedded in the R3d files, one could consider an alternate workflow tailored to the needs of the offline/online approach. For instance, a quick debayer (say, "standard" quality as opposed to "full") takes much less time (and looks essentially identical to full, at least on most video monitors), so theoretically one could convert all of the footage for offline - staying in the file based world - using this approach. While offline editorial is taking place, the "full" version could be processed and later played out to SR tape using the original time codes (there are a number of ways to do this). True, this requires two processing passes, and a longer total turnaround for the post facility, but it's much friendlier to the needs of the overall post process for these type of projects. The other facility alternative, of course, is multiple processing stations (basically a render farm approach), but a facility has to be doing a lot of Red work to justify this.

 

[M Most]

Jashunter, sounds like you could get some kind of hybrid solution going using a FCP / RedCine station that outputs digi and HDCAM SR tapes from the Red footage? Surely that can be done in less than the time and cost it takes to develop the 35mm film and telecine?

Not if you're using Full quality debayer. Not even close, especially if you're printing 2 hours of footage (quite common in both narrative and commercial work these days).

 

[Bruce Allen]

Not if you're using Full quality debayer. Not even close, especially if you're printing 2 hours of footage (quite common in both narrative and commercial work these days).

1. What if he had a bunch of PCs unning RedCine, feeding the FCP output station?

2. If he is getting two hours of footage per day, what if he doesn't do full quality debayer for the digibeta offline? How about if the full-quality HDCAM SR version could be delivered a little later?

3. What was your price estimate for the telecine, again?

EDIT: oh, you also raised #2 as a point. Cool. Thought we were in more agreement...

Bruce Allen
www.boacinema.com

 

[M Most]

1. What if he had a bunch of PCs unning RedCine, feeding the FCP output station?

2. If he is getting two hours of footage per day, what if he doesn't do full quality debayer for the digibeta offline? How about if the full-quality HDCAM SR version could be delivered a little later?

3. What was your price estimate for the telecine, again?

EDIT: oh, you also raised #2 as a point. Cool. Thought we were in more agreement...

1 and 2: Already asked and answered.

3. Irrelevant. Not a part of this discussion. But, it should be kept in mind that processing time is not free. If it's going to take 20 hours, that's a lot of processing time that someone has to pay for. And that doesn't even include playing it out to SR tape, the tape stock, or the cost of the machine itself. The cost of putting together a render farm, the storage required, possible SAN licenses, and render management software if it is needed is not free, either. So one way or another, there are continuing costs involved, whether you're processing film or processing Red files.

 

[Bruce Allen]

Mmost, I totally agree with every point you make.

I originally recommended for Jashunter to consider shooting film instead.

But I am playing devil's advocate here, based on my good experience with Red on the Ringo video.

We were able to transcode everything to ProRes over two days... on a single iMac. That wasn't full quality debayer, but it was good enough for the CBS Daily Show to broadcast the whole thing every day for a week. We have seen the full HD version projected and it looks good. I think it would compete with a lot of HD programming we have seen originated from other cameras.

I know film telecines aren't that expensive. The last time I was at R!OT was for a $10,000 music video - they squeezed us in at midnight and did a great job within our budget.

But I think your comparisons are a little unfair. Of course you need an HDCAM SR deck, and tape stock, and operator time, but you'd need that for a telecine too.

I don't understand why a telecine house can't buy say 20 fast PCs for $1000 each (my $450 "Craigslist special" PC runs RedCine quite nicely), get an intern to render baby-sit, put the files on their SAN which already exists, do a one-light, then write it out to tape on the HDCAM SR deck they already have - and offer Red footage "processing" at competitive cost, time and quality to a telecine?

Bruce Allen
www.boacinema.com

 

[Anthony Gratl]

I don't understand why a telecine house can't buy say 20 fast PCs for $1000 each (my $450 "Craigslist special" PC runs RedCine quite nicely), get an intern to render baby-sit, put the files on their SAN which already exists, do a one-light, then write it out to tape on the HDCAM SR deck they already have - and offer Red footage "processing" at competitive cost, time and quality to a telecine?

