Dave Haynie
Forum Replies Created
-
MJPEG is a very primitive CODEC. Video formats like MPEG-2 and AVC use interframe and intraframe compression. Some individual frames will be compressed directly, just like a frame in MJPEG (which is just like a single JPEG frame from a still camera), but others encode the differences between themselves and previous or successive frames in a very small vector format, and only have to compress the differences in a conventional way. This makes MPEG and AVC much, much smaller per unit quality… or much, much better quality at the same file size. You would have to go maybe 10x-20x large in MJPEG to deliver the look you’re after.
Compression should be chosen according to your application, not at random. For the web these days, AVC is the right answer. For DVD, you want MPEG-2, at 720×480 (NTSC) or 720×576 (PAL), etc. There’s really nothing much around today that needed MJPEG… I have run into a few robots at work using that, but it’s pretty much been superceeded by other formats.
I would start with the pre-defined H.264 templates in Vegas. There are a few under the Sony AVC CODEC (Vegas Pro has both Sony’s and Main Concept’s AVC CODEC), with a couple “Interent” presets that work very well. If you run one of these, and think the result is wrong, post the specifics here, and maybe we can at least check you math. Note that the Sony CODEC is constant bitrate, the Main Concept may be variable. You can actually use either for low and high resolutions, but Sony’s ships with higher resolution presets, Main Concept with lower resolution presets.
-Dave
-
I would say, upgrade, if you can figure out why you need the upgrade. And also, given the time of the year, I do wonder how far away Version 11 is….
The main reasons I upgrade. First, AVC rendering, while not GPU accelerated, is a bit faster in Vegas 10. I think we’re in a cycle, much like the dawn of HD editing with Vegas 6 or so, where each new release has that code just a bit better. Still hoping for GPU acceleration in V11, that’s now kind of a given in any high-end NLE, but for now, V10 is a win over V9.
The other things was BCC7 support. The old DirectX Filter-as-plug-in mechanism in Vegas was nice and standardized, but the API was weak compared to the various proprietary plug-in standards out there (Adobe, etc). Vegas’s use of OFX (https://openfx.sourceforge.net/) in V10 is the right answer to both issues.
While it may be fairly minor, Vegas now has a feature I suggested years back (and got) for Acid: track grouping. This comes in handy when you have many dozens of tracks. That’s pretty much guaranteed in Acid, particularly if you’re doing a complex piece (I was working on a 20 minute bit of music in four movements at the time I started pestering a buddy at Sonic Foundry about this)… these days, I like it for complex animations. Just makes the whole workflow cleaner.
I also use Vegas for live 16 track recording, and some of the new features (VU meters, input buses) make that much better. I’m still not 100% with Vegas in this capacity, and I have occasionally had it just stop during a recording, no idea why. But until I find a better tool for multitrack recording, it’s pretty good.
What I didn’t like: the new “special” Cineform stuff. Cineform, if you’re not familiar, it a DirectShow/Video for Windows/Quicktime video CODEC, based on wavelet compression, which does an amazing job as an intermediate CODEC. It worked perfectly under VfW and Vegas 9. Sony added support for some custom interface in the newer versions of Cineform in V10… don’t know just what they get there. But this completely killed older versions, cost me an extra pile of $20’s for the upgrade, on top of the Vegas 10 upgrade, that I had no reason to expect. And it just didn’t work. The new Cineform was, as before, just dandy via the VfW interface, but it crashed on output, every single time, in Vegas, until Vegas 10e. And it’s still acting flakey, from time to time. I do like new features in software. I don’t like it when a new feature isn’t working, and yet, there’s no way to disable it and go back to what did work — that’s not the way way to write professional software.
I’m also seeing some flakiness with loading projects that have a very large number of nested .veg files. However, that could have always been with us… I didn’t try to use as many of those in the past (I did recently, largely because I wasn’t able to use Cineform for intermediate files).
The stereoscopic (3D) stuff is in there too, and apparently, that was the big new feature. I have no use for that yet.. nice to know it’s possible, but I’m not convinced the consumer gear is good enough for this to be a popular thing for awhile. Plus, there’s no support in DVDA for Blu-ray in “3D” yet, though you can burn a very basic BDAV discs in “3D” from the Vegas timeline.
-Dave
-
Do they have any recommendations for nVidia chips that are still in production? The GTX260/280 are from 2008, and were discontinued early last year. Current versions are the GTX 400 and 500 series. I’ve been sort of circling the idea of upgrading my old (8800GT) graphics card, but can’t really figure out where performance can be optimized with respect to things useful for video.
Same with AMD.. the HD6000 (Northern Islands) series was out about a year ago, the HD5000 series (Evergreen) late in 2009; the HD 4000 series from 2009 is discontinued.
