Noisy blue channel with 5218 after HD telecine

[Thomas Worth]

I was fortunate enough to be part of a shoot that ended up transferring the 35mm footage (all 5218) to HD, and noticed that the blue "channel" or layer was very grainy in comparison to the other two layers. Now I understand this is characteristic of tungsten balanced stock, but I found it very distracting. My question is, would shooting on daylight stock do much to alleviate this problem, or would I just have the same problem but in reverse (the red channel being grainy)? I tend to prefer slower stocks in general, but if there's a possibility of reducing the grain in the blue channel by using a daylight balanced stock, I may opt to do that in the future.

 

Here are the specs of the shoot:

 

Film stock: 5218, rated at 320 ASA, processed normally

Telecine: C-Reality to HDCAM, flat transfer

HDCAM footage was then transferred to disk as an uncompressed QuickTime using a Kona board on a Mac

[David Mullen ASC]

Perhaps something was wrong with the telecine -- you should do a test on another telecine to see if that's the problem, because blues should not be overly noisy/grainy on well-exposed 5218.

[Phil Rhodes]

Hi,

 

I think if we'd exposed that 5218 any more "well" it'd have burned a hole in the stock!

 

Perhaps Mr. Worth could be persuaded to post a couple of stills.

 

Phil

[Stephen Williams]

I was fortunate enough to be part of a shoot that ended up transferring the 35mm footage (all 5218) to HD, and noticed that the blue "channel" or layer was very grainy in comparison to the other two layers.

 

Hi,

 

Blue was always the noisy channel on tube telecines in the past. Personally I find 5218 just a tad too grainy for my taste.

 

Stephen

[Dan Goulder]

Here are the specs of the shoot:

 

Film stock: 5218, rated at 320 ASA, processed normally

Telecine: C-Reality to HDCAM, flat transfer

HDCAM footage was then transferred to disk as an uncompressed QuickTime using a Kona board on a Mac

Although it wouldn't be the source of the problem, would you mind sharing which Kona present you used for the HDCAM ingest? Thanks.

[Thomas Worth]

Ok, hopefully this will help you guys see what I'm seeing. Excuse the long loading times if you're on a slow connection, but the JPEGs haven't been compressed much to preserve as much quality as possible.

 

Here's a full frame, resized to fit the screen, no color correction:

 

Here's a portion of the frame at 100% to show the grain:

 

Here's a portion of the frame at 100%, with some color correction. Now you can really see the blue grain:

 

And here's a link to an uncompressed TIFF (zipped) for closer inspection:

http://rarevision.com/pics/avatar_still_nocolor.zip

[Eric Steelberg ASC]

Can you post examples? If you were exposing at 320 you should have no problems. Even when I've exposed at 500 I've never had issues, even in HD.

 

The issue, as far as I know, with the blue channel being noisy is a video quality and has nothing to do with the film stock. The telecine machine sounds suspect to me. Where did you have it done?

 

What are you using for your software color correction?

[Thomas Worth]

Although it wouldn't be the source of the problem, would you mind sharing which Kona present you used for the HDCAM ingest? Thanks.

It was 1920x1080 @ 10 bit, uncompressed.

 

QuickTime reports the data rate as 1.07 gbits/sec, if that helps.

[Eric Steelberg ASC]

Seems like a problem with the transfer. You can already see a lot of video noise in the FLAT transfer beofore color correction. A FLAT HD transfer of film properly exposed should look silky smooth, even 5218.

[Thomas Worth]

The issue, as far as I know, with the blue channel being noisy is a video quality and has nothing to do with the film stock. The telecine machine sounds suspect to me. Where did you have it done?

We had it transferred at Level 3. I asked about grain reduction halfway through the session (duh), and the colorist chick said, "Uh, yeah we can do it, but I have to connect it."

 

What are you using for your software color correction?

The still here was corrected with Photoshop, but typically I use Combustion.

[Bryan Darling]

We had it transferred at Level 3. I asked about grain reduction halfway through the session (duh), and the colorist chick said, "Uh, yeah we can do it, but I have to connect it."

 

I think that alone says a lot about the operators and facility. I'd have been very weary. Equipment is nothing without skilled and meticulous operators.

[Phil Rhodes]

Hi,

 

I'm sure it's being exacerbated by HDCAM compression.

 

Phil

[David Mullen ASC]

Hi,

 

I'm sure it's being exacerbated by HDCAM compression.

 

Phil

 

That's my thought too -- it looks like the typical problem of trying to color-correct HDCAM.

[Joshua Reis]

Was the scene shot mostly color corrected in camera via filtration or in telecine? If the telecine operator had to pump up the blue channel for correction, that could defintely be a cause for seeing a dramatic increase in grain/noise in only blue. I've seen this before on uncorrected 5218 footage which favors reds and earth tones. I'd be curious to see what the footage looked like on a Spirit 2k.

[Phil Rhodes]

Hi,

 

Nope, all under tungsten. Some (most?) of the tungsten was corrected with - um, what was it Thomas, quarter or half plusgreen?- to match practical flos.

 

Phil

[Thomas Worth]

Nope, all under tungsten. Some (most?) of the tungsten was corrected with - um, what was it Thomas, quarter or half plusgreen?- to match practical flos.

I'm pretty sure tungsten sources were gelled with 1/4 plusgreen. But we overexposed by 2/3 stop, so wouldn't that require dropping other colors, not pumping up the blue?

 

I'm still not convinced it's not video noise. If I cycle each channel as grayscale while in combustion (Red, Green, and Blue), you can see where the grain (or noise) is identical across all channels. I know film isn't exactly RGB (more CMY), but would the grain intrude on other channels after RGB conversion to point where it was very obvious?

[bridgett roh]

just wondering how old your 5218 was, did you buy new? if you have access to a densitometer read an unexposed section on the neg, the blue should read about .9 to 1.0 for good stock. The 500 negs don't keep for to long even refrigerated, even so telecine should be able to correct any minor fog in the blue as long as its not to high.

[Sam Wells]

I'm not sure why you didn't color grade it in the transfer. Instead of dealing with the already-compressed HDCam.... ?

 

The uncorrected frame grab has the C-Reality texture (which I personally like) but I can't really judge the specific machine from this frame grab.

 

Spirit has more of a look I call "metallic pointilism" but I think the same problems would occur with that path..

 

-Sam

[Phil Rhodes]

Hi,

 

> I'm not sure why you didn't color grade it in the transfer. Instead of dealing with the already-compressed

> HDCam.... ?

 

Time, money, flexibility.

 

Would have much preferred to get it on a hard disk.

 

Phil

[Thomas Worth]

Would have much preferred to get it on a hard disk.

You and me both. About the best solution I forsee at this point without straight to disk transfer is telecine to HDCAM SR, then from HDCAM SR to disk via the Kona at 1920x1080, 10 bit.

 

I hope to test that soon.

[Sam Wells]

Time, money, flexibility.

 

Phil

 

Pick two :(

 

-Sam