Telecine Tips

[Damien Bhatti]

I am about to have some telecine done, hopefully at the Mill in London, I have been talking to them a little and they seem most accommodating. However they can offer me a scan to hard disk at 2K to create a 10bit RGB LOG scan. Now the program I am using is Avid Liquid, but as I have learnt Liquid is only a 8 bit program and would not be able to handle resolution at 2k. What kind of external hard drive would be useful at such a range? I was thinking of investing in a G tech drive, but someone advised me that I would need something more substantial at these kinds of levels.

 

As I am still learning about these things, do you think that it would make sense to simply telecine the footage out to digi beta/ dvcam and take it from there?

 

Would I be right in assuming that a quality scan could still hold its own against a one light transfer even on a lesser format, say Mini DV?

 

The work flow I am imagining is a good telecine transfer onto digi beta (for posterity) and dvcam - onto Liquid, and if the project warrants it for festivals back out to digi beta. Would this be viable?

 

I simply cant afford to take it from digi beta to another post house to get it transferred back to me to edit.

 

Thanks

[Daniel Porto]

I am about to have some telecine done, hopefully at the Mill in London, I have been talking to them a little and they seem most accommodating. However they can offer me a scan to hard disk at 2K to create a 10bit RGB LOG scan. Now the program I am using is Avid Liquid, but as I have learnt Liquid is only a 8 bit program and would not be able to handle resolution at 2k. What kind of external hard drive would be useful at such a range? I was thinking of investing in a G tech drive, but someone advised me that I would need something more substantial at these kinds of levels.

 

As I am still learning about these things, do you think that it would make sense to simply telecine the footage out to digi beta/ dvcam and take it from there?

 

Would I be right in assuming that a quality scan could still hold its own against a one light transfer even on a lesser format, say Mini DV?

 

The work flow I am imagining is a good telecine transfer onto digi beta (for posterity) and dvcam - onto Liquid, and if the project warrants it for festivals back out to digi beta. Would this be viable?

 

I simply cant afford to take it from digi beta to another post house to get it transferred back to me to edit.

 

Thanks

 

I have a Firewire 800 external hard-drive and that works fine for editing my 10-Bit footage, on Final Cut Pro on a Mac Book Pro.

 

Dont transfer to DV. Its like buying a ferrari and taking out the motor and then pushing it.

 

You should be fine with the footage. Just tell your telecine guy/girl to retain as much detail in the image when transferring and then just color correct it all you like in post. Its the cheapest way to get awesome quality images.

[Phil Rhodes]

2K to create a 10bit RGB LOG scan

 

This will be quite demanding. Usually you would create lower resolution material to do your edit, then render the higher resolution stuff once you've made all your decisions in the easier-to-handle format. You can do that by creating low resolution proxies from your high resolution stuff.

 

Now the program I am using is Avid Liquid, but as I have learnt Liquid is only a 8 bit program and would not be able to handle resolution at 2k

 

I don't know Liquid, but if what you're saying is true then you'll have to find another platform to do your high resolution work.

 

What kind of external hard drive would be useful at such a range? I was thinking of investing in a G tech drive

 

No, your silly white-painted, style-conscious, look-at-me-I-own-a-mac external firewire drive will not be suitable for 200MB/sec high def online editing!

 

As I am still learning about these things, do you think that it would make sense to simply telecine the footage out to digi beta/ dvcam and take it from there?

 

That would be easier, if vastly lower quality.

 

Would I be right in assuming that a quality scan could still hold its own against a one light transfer even on a lesser format, say Mini DV?

 

Hold its own? The high quality scan would obliterate the standard-def transfer in terms of quality. Needless to say, the scan would need grading.

 

The work flow I am imagining is a good telecine transfer onto digi beta (for posterity) and dvcam - onto Liquid, and if the project warrants it for festivals back out to digi beta. Would this be viable?

 

Entirely viable and done all the time but vastly different to the scan option in more or less every way - different price, different quality of result, different technical implementation.

 

I have a Firewire 800 external hard-drive and that works fine for editing my 10-Bit footage, on Final Cut Pro on a Mac Book Pro.

 

In what format?

 

P

[Damien Bhatti]

It does seem like a regression paying all that money for a nice scan and then taking it to mini dv.

 

My next question, and thanks for the responses, would be what kind of drives do you guys use that would be suitable for my predicament?

[Phil Rhodes]

Depends how you want to do it.

 

If you want to build a system from piece parts, there are two options:

 

Get an 8-port serial ATA hardware RAID controller and load up with eight of the biggest, fastest hard disks you can find. This is probably the best way of doing it; given a decent RAID controller (a hardware one, several hundred units of currency, not one of the £100 ones) you will have reliable performance and redundancy in case of failure.