Bruce Allen
www.boacinema.com

well...uh....er...it's complicated.....because....and uh....then we'd have to....well uh the super deluxe espresso machine ain't cheap...er...not make as much money....uhhmmm...render intern huh.....could be a problem....dammit bruce....you make way too much sense.

 

[M Most]

I don't understand why a telecine house can't buy say 20 fast PCs for $1000 each (my $450 "Craigslist special" PC runs RedCine quite nicely), get an intern to render baby-sit, put the files on their SAN which already exists, do a one-light, then write it out to tape on the HDCAM SR deck they already have - and offer Red footage "processing" at competitive cost, time and quality to a telecine?

That could be done. However, as I said before, you'd have to see enough return in it - in other words, be working with enough Red footage - to justify that kind of expense and trouble. For one thing, the SAN licenses for that many CPU's would likely cost considerably more than the blades themselves. And with the size of the files involved, using Ethernet would probably slow down the process considerably, so you're talking about a lot more than $20K for a high performance solution, although there are workarounds. But one has to keep in mind that there needs to be enough business from sensibly budgeted projects (in other words, the ones that would use facility services) being shot on Red to justify a solution that only applies to material shot with Red cameras. At this point in time, the need is unclear. But ask me again in about 6 months.

 

[Hans von Sonntag]

Hans, the F23, F900, genesis are all catching on here quite strongly, although most stuff is still 35, its really starting to get a lot of use and many DP's here (all with significant film feature credits) are happy to dabble and enjoy the challenge, my experiences with all of the above have been generally good, with well chosen locations etc you can get some great images. I generally choose the right tool for the job, shooting digital only where theres a reason for it but its slowly becoming mainstream. There's at least 2 or 3 features shooting here and in NZ on the red as we speak and Superman returns was shot here on genesis.:

Jashunter,

DOPs tend to be more experimental than post houses. Digital isn't bad at all and the cameras you mentioned are far from "video". It's just that you do not get supported by the posthouses. Posthouses tend to hate directors especially if they want a change the workflow. This is what they say:

1. We've done always this way.

2. See above.

3. Nice idea but this is going to be too expensive.

4. We don't take any responsibilty.

As Michael Most mentioned above: Debayering in a acceptable turn around is not cheap (so is lab). Its for the indies and the brave to dig into this.

Hans (who is brave in this regard)

 

[Bruce Allen]

But ask me again in about 6 months.

Sounds good, Mr. Most, will do! So many good points raised here, thank you.

Hopefully someone will put together some Ghetto Farm tests and give us some speed / cost / quality comparisons for the indie crowd between:
1. PC-based Ethernet RedCine farm -> CineForm -> Premiere -> tape
2. PC / MAC -based Ethernet RedCine farm -> DNxHD -> Avid -> tape
3. MAC-based Ethernet RedCine farm -> ProRes -> FCP -> tape

...and a high-end uncompressed SAN system? How much speed and quality do you gain for the (presumably high) increase in gear cost? Of course, there would be labor savings (and frustration savings too

I guess it's a bit of a moving target at the moment, seeing as CineForm's outputter just got temporarily deep-sixed by whatever Faustian contract Red had to sign in order to deliver us their magic wonders, and that RedCine is still in a state of development? Does it even output timecode in its QT files yet?

Bruce Allen
www.boacinema.com

 

[Dylan Reeve]

For what it's worth, I just tested with some RED footage...

In Avid Xpress Pro HD 5.8 on OS X 10.4.11, with the RED QT codec installed, I can import the RED proxy QT Reference files directly into Avid quite easily. Obviously they are only SD res (the 'H' 1k file on our samples is 1024x576, or PAL 16:9 Square Pix).

Importing a 1'50" clip to DV 420 took a bit under three minutes on our Octo-core 3GHz machine. To other resolutions may be quicker.

I believe this capability combined with the MetaCheat ALE utility makes it fairly simple to offline RED footage in Avid anyway. We'd probably make the media and bins on the Mac and then edit on our other Avid suite.

 

[Jabez Olssen]

For what it's worth, I just tested with some RED footage...

