[Dave Haynie]

The usual cause of crashing after some specific editing time is loss of resources. It could be that your PC just isn’t up to the task, it could be a bug in that version of Vegas. And of course, in the case of Vegas 7, you’re just fooling yourself if you expect to edit today’s camcorder formats with yesterday’s software or hardware.

Each new advance in camcorder encoding has delivered higher quality video, but at the price of complexity. HDV was much better than DV in quality, but it wasn’t 2x or 4x as complicated, more on the order of 100x as complex to edit. It’s similar going from HDV to AVCHD.. it’s on the order of 100x more complex to edit than HDV’s MPEG-2. A PC that will verry easily handle HDV can be barely functional at AVCHD.

Fact is, quality costs. The price falls every year, but there is no magic bullet. You spent money on a new camcorder, but that’s only one piece of the toolchain. I spent $2600 on my first HDV camcorder, but that was just the beginning. It took three years, new PC hardware, more storage, Blu-ray gear, etc. before I was done. For a year or so, then I started to mess around in AVC.

If you can’t handle AVCHD today, you can use the proxy editing solution or the intermediate format, both let you get by with fewer resources, for a price. You can even try Avid’s DNxHD, or Sony’s MXF formats, as alternatives, each with its own issues.

Another option is taking back the AVCHD camera, maybe find a high quality HDV model, until you can afford to fully support your whole editing toolchain.

-Dave

[Dave Haynie]

Depends on your computer. Stephen’s suggestion of Gearshift offers the simplest solution for a lower performance PC. Note that AVCHD records at up to 24Mb/s (28Mb/s, maybe a bit more, for AVCHD 2.0), which is a slow stroll for a modern PC, same basic disk requirements as for HDV and DV. A good PC from 1995 has enough disc speed.

But just decoding AVC is heavy lifting even for a modern PC. Just to playback AVC, I’ll see 30% CPU use on 1080/60i, 60% CPU for 1080/60p, using a CPU only player. Editing is harder still on your system. And I run a six core, 3.2GHz AMD system… pretty fast. You need this, or better still an i7, for really nice AVC editing.

Another alternative is an intermediate CODEC like Cineform. You convert AVC to Cineform, edit the Cineform tracks… much less CPU needed. But disc performance and storage is the price you pay… Cineform is about 100Mb/s. Technically, you can lose some quality, but realistically, very hard to tell. You can download the Cineform Neo demo and give it a try.

-Dave

[Dave Haynie]

Stephen has the best answer I’ve come up with. If the noise is actually random, multiple sampling is a time-honored signal-processing technique. You basically scale each sample by 1/N, then sum the N samples. The result is the signal at normal levels, but the noise at 1/N. Now, if the noise isn’t random, but due to fixed defects in the tape, this won’t help.

You might see some advantage to trying it on different decks. I found this to be very true, restoring an old 8mm video I had shot back in the 90s. The original camera wasn’t much good, but another I tried did much better. That old analog thing at work.

Now, back to the hurricane….

-Dave

[Dave Haynie]

The CPU use percentage is an extremely good indication of your system and project tuning, simply because that’s the one bottleneck you can’t get past. Sure, you may be stuck with a single-threaded plud-in, but that’s [a] rare, [b] possible to replace with a better one, and [c] easy to discover… by using the CPU performance monitor. You bottleneck could be hard disc (usually due to seek times coupled with too many high bitrate intermediate files on the same HDD), but at worst, shuffling assets to another HDD or two, and in particular your output, solves that problem. As John suggested, it could even be the GPU, but that is also easily fixed.

Any way, if the CPU isn’t over 90-95% during a render, you have a bottleneck that is probably easy to fix, though it can take some trial and error. I’m on six cores, and I can always keep the CPU as the weakest link. And doing so is the bottom line.. if you’re at 50% CPU, the render will take twuce as long… at least until the GPU becomes a useful bottleneck.

-Dave

[Dave Haynie]

Vegas 10 understands Quicktime just dandy. But like most mulitmedia frameworks, a .mov file is just a wrapper… it can contain all kinds of different video and audio formats. Vegas knows onky the formats distributed by Apple that come with standard Windows Quicktime. You have a larger set available in the standard MacOS set. You can get a larger set of these in Quicktime Pro for Windows, but unless you know the actual format you’re using, there’s no way to be sure you have that support in Vegas.

The best format I have found for interchange at pro quality in Avid’s DNxHD. This is a free download from Avid; it’s high enough quality to support 4:2:2 color and work as an intermediate file format, like Cineform or Apple’s ProRes. Install this on both systems, interchange will become easy.

-Dave

[Dave Haynie]

I had not even noticed that until now… what IS the deal with AAF export being gone from Vegas 64-bit. This isn’t a Vegas 9/10 thing… it follows the 64-bit version of the program.

So, that got me curious. I looked up the AAF stuff. The Advanced Authoring Format is a project of the Advanced Media Workflow Association. The AAF support libraries live over top the Microsoft Structured Storage System, but the main APIs are managed as a project on SourceForge by the AMWA.