 

On the other hand it can be done with some of the six-port controllers you get on various low-cost PCIe cards and PC motherboards. You won't be able to use a protected RAID mode, so don't rely on it, but the performance can be more than good enough.

 

The other way is just to buy a fiberchannel-connected external storage rack, but that's a higher priced option.

 

No you cannot get an iHighDefinitionFinishingStation from the Mac store and they do not come in a variety of colours.

 

On the other hand, as I say, you can figure out a way to make low-res Quicktime proxies of your material and edit that, then conform it back to the HD/2K stuff later on. You can achieve a 2K finish without ever needing 2K-capable storage.

 

P

[Damien Bhatti]

Depends how you want to do it.

 

If you want to build a system from piece parts, there are two options:

 

Get an 8-port serial ATA hardware RAID controller and load up with eight of the biggest, fastest hard disks you can find. This is probably the best way of doing it; given a decent RAID controller (a hardware one, several hundred units of currency, not one of the £100 ones) you will have reliable performance and redundancy in case of failure.

 

On the other hand it can be done with some of the six-port controllers you get on various low-cost PCIe cards and PC motherboards. You won't be able to use a protected RAID mode, so don't rely on it, but the performance can be more than good enough.

 

The other way is just to buy a fiberchannel-connected external storage rack, but that's a higher priced option.

 

No you cannot get an iHighDefinitionFinishingStation from the Mac store and they do not come in a variety of colours.

 

On the other hand, as I say, you can figure out a way to make low-res Quicktime proxies of your material and edit that, then conform it back to the HD/2K stuff later on. You can achieve a 2K finish without ever needing 2K-capable storage.

 

P

 

And this is where my lack of telecine wisdom comes to the fore - how would you go about making QT proxies of the footage? Would i ask at the house to put the scaled down version onto an external hard drive and then take it back to them to let them put the ribbons on? Or would it involve working in my software somehow? Thanks Phil!

[Phil Rhodes]

Taking you through the nitty gritty of a full data based online/offline workflow is really a bit beyond the scope of an internet forum, so I would suggest that you simply tell the post house how you want to cut it, and get a price from them for creating your offlines and then conforming the online to your edit. You will need to have at least something of a technical discussion with them about this to minimise SNAFUs.

 

P

[Damien Bhatti]

Taking you through the nitty gritty of a full data based online/offline workflow is really a bit beyond the scope of an internet forum, so I would suggest that you simply tell the post house how you want to cut it, and get a price from them for creating your offlines and then conforming the online to your edit. You will need to have at least something of a technical discussion with them about this to minimise SNAFUs.

 

P

 

Thanks, need to do some more reading really. for a moment I thought snafu was telecine code for something, before looking it up on wiki (!)

[Tim Terner]

We know you are the 'king of post' Phil, but what doe's SNAFUs mean ?

[Damien Bhatti]

We know you are the 'king of post' Phil, but what doe's SNAFUs mean ?

 

 

 

"SNAFU is an acronym meaning roughly, "things are in a mess ? as usual". The most commonly accepted rendering is "Situation Normal: All **(obscenity removed)**ed Up". In computer jargon, it sometimes is intended to mean "Systems Neatly All **(obscenity removed)**ed Up". It is sometimes bowdlerized to "Situation Normal: All Fouled Up" or similar,[1] in circumstances where profanity is discouraged "

 

Maybe?

[David Bradley]

A 2K will cost a fortune and nothing short of an Avid DS will be able to handle it in real time. You would definately need to make a low res copy to edit offline.

 

You would be better off getting a technical transfer to HDCAM-SR (if your looking for high-res). Most labs can transfer to .DPX files for you so you may be able to whack it straight on to a raid array. HDCAM-SR runs at either 440mbps or 880mbps so you will need a firewire 800 or fibre optic raid that can push data at that rate.

 

I don't understand why you would do a 2K and then print to digibeta. Stick with SD.

 

I would do this

 

- Best light or tech transfer to Digibeta or .DPX files if you don't have a deck

- Edit / grade / conform SD online

- Print to Digibeta

 

If you want to blow up to 35mm later then export an EDL from avid and take it to the lab - they will be able to scan the relevant portions of the neg at 2K, grade it and then print to 35mm (digital intermediate).

 

Digi Beta runs at 90mbps so a decent firewire sata should handle it.

 

Don't scan at 2K and then print to digibeta. Its a waste of money and quite frankly - its a sin.