And yeah, I know these things get upgraded and replaced kind of on a daily basis, and it’s important to list support for GPUs people already have in their systems. But I think there are a bunch of video heads around looking for recommendations on new hardware, too. And in particular, it’s easy to find benchmarks for gaming on any non-pro-market GPU, but very little is published about even general video support.
-Dave
-
Some of it’s just history. When most digital still camera makers decided to support video, they took the path of least resistance. They already had camera processor chips that could generate JPEG, so they supported Motion JPEG.
But, of course, it has to go into a wrapper. The problem was, there is no standard for MJPEG on Video for Windows… different vendors produced different incompatible versions, and Microsoft never bothered with it. But Apple had standardized MJPEG under Quicktime. So that’s why so many digital cameras create .mov files.
And many stuck with it when moving to AVC. And why not — they had the encapsulation code done already (though the MPEG-4 container is very, very similar, having been derived from Quicktime), and Apple already supported H.264 encapsulation.
Then there’s AVCHD. Some HDSLR/EVIL manufacturers do use AVCHD… Panasonic, for one, on at least some cameras. But the bit rate in AVCHD is limited to a peak of 24Mb/s (goes to 28Mb/s for 1080/60p in AVCHD 2.0, but that’s very new). But video out of my Canon 60D can go to 45Mb/s or thereabouts. The AVCHD format is no match for this, and I do think I’d rather have the higher quality video than an easier standard. Wouldn’t kill them to offer an AVCHD mode… or for that matter, continuous video (the Canons all stop recording after 4GB, even on exFAT file systems, which like NTFS and ext4 can support gigantic files).
-Dave
-
Just one more data point… I’ve been doing quite unusual amounts of mixed media animation in Vegas, leaning fairly heavy on BCC7 for effects, and would definitely use the 2D particle filter.
I do realize that the Vegas port was new, and that if I had to wait for every effect to be ported, I’d still be waiting — you do have to prioritize.
-Dave
-
I have most of the NewBlue audio plug-ins, a few of the video plug-ins… no issues.
One common thing: most Sony tools will load the same plug-ins, probably in the same way, particularly audio tools. At least, I assume, if you’re seeing this in Acid or Forge, it’s related to something in the audio world, since they wouldn’t open any video plug-ins, far as I know.
Any chance you’ve got some evil audio plug-in that’s causing problems? Try uninstalling all plug-ins (for VSTs, just try hiding the VST directory by changing its name), but specifically, everything non-Sony (well, you can try one-by-one, but it could take a very long time).
-Dave
-
In Vegas Pro, the action safe and other guides are found on the Video Preview window. Looking across the top of the window (easier to see if you de-dock it from the main Vegas window), on the left, there are a series of icons. Moving left from the rightmost icon, you have “Save Snapshot to file…” (looks like a floppy disc), “Copy Snapshot to clipboard” (look like two tiny windows) and then “Overlays”. Select “Safe Areas”, and you’ll the outer area, usually about 90% of the image, as the “Action Safe” boundary, and the inner area, which marks the “Title Safe” boundary. I can only guess it might be similar in Vegas Studio Platinum.
You can always scale your video to fit within the safe region… zoom out a little in an “Event Pan/Crop” dialog, and you’ll be able to make it fit. The question is, did you shoot it so tight that you’re now left with empty space in the video?
You might actually tweak that with something like a displacement map plug-in… this lets you distort the video in various ways to move parts around without cropping or shrinking (but stretching and bending) the rest of the video. The one I use is part of Boris BCC7, though, not compatible with Vegas Studio, and expensive. Not sure if there’s a free displacement mapping plug-in around.
On a good camera, you generally have the option of setting up some of these overlays while shooting. For example, my Panny HMC40 can show an action safe 90% guide, a 4:3 guide (not likely, but I used to still occasionally shoot 4:3 a few years ago, based on project requirements), things like that. And in general, just be careful about shooting too close.
The other option… just cut people’s heads off, add some artificial camera shake, and pretend you’re copying that evil hand-held camera style that was popular some years back. Ok… maybe a reshoot… but only as a last resort.
-Dave
-
-Dave
-
I’m not sure the world is leaving behind optical media as quickly as Apple would like use to believe.
