Brendan Coots
Forum Replies Created
-
“One post of interest pointed to this: Blackmagic Decklink card, broadcast monitor with BNC input.
Before I invest, what can expect from this combination?”
If you get a decent monitor and properly calibrate it, you can expect to see an accurate rendition of what your video looks like, not necessarily comparable to perfect “black box” conditions but good enough for common broadcast work. You can also trust it for broadcast-quality color correction work. For sensitive projects a fairly costly monitor (possibly even HD CRT) is in order.
The missing ingredient needed to answer your question is “which broadcast monitor?” They vary dramatically in quality, resolution etc. The first monitor I ever bought was a JVC at around $600. Although it’s a little small it works great and is reasonably trustworthy. If you’re on a budget Ebay is an ok way to go, but be aware that many people run monitors into the ground so you never know what you might get. It may need a going-over by a technician before you can properly calibrate it, it could have screen burn-in, missing cord etc. so just make sure you know its story, specs and condition. If the Ebay postings are at all vague, avoid them. They are probably some liquidator that doesn’t even know how to test them so they just toss them on Ebay “as-is.”
Also, make sure it has at least the specs of the monitor I linked to above, anything less and there isn’t much point. It must have a Blue-Gun feature – this disables the red and green guns in the monitor to aid in calibration, and any monitor without this feature probably isn’t a serious option. Also make sure it has (at minimum) BNC connectors, ideally SDI but that may be out of your price range. I would probably focus on Sony, Ikegami, Panasonic and JVC brands, with Sony’s Trinitron monitors being the standard bearer to some degree.
Many will disagree with me and have other, equally valid points. Just make sure you understand the setup, why each component is needed and what it does, and what features are important to getting the quality you need.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
Provided you are talking about RAM and not hard drive issues (as Dave points out) 4GB of RAM is enough to do feature film work. It’s all in how you work. And for the record, you may be confused – If you are rendering out files that end up being 8GB in size, that doesn’t mean you need 8GB RAM.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
Ditto ditto, me three. Render out to Animation codec (choose half resolution in render settings if needed). From there your options are pretty broad for how best to compress, but Quicktime Pro ($29.95) compresses quite nicely and quickly. Using this to create h.264 versions is a handy option. If you have flash video encoder installed, it will even do FLVs. I have compared the resulting quality and render times of various compression apps, and QT Pro is in the top three.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
I did see after writing my post that your camera does some really strange stuff regarding sharpness so sorry for the tone of my note. It sounds like your camera does 30p or 60i (29.97 interlaced) which should be sufficient to help with motion blur, but it’s doing whatever it wants based on your lighting, which is just lame. Your camera is obviously holding you back because it lacks manual settings.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
Select all layers to nest, then press CTRL>SHFT>C (on PCs) or CMD>SHFT>C (Macs).
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
I pointed this potential danger out to you in your previous keying thread, you shoulda listened to me. As I said then, shoot the highest frame rate you can when keying.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
Grant,
Sorry to disagree with you again, but shooting in HD, especially with a consumer-grade camera, won’t necessarily translate into a better key.
Problem 1 – Sure you technically have more pixels to work with, but almost all consumer/prosumer HD cameras, including the Panasonic HVX200, use tiny sensors with a native shooting size of 960×540 (barely larger than standard definition). The camera internally uses scaling tricks to get that size up to 960×720, resulting in the first level of stretching. You can pretty much double the scaling required for 1080p.
Problem 2 – Virtually all consumer/prosumer cameras use the 1.33 aspect ratio to get the “true” HD width of 1920 or 1280 from their starting point of 1440 or 960. This is your second level of stretching used to arrive at “HD” sizes.
Problem 3 – All consumer/prosumer HD cameras have notoriously hideous low-light response when shooting in HD mode. Wild pixelation in the shadow areas is common unless you use several thousand watts of lighting, allowing the DP to stop down substantially.
Problem 4 – Consumer/prosumer HD tends to compress using HDV or similar, which has the same 25Mbps datarate as mini-DV. How a 1920×1080 image can contain only the same amount of data as 720×480 DV and not fall apart in front of your eyes is a miracle, but it certainly tells you what to expect if you try to do any color correction to the footage to match it into your background plate. There just isn’t enough data there to push or pull color values around without severe damage to the image.
In the end, you have a camera pushing, pulling, scaling and stretching your pixels like mad, and when you throw in the motion blur you will get from 24p (or even 30p) you end up with a softer, grainier image with poor edge detail. It may be larger, but the problems this format introduces almost offsets any gains to be had from the larger physical size of HD. I would rather have a DVCPRO50 (4:2:2) 30p file to key than a consumer HD file any day.
—-
I should add that not many of these issues matter as they relate to shooting live footage in consumer HD. These problems are usually fairly buried away in those cases and imagery can look quite nice. It is with keying that you start zooming in very close and examining edge detail, and you start to see real problems.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
“The Hue/Saturation effect works like this: it takes the hue of the individual color channels and manipulates them independently from the other channels, thus you can change the relationship that pixels have with other colored pixels.”
Sure, but boosting the greens in the background boosts the greens in the actor, resulting in a net “nothing” in terms of ability to pull a better key. Trust me, there is no way to color correct your way out of a keying situation.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
Windows Media 9 Advanced Profile (VC-1) is a great option. The bitrate depends on your source material but I find around 5Mbps to be a nice starting point. I also tend to use 1 or 2 pass CBR, Variable bitrate takes way longer to encode and I tend to get more stable, reliable results with CBR when using VC-1 codec.
I would highly recommend using a third party application for creating the VC-1 video rather than letting AE create it. AE is not as good at encoding well compressed, efficient videos as some third party (even free) options can be.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com -
“1)Duplicate layer
2)Remove grain
3)Increase contrast
4)Use Hue/Saturation to change purple to blue (may need to create a mask around the character to keep that from becoming blue)
5)Increase saturation
6)CREATE GARBAGE MATTES
7)Take the key(s) from this and use it as an alpha matte for the original “This technique wouldn’t actually work that well in keylight because it is a color difference keyer. You can push and shove the pixels all you want, but it’s the relationship between them that Keylight pays attention to. Boosting color or saturation to the entire image results in the exact same “difference” calculation.
For this to work and deliver a clean result, you would have to create a detailed, frame-by-frame animated mask around your character to prevent the color corrections from affecting him/her. At that point, you’d have a very nice frame-by-frame rotoscope and keying wouldn’t even be necessary.
Brendan Coots
Splitvision Digital
http://www.splitvisiondigital.com