Some lessons learned as I moved into HD

smhontz wrote on 3/30/2012, 7:58 PM

I've been using Vegas since V3, doing SD for the last 10 years or so. This past October I got together with a friend that I've known for 25 years and we decided to follow our passion to create videos. We took the plunge into HD and now that we finished our first big project I thought I would share some of the things (good and bad) I've learned that may help others going down this path. Note: I am not promoting any of the add-on products here; I just thought you might like to know what I used to solve some of the common problems with HD.

(As a point of reference, I have an Canon XF 105 and a Canon XF 305. Everything I've shot so far has been 1920x1080 24p. I just worked directly with the MXF files on the timeline.)

The Good:
1. Having a camera that uses CF cards is great. Much quicker to import, and I can use the Canon XF utility to look at what I shot without actually importing. Plus, the camera keeps meta information so I can see at any point what settings the camera was using at that exact moment for aperture, focal length, white balance, custom settings, etc.

2. You can really push in on an HD frame in Vegas without losing quality. This is great if you need to recompose a shot after the fact.

3. You can pull some really nice freeze-frames from the HD footage. To deal with any noise in the image (see the Not So Good below), I use the fantastic Topaz bundle for Photoshop.

4. I shot a lot of green-screen interviews. The XF 305 shoots 4:2:2, so it has a leg up in the green-screen department, but the biggest bang for the buck I got was switching from the Sony ChromaKey to the one in Boris Continum Complete.

The Not-So-Good
1. The camera breaks shots up into 2GB chunks. This makes it kind of a pain on the timeline because you have to remember to group the chunks together. And, if you're using a cut that's at the very end of one chunk, and you want to extend it, you have to go get the next sequential chunk and butt it up against the previous one.

2. HD takes up a lot of disk space - it takes about 32GB to hold 82 mins of HD footage from this camera. I ate up my 1TB disk on basically two projects. Now I have a lot of portable hard drives that I'm going to archive footage to.

3. HD shows a lot of noise in low-light situations. This camera is pretty sensitive, but you do have to boost the gain quite a bit. This problem can be solved in post using the great Neat Video denoiser plugin.

4. HD shows everything. It's not very flattering on faces (particularly women) so it's good to have a plugin that can be used to soften the image a bit. In high-noise situations, the NeatVideo plugin also helped soften the faces; where I had good lighting, I used the Boris Gaussian Blur. Others have suggested Magic Bullets "Cosmo" look and I'm going to be checking into that.

5. Vegas does a pretty crappy job when you want to downrez from HD to SD to make a DVD. I read a lot on this forum and at DVInfo.net, and ended up with a workflow where I render my project using the Lagarith lossless compression codec, process it through VirtualDub to rescale to DVD size, then bring it back into Vegas to render it to MPEG2. This made a great-looking DVD, and cleaned up all the twitter/shimmer/weirdness I got when I tried to go directly from Vegas to MPEG2.

Hope someone finds this useful.

Comments

PeterDuke wrote on 3/30/2012, 10:25 PM

"1. The camera breaks shots up into 2GB chunks. This makes it kind of a pain on the timeline because you have to remember to group the chunks together. And, if you're using a cut that's at the very end of one chunk, and you want to extend it, you have to go get the next sequential chunk and **** it up against the previous one."

This is another well-worn topic.

You are strongly advised to not simply drop multi-chunk clips straight onto the timeline.

I prefer to use the transfer software that came with my camera to transfer the files from camera to computer. All multi-chunk clips are glued back correctly into proper clips again without me using one brain cell. I suggest you look at the software that came with your camera. There are other ways to concatenate files too, including the old COPY command.

smhontz wrote on 3/30/2012, 11:31 PM

It's not that big of a deal - I just wish the camera kept everything as one file per clip. (Although I do understand the technical reasons why they can't.)

I really don't see the point in using the Canon XF utility to export the clips into one file. Then I have yet another copy of all the files, taking up space. Yes, it would be easier to work with, I suppose, but I can't afford the disk space, nor do I want to wait for it to do it.

My workflow is to copy the entire Contents directory from the CF card to the hard drive. With the Canon XF utility, I can set that directory as virtual media, so I can freely browse the clips and look at the meta data. When I want to work with the clips in Vegas, a simple windows Search for MXF against the Contents directory gives me all the clip files from all the subdirectories, which I can easily order by name to preserve the sequence. Then I just drag and drop them into Vegas.