Do consider: Apple makes a great deal of money selling folks on the idea of video downloads being the future. They’re selling, sure, but just like the iPod AAC downloads of last decade, they’re selling an inferior product for the same price. Apple’s “HD” is 720/24p or
540/30p, encoded in AVC/H.264 at up to 5Mb/s… and it can’t really change, since they have a few generations of iPods, iPhones, iPads, and Apple TVs to support. This is substantially lower resolution than broadcast HDTV (ATSC or DVB), much less Blu-ray.I think online video has a stronger niche in rentals. And that’s the real thing that’s eating into hard media sales. Think about it… when Blockbuster charged $5.00 for a rental, and I could probably buy the DVD on release-week sale for under $15.00, it was pretty easy to figure that, if I wanted to see that film more than once, I’d just buy the disc. With flat-rate on-demand services at under $8.00 a month, or physical rentals via RedBox now $1.00, the effective cost of temporary video has all but vanished. In fact, these services are having as much impact on cable/satellite TV subscriptions as they are on physical media.
Anyone who was around in the 80s and 90s probably remembers the various computer industry disc formats that arose, lived a short life, and then either died or lived on in some niche. CD pretty much replaced most such efforts, DVD and BD ground these into the ground. Not that there won’t be computer-only formats, but that such formats (like DVD-RAM or BD-XL) are more likely to be based on consumer formats.
I think the M-disc is inherently shooting for a niche: archival. And, just as these others, they’re leveraging the existing consumer format, rather than striking out on their own. Which is smart. If they really hit that $3.00/disc price… well, that’s “only” 10x more expensive than BD-R for backup. But for archival… if it’s that reliable, I’d absolutely use it for “gold masters” of every DVD5 I make. I’d still keep ’em all on the RAID, but that’s a level of security you’ll never get with standard DVD-Rs.
As for web-based storage.. it’s for a different purpose, right now anyway. Most people don’t have Dropbox or other “cloud” storage accounts that’ll hold more than a single DVD5, much less a BD25’s worth of data. The point of that is to have a few things available that you can access anywhere. The speed of the network is probably going to be a limiting factor on this kind of storage, no matter how many free GB these services are about to offer you.
-Dave
-
There’s no single answer.
Yes, technically, if you render based entirely on resolution, you’d simply scale everything up by 5x (PAL) or 6x (NTSC) to deliver a full 1920×1080 HD output at the same frame rate.
However, that’s not really done. For one, it’s not strictly necessary. Even for a fairly simple CODEC like MPEG-2, scaling up by 5x or 6x the bitrate isn’t really necessary for a perfectly acceptable video output. This is due to much smaller, relatively, macrocells each having less of an effect on the overall video. So, for example, while DVDs tend to run up to about 8Mb/s (peaks are higher, and of course, the actual value is very carefully tweaked on a professionally encoded DVD), while broadcast ATSC at 1080/60i has a maximum rate of 19.4Mb/s. Both are MPEG-2.
Blu-ray is more involved yet. For one, it can support video bitrates up to 40Mb/s or so, though it’s rarely necessary to go so high. And while you can encode in MPEG-2, the option to use AVC (aka H.264, MPEG-4 part 10) is a big advantage. Well-encoded AVC has about twice the coding efficiency as MPEG-2… that means that, what looks good at 20Mb/s in MPEG-2 ought to look about the same at 10Mb/s in AVC.
Most camcorders, of course, use higher bitrates. The typical consumer/prosumer MPEG-2 HD camcorder records at 25Mb/s, the typical more advanced MPEG-2 camcorder may run 35-50Mb/s (probably added 4:2:2 color once you’re up at 50Mb/s), and AVC camcorders typically support recording peaks up to 24Mb/s for 1080/60i video. That doesn’t necessarily deliver video that’s twice as good as that 25Mb/s MPEG-2 camcorder, though. There are some limitation on just how complex an on-camcorder AVC encoder can get to-date, which is not necessarily true when you’re rending on a PC.
To actually answer your question, you have to flip it around. The best possible rendering of your DVD material, as you note, is when you’ve pretty much filled up that 8.5GB on a DVD9… or, the 4.7GB available on a DVD5. The best you’re going to get on a BD25 is when you’ve rendered the video to fit in just about 25GB worth of space. With that said, there’s certainly not much to be gained by cranking your 24Mb/s original video up to 35Mb/s or something, just to fill the BD25, if it’s a shorter video. Rendering to AVC will probably give you a better result than rendering to MPEG-2.
But keep in mind, Vegas sets some limits on AVC rendering, and the situation is a bit confusing. Sony actually supplies two AVC encoders: Sony’s own, and the MainConcept AVC encoder. The suggestion is to use the Sony encoder for Blu-ray and the MainConcept for small devices (smartphones, iPod, etc). However, the Sony CODEC only supports constant bitrate encoding and single-pass. It delivers a good result, but you could nearly always do better with variable bitrate and dual-pass (the two are related — without VBR, there’s nothing for a dual-pass encoder to do). Pretty much all commercial DVDs and BDs are encoded VBR, which ensures that bit rates go high for fast motion, but slow down when not needed to save disc space.
-Dave