1 person found this helpful
I have neither a 5D2 or NeoScene, so I can't comment on the workflow, but I want to point out an error/deficiency in your explanation of NeoScene:
That's because NeoScene increases the 5D2 raw file up to 10-bit, arguably making a bit better looking clip from the 4:2:0 file you get from the 5D2.
That's a nice thought, but it doesn't work that way. You can't increase visual quality by changing the compression of clip, or even by changing a clip from a compressed state to an uncompressed state. That's sort of like saying you can create a quarter by rubbing two nickels together. In other words, you can't create something that isn't there to begin with--it's sort of like the Law of Conservation of Matter for video quality. The best you can hope for is for better results when applying effects or transformations to the converted video than to the original more highly-compressed version, or for a better editing experience. It likely won't get worse when you do a conversion using NeoScene, but it won't get better, either.
Apologies if I come off as a buttinski, but I think it's an important clarification.
A side note: when using NeoScene a i7 is recommended as they are very large files.
LOL! No, you're not a buttinski, I had wondered the same thing which is why I preceded that statement with "arguably". I got that info straight from Cineform, and they state that the video quality actually improves... but it made me think of upsampling in Photoshop: You get a bigger print, but print quality doesn't improve, right?
Glad you responded, thanks!
Ann, you're right about the i7, I'm glad I made that investment. I think we all need Harm's rig to edit 5D2 clips!
Wow, Cineform actually makes that claim? Pretty bold, I must say. I'm not a computer/electronics/software/rocket ship engineer--just a lowly video hacker--but I can't really imagine this is true. If it can, I'd have to say that NeoScene is performing black magicks because if the pixels aren't there in the original, as-shot footage, how is the codec to know what's missing? The best it could hope to achieve is a pixel-for-pixel copy of the original footage... at least, that's the only thing that seems logical to me.
Now to say that it stands up better to processing in post, that may have some creedence, and is measurable... so I'll buy that.
Colin, believe me, I hear ya, if data isn't there, it isn't there. So, I went to their site to see if I actually read it right, and they do state that they convert a 4:2:0 5D2 file to a 10-bit 4:2:2 file using chroma interpolation... probably similar to Photoshop's ability to upsample by grabbing the nearest surrounding pixels and extrapolating an averaged value to create the missing pixel. Here is Cineform's spec list on NeoScene talking about the chroma interpolation (briefly):
Ba da bing.
and they do state that they convert a 4:2:0 5D2 file to a 10-bit 4:2:2 file using chroma interpolation... probably similar to Photoshop's ability to upsample by grabbing the nearest surrounding pixels and extrapolating an averaged value to create the missing pixel.
Oh, I'm not disputing that. I'm sure that's exactly how it works, and I'm sure it works well. It would make for smoother color gradations and less aliasing along distinct edges in a shot. But is it higher quality? Seems like a pretty subjective thing. What if the interpolated pixels are way off from what the "real" color values would have been? Interpolated data is still data that wasn't there.
(I'm not trying to be a jerk... just making conversation!)
They do say this on the product page you linked to:
When Neo Scene detects 4:2:0 chroma it properly interpolates the source chroma to 4:2:2 for more accurate color processing during editing and effects work. And if you ever "key" your material, CineForm’s chroma interpolation will substantially improve your resulting visual fidelity.
It's those "during editing and effects work" and "resulting visual fidelity" parts that are important; that I can live with.
Colin, you're not being a jerk, I'm totally on the same page with you! I have several years' experience with Photoshop so I know that "fill in" pixels sounds nice in theory, but in reality... well... we aren't there yet. So when I saw the claim "4:2:0 --> 4:2:2" I should have known better and not been so enthusiastic in my original post.
Video is a new world for me, so I'm still in the kid-like stage where everything you hear sounds great. I feel kind of bad for the long-time serious video folks, because their world is now getting stormed upon by all of us who went out and bought a Canon 5D Mark II and all of a sudden have a serious device to capture HD video (albeit, quirky to hold) with our lenses ranging from 16mm to 600mm... and now there's these overnight video folks who, like me, had no CLUE all that it entails. Some think their Dell P4 from 2003 with integrated video and 512mb of ram should be able to edit video... others (like me) plunged into building a gamer rig and spent millions on that and Production Premium CS4, and yet we're all totally clueless. But I tell you what: when the video bug hits for real, WOW, is it all ever exciting! I can't believe what I've been missing all these years. These folks getting these DSLRVs better be passionate about video making, because it's going to drain their time and bank account, lol!
So anyway, I appreciate the seasoned veterans of the video game who are having great patience with us newcomers.
As I mentioned in another post, I'm having no trouble importing 5D Mark II files natively into PrPro CS4. They're choppy while the yellow line is visible, but after rendering the timeline and getting a green line, they playback and edit perfectly. No conversions or additional codecs needed.
I do have to agree that handheld shooting is somewhat precarious since you have to hold the camera away from you to see the screen and that puts the focus ring a little far away on some lenses, but using Canon L lenses more than makes up for the awkwardness when you see the results!!!