In Avid Xpress Pro HD 5.8 on OS X 10.4.11, with the RED QT codec installed, I can import the RED proxy QT Reference files directly into Avid quite easily. Obviously they are only SD res (the 'H' 1k file on our samples is 1024x576, or PAL 16:9 Square Pix).

Importing a 1'50" clip to DV 420 took a bit under three minutes on our Octo-core 3GHz machine. To other resolutions may be quicker.

I believe this capability combined with the MetaCheat ALE utility makes it fairly simple to offline RED footage in Avid anyway. We'd probably make the media and bins on the Mac and then edit on our other Avid suite.

Yes exactly, although the news is even better because the 'H' quality quicktimes are actually 2048x1024, so they are more than enough res for importing into HD in the Avid. (The 'M' quality ones are 1024x512). And if you're editing in SD then the 'P' quality ones are fine, and they import in just over 2:1 realtime (on an Octo core Mac).

If you are just offline editing in the Avid (and plan to online later in something else) then all this talk of doing a proper debayer or going via HDCAM-SR seems strange to me. Just edit in an Avid 1080p HD project, import the 'M' quality quicktimes (they look great) and you can get rushes into the Avid much quicker than the 24:1 timeframe that the original poster on this thread was quoted. I time it at about 4or5:1 realtime... And here's the thing... that is assuming that the lab only has one mac to import rushes on. If they get a second one (plus another copy of Avid Xpress) then the time halves.

In fact, as I said, if you edit in SD its almost 2:1 already (although I will probably have to add the NTSC timecode fix to metacheater to make that work... but I recommend all Avid offlining be done in DNxHD36 anyway

Processing footage via RedCine just for the offline rushes is not the smart way to go at the moment. Instead, use RedAlert to instantly make QT wrappers, make an ALE of the metadata and import them into the Avid. It is faster than a 35mm film telecine workflow. All this talk of Red and Avid not being possible just yet seems crazy to me, when a couple of Macs with Xpress Pro makes it very quick. Obviously its not ideal, or suited to really low budget projects (but they would probably be better off with FCP anyway), but for bigger features, and high end commercials, there is a fine and fast workflow available right now.

 

[Jason Wingrove]

Progress

Progress

Both editors I work with here in Syd are now onto it big time and keen to get it happening, everyone is all for a SIMPLE solution as there's not just this job but research for an upcoming mainstream feature on Red to evolve as well.

Im confident I'll get a solution soon and of course when we get to the other side of this job in a couple of weeks I'll feedback what we learnt along the way. The idea really is to avoid tape whilst ensuring we're well and truly covered with not only backups but that footage is easily accessed by others in the future without having to to resort to alchemy or the dark arts.

 

[Jason Wingrove]

well said Cogito, we're looking in that direction too.
j

 

[M Most]

Im confident I'll get a solution soon and of course when we get to the other side of this job in a couple of weeks I'll feedback what we learnt along the way. The idea really is to avoid tape whilst ensuring we're well and truly covered with not only backups but that footage is easily accessed by others in the future without having to to resort to alchemy or the dark arts.

In my experience, the biggest hurdle, and one that has yet to be overcome, is the clients' perception that dailies must represent final release quality if the program is being shot on an electronic format - when in fact, that is never the case with film, and it's never been an issue. My feeling is that this all started because the first recent use of electronic capture in high end production was with video cameras, primarily the F900, in which what you see at the time of shooting needed to be fairly close to what you wanted, because the latitude is limited and the ability to alter the resulting footage is also limited due to both the nature of 4:2:2 video and the amount of compression used and its effects. On the one hand, this provides a nice workflow because the look of the images is never in question. On the other hand, on a professional production, there should be enough confidence in the director of photography that the look of a daily videocassette shouldn't be an issue. With a raw capture like that of Red, the look of the dailies is only somewhat important (there is more latitude and better color depth, so there is more flexibility), but the quality of the offline image is not - because the intent is to go back to the original and do high quality processing for the final deliverables, much as you do with film. So the notion that the online is created from the offline, which is the common workflow with HD video capture and most network television post production, is not applicable.

But good luck in trying to get people to understand that.