Is there some technical reason why you say "You are strongly advised to not simply drop multi-chunk clips straight onto the timeline"?

PeterDuke wrote on 3/31/2012, 12:19 AM

You missed my point. You don't start by copying files from your CF card, you start by copying from your camera (with card inserted). So you only get one set of files on your hard disk. The camera+card combination contains info to tell the transfer utility how the files should be concatenated.

If you use some other software or utility to concatenate files then yes you will end up with two copies, and furthermore you have to think about which files to concatenate. You only need two CF cards worth of files maximum on your hard disk at any one time, and after concatenating you can delete the chunks. If hard disk space is so tight that this is a problem then I suggest you immediately add another. You should always have plenty of hard disk space when video editing. 2 TB hard disks are only about $100 to $150.

If you don't concatenate first then you run the risk of introducing glitches at the chunk boundaries. A second or subsequent chunk does not start at an I-frame necessarily, so the preceding frames could be garbled. The 2GB chunks are merely what you get when you cut a larger file at a certain byte count regardless of the file's content at that point.

amendegw wrote on 3/31/2012, 6:25 AM

Good summary!

Re: the 2GB "problem" - have you tried the Sony Vegas "Device Explorer" (Ctrl+Alt+7) to import your clips? That should solve the problem.

Re: HD->SD reszing - since you're shooting in 24p, have you tried using the MainConcept "DVD Architect 24p NTSC Widescreen video stream" template for producing your SD DVDs? I've had great success with this, however, YMMV.

...Jerry

TeetimeNC wrote on 4/1/2012, 6:09 AM

Re: the 2GB "problem" - have you tried the Sony Vegas "Device Explorer" (Ctrl+Alt+7) to import your clips? That should solve the problem.

I use this with my Panasonic HMC150 footage and it works perfectly.

/jerry

brant wrote on 4/15/2012, 5:48 PM

Device Explorer doesn't even work with all versions of 11.0. I too have a HMC150, it neither recognizes the camera, nor the external USB connected SanDisk SDHC reader, nor any of the hard drives. It always says "No device connected", no matter where I "browse". It never sees the files so that it will paste them all back together again.

I use my old Vegas version 9 to use Device Explorer. It works fine there. Sees the camera, external reader, and all the 4gig multiple spanned .mts files even if I copy them to the hard drives. It pastes them back together just fine.

And yes, I did report this to Sony. I'm using Win7 64 bit.

TheHappyFriar wrote on 4/15/2012, 8:39 PM

I've started capturing in chunks if I have the option (I'm using HDV tape). Why? Because Vegas seems to handle it on the TL better then one large chunk (I'm talking ~an hour+ here, TL response seems greatly improved).

I just drop/drag the whole kitten caboodle to the TL (the file you have the pointer on when click/dragging is the first one on the TL).

There's the same disadvantages listed though, but I like the speed better.

I've never had an issue rendering my HDV projects to a custom Mpeg-2 DVDA compliant file & playing it back. No complaints from anybody (and I've got some complainers. I get complaints over a fraction of a second timing, any type of judder, brightness/darkness, size of subject in relation to frame size, etc).

PeterDuke wrote on 4/15/2012, 9:16 PM

For HDV the best approach in my view is HDVSplit, which transfers each scene into a separate file, named according to shooting date and time.

The chunks in AVCHD, however, are not necessarily based on scenes. Any scene file which would be longer than 2 or 4 GiB is arbitrarily split at that boundary, regardless of content.

MikeLV wrote on 4/17/2012, 7:46 PM

The utility that comes with the Canon XA10 does not stitch the files together, I think it's called Pixela or something like that, it basically stinks. The only way you know what clip goes where is that the camera will save the file with a _1 indicating that it's the next part of the previous file of the same name. I had to download some other freeware utility to stitch the clips together, just adds time to the work but at least it's not real time capture like it was with miniDV tapes......

Former user wrote on 4/17/2012, 8:56 PM

Yeah, the Pixela software is pretty worthless. It does not give you the option to select the files to download. You can either download all of them, or the ones that have not been downloaded yet.

Really cheesy.

Dave T2