And guess what… they don’t have a 64-bit version of the AAF libraries yet. Last year, they had targeted this for November, but as of this June, they still did not have a supported 64-bit version, for Windows or for MacOS. I’m fairly certain this is the only reason Vegas 64-bit doesn’t have AAF support.

Sure, Sony could do the work themselves, but it’s less critical, given that I can load up any Vegas project in 32-bit Vegas as produce the AAF file I need. Presumably, this is just what Adobe did for AAF support in Premiere CS5/CS5.5, which has no 32-bit alternative. The AAF SDK is distributed under what looks like a version of the Apache license. If Adobe made changes to the code, they have to supply that code back to the AMWA, but they don’t have to post it themselves.

So this is very clearly NOT anything like “let’s remove all professional features and concentrate on the prosumer” as we’ve seen in FCP-X. I used Premiere in the past… not a fan, not even remotely. And it’s only Premiere that’s evil… I love Photoshop.

However, I think Sony needs to get their act together on some things Premiere does very well — in particular, Premiere’s high speed “Mercury Playback Engine”, which puts Vegas editing speed to shame on MPEG and AVC native edits. And this, from a program that didn’t even handle native edits, not too long ago.

-Dave

[Dave Haynie]

Only the stuff I suggested in my first post: re-mux the AVC/AVI as AVC/MP4 or AVC/TS. That will let Vegas use its own AVC CODEC, rather than counting on x264/FFMPEG, which isn’t really delivered for editing proposes. Try YAMB, VLC, or ts-muxer for these re-muxing tasks.

-Dave

[Dave Haynie]

I’d probably look for USB 3.0 these days if I could get it… nearly as fast as eSATA, but way more useful. Not a deal breaker, but that would be on my list.

The GPU is a lower-end dedicated graphics unit… better than a shared-memory GPU like the Intels, but it’s not for serious 3D gaming. That product number (GT520M) is either based on a GF108 core with 128-bit private memory bus, 1200MHz internal, or a GC119 core with 64-bit private memory bus and 1480MHz internal clock. Here’s the Notebook Check page on it: https://www.notebookcheck.net/NVIDIA-GeForce-GT-520M.43104.0.html

It should be fine as a video accelerator… it has 48 unified shader cores, and can accelerate at least two simultaneous 1080p decodes. I have a GeForce 8400M in my laptop, which has only sixteen of the nVidia older stream processors, and even with just a dual-core 2.4GHz Core2, it can decode a 1080/60p stream in realtime. Never tried two, but it does do at least three 6Mb/s 1080/24p streams simultaneously (part of a networking demo I ran a few weeks ago). The CPU on mine is not fast enough to do this without GPU acceleration.

If we had reasonable GPU acceleration in Vegas, or it was known to be coming, I might hold out. But at this point in time, that’s a big unknown. That GPU should be fast enough to be a big help on things that use OpenGL like some of the Boris BCC7 plug-ins, if you ever use that stuff.

CPU, memory, and screen are close to the best you can get on a reasonable laptop these days (sure, there are some totally nutter machines built for gamers that use desktop chips and essentially have no useful battery life).

-Dave

[Dave Haynie]

I did something like that ages ago. If you’re stuck with that footage, here’s what you can try. Double the video layer. Now offset the bottom layer down by one scan line, if possible. You can try making the top layer transparent, or compositing as an add or a screen. Here’s an example:

This is actually a photo with simulated scan lines… I don’t have any analog televisions hooked up these days. So here’s a snap-shot from Vegas, with the two layers offset, using 50% transparency on the top layer:

And here’s the two set up as a composite set, in “screen” mode:

You could probably do a better job lining things up.. this was a quickie just to demonstrate the technique. And of course, if you have more than just the screen in your shot, you may have to use mask out the TV and do this in a few separate layers.

This is actually a crude attempt at building a simple comb filter, which is what modern televisions, particular digital televisions, do to de-interlace, remove scan lines, and filter out chroma, all in one operation.

-Dave

[Dave Haynie]

I see part of your problem here… you’re doing things with AVC in a AVI wrapper… that’s where FOURCC codes come from. That’s very non-standard, and you’re probably going to get what you get there.. could be good, or it could be generating defective video files. For professional and most consumer use, AVC files are encapsulated in MPEG-4 (.mp4) files, MPEG-2 transport streams (.ts, .m2t, .m2ts, etc), or Quicktime files. There is no FOURCC concept for any of these formats, the formats are coded internally as streams.

You might find Sony handles the files better if you re-mux to AVC. There’s a free program, YAMB (https://yamb.unite-video.com), which might help here. You can also use the VLC media player to re-mux into a .m2ts or .mp4 wrapper. In both cases, they’re just changing the wrapper, video is not re-coded.

You may not see thumbnails when Windows itself doesn’t understand the file format. Some of these things (well, many of them, these days) are actually handled through the ffMPEG subsystem (another freeware project), which is rather dubiously hooked into the Windows media subsystems. In fact, since Vegas will use Windows resources for any of these things, you’d expect the Windows shell (aka Explorer) and Windows Media Player to have some of the same issues you see in Vegas… this isn’t specifically a Vegas problem, it’s system-wide.

-Dave