[Phil Rhodes]

> A 2K will cost a fortune and nothing short of an Avid DS will be able to handle it in real

> time.

 

In general yes, although as a point of order it is now possible to do this on a fairly modest desktop computer which will cost you less to buy than a couple of hours in a DS suite. I'm not advocating this for everyone, but it can be done.

 

> HDCAM-SR runs at either 440mbps or 880mbps

 

Yes but this is irrelevant as you will (unless you've found an Xpri edit station, in which case all of this is moot) be capturing it uncompressed, which is higher bitrate.

 

P

[Damien Bhatti]

I don't understand why you would do a 2K and then print to digibeta. Stick with SD.

 

I would do this

 

- Best light or tech transfer to Digibeta or .DPX files if you don't have a deck

- Edit / grade / conform SD online

- Print to Digibeta

 

Don't scan at 2K and then print to digibeta. Its a waste of money and quite frankly - its a sin.

 

To be honest with you I am new to editing and what I want is a good copy of my project for festival circuits, hence the digi beta.

 

So 2 k is too much for digi beta?

 

- Best light transfer to .dpx files,

- Edit with a low res copy

- Hand the Bins/Edl back to post house

- Get master on digi beta

 

Perhaps I was going over board with such a high grade scan.

 

What level would you guys suggest for a good copy?

 

Thanks

[David Bradley]

Stick with digibeta.

 

Its cheap robust and and works in a 4:2:2 RGB colour space - this means you can tweak the grade in post without running in to too many compression artifacts. HD is great but at the lab end it still costs about about three times as much as SD.

 

Festivals are tricky because they often require 35mm for projection. Even if they didn't projected SD looks very mucky.

 

This is why I suggested holding on the EDL. With it you can return to the lab who will re-scan selected portions of the neg at 2k or 1080 HD and print it to 35mm. This costs alot (relative to your initial processing and SD dailies route) but if you want your film to be screened it may be a necessary evil.

 

It doesn't make sense to best light to DPX and then get the lab to master to digibeta. Why not have your dailies delivered on digibeta and just edit the lot online. The workflow you are suggesting will just cost more.

[Phil Rhodes]

Personally I would not aim to finish to anything less than very good or uncompressed HD; that's probably because I know how to do it fairly cheaply, but even then I would not go for anything less than at least some form of HD.

 

P

[Mike Nichols]

Personally I would not aim to finish to anything less than very good or uncompressed HD; that's probably because I know how to do it fairly cheaply, but even then I would not go for anything less than at least some form of HD.

 

P

 

Phil, what are thoughts on DPX files within the FCP environment? I am having nothing but headache after headache.

[David Bradley]

Personally I would not aim to finish to anything less than very good or uncompressed HD; that's probably because I know how to do it fairly cheaply, but even then I would not go for anything less than at least some form of HD.

 

P

 

HOW?! For an upcoming production I had considered a technical transfer to HDCAM-SR as a master from the lab but deck hire, lab costs and storage devices capable of keeping up with insane bit rates scared off the producer. Even editing offline and conforming was going to cost an arm and a leg.

 

It seems HD is cost effective if you shoot HD in the first place but going from 16mm/35mm invariably costs more.

 

If you have a cheap solution please share

 

Cheers

 

Dave

[Phil Rhodes]

If you want to emulate high-end practice you have little choice but FCP and color.

 

Final cut does not speak DPX at all, which is indescribably stupid, but there you go. Color does, although I seem to recall that it will only let you export DPX if your input is DPX - that's secondhand information, though, so double check.

 

Color will read filename timecode and doesn't need it to be embedded in the headers.

 

Personally I would do all of this in Adobe software on a PC, partly because that's what I know, but mainly because it has rather greater flexibility. It's nothing like high end practice but it actually works somewhat better, since you can simply rely on filenaming and common start points and never have to care about timecode or EDLs since you're working solely in the internal format of the NLE.

 

After Effects can be pressed into service as a fairly flexible file conversion utility (and it's scriptable) and Premiere has had the ability to perform file-centric onlines since version 4 in the late 90s - it's just that nobody ever thought to use it that way. This will require handling your material as AVI or Quicktime movies, but really, the only plus side to handling it as DPX sequences is the ability to recover only part of the take from a magtape. If you're not doing that, it's irrelevant.

 

The real gotcha here is 10-bit support, which is extremely patchy in general on desktop apps, and what's worse, may fail silently, truncating your data to 8-bit but saving it in a 10-bit container. It's worth verifying every stage of your workflow for this, usually by dragging an eyedropper over the frame in a 10-bit capable application. If all the pixel values are suspiciously multiples of four, you know you have a problem. I have somewhere a matrix, created by the estimable Thomas Worth, which shows what works in what circumstances in this regard. I'll try and get it displayed here.

 

Persuading things like Final Cut, After Effects and Premiere to do HD onlines from offlines cut in the same software has big advantages in that you can freely use any of the effects controls or plugins the thing has, and you will not have trouble translating those plugins or their settings to the online, notwithstanding however many of those plugins are 10-bit capable. The only thing to watch out for is if you set some plugin variable as a pixel count, since most of them will not intelligently scale that value when you come to online and suddenly all your gaussian blurs start to look a lot less blurry.

 

As I say the major advantage of this is that you can let it clunk through your online unattended and create a full-on HD final which you can step through and at least basically QC, without ever needing HD-capable storage.

 

P

[Phil Rhodes]

HDCAM-SR is a very very expensive tape format.

 

Get it on a hard disk and go from there.

 

P

[Paul Bruening]

Hey Phil,

 

I love to hear you talk PC/Adobe. It can seem like everyone only uses Mac since those users are most vocal. I love how Adobe maximizes crossover power between their softwares. I love the versatility and build variety of PC. I can only take people's word for the superiority of Mac/FCP. Personally, I don't really get it.

 

AE's free render engine is excellent. Buy cheap MoBos with 2-core chips and 6 ports of SATAII. You get another unit towards your render farm and 6 drive stacking in one box. Every one you add multiplies your productivity and at a pretty darn cheap buy-in. You want insane resolutions and multiple workstations to work everything from work print, CGI, high res conforms, dialog polish, music, SFX, then PC is a can-do approach.

[Rolfe Klement]

based upon my experience with 35mm film I would say

 

Get a great 2K telecine to DPX files on a hard drive

Make a back up of the data drive and store it

Buy gluetools and run off 10 bit compressed and uncompressed SD quicktime mov from master 2k DPX (this can take a while)

Edit in FCP using compressed SD (and do audio SFX etc)

Attach uncompressed SD footage - render to 10 bit uncompressed SD mov file (this can take a while)

Take into DVD Studio pro and render out amazing overnight SD DVD at high bit rate (7mb area)

 

Then if someone picks up your idea - you can reconfrm with edl (from FCP) using 2k DPX and render out to film at Capitalfx etc

 

thanks

 

Rolfe

[Damien Bhatti]

based upon my experience with 35mm film I would say

 

Get a great 2K telecine to DPX files on a hard drive

Make a back up of the data drive and store it

Buy gluetools and run off 10 bit compressed and uncompressed SD quicktime mov from master 2k DPX (this can take a while)

Edit in FCP using compressed SD (and do audio SFX etc)

Attach uncompressed SD footage - render to 10 bit uncompressed SD mov file (this can take a while)

Take into DVD Studio pro and render out amazing overnight SD DVD at high bit rate (7mb area)

 

Then if someone picks up your idea - you can reconfrm with edl (from FCP) using 2k DPX and render out to film at Capitalfx etc

 

thanks

 

Rolfe

 

Surely a 35mm from hard disk is going to cost Mega $$$ ? What kind of ball park are we talking roughly, just for hypothesis sake. (10 min BW)

[Chris Burke]

Surely a 35mm from hard disk is going to cost Mega $$$ ? What kind of ball park are we talking roughly, just for hypothesis sake. (10 min BW)

 

 

Is this 10 minutes of finished material, where the colorist is going to be shuttling through several flats looking for your shots or is it one 1000 foot flat with all your footage? If the later, then by all means have the whole thing scanned and use the afore mentioned workflow for creating offline clips. But in all honesty, if it is ten minutes of film, why not do a photochemical finish, especially for black and white? How are you shooting it, academy or scope? In either case you can contact print and get the equivalent of a 4k or more image, probably much more than 4k.

[Damien Bhatti]

Is this 10 minutes of finished material, where the colorist is going to be shuttling through several flats looking for your shots or is it one 1000 foot flat with all your footage? If the later, then by all means have the whole thing scanned and use the afore mentioned workflow for creating offline clips. But in all honesty, if it is ten minutes of film, why not do a photochemical finish, especially for black and white? How are you shooting it, academy or scope? In either case you can contact print and get the equivalent of a 4k or more image, probably much more than 4k.

 

My project is 16mm. Someone mentioned optically blowing up to 35mm from a 2 k scan, and I was curious to see how this much this would be, and also using a 35mm original. If I were to be using 35mm negatives then your suggestion of a photochemical finish is the one I would take.

[Phil Rhodes]

I was curious to see how this much this would be

 

One kidney and your first-born son.

 

P