[Dave Haynie]

[Shoestring Videos] “I could be wrong but I thought 5DII shot AVCHD and when imported onto a windows system would create MTS files.”

Nope. Canon video is not AVCHD compliant. It’s AVC + PCM audio in a Quicktime wrapper.. that’s the format recorded to the SD card. That’s 44Mb/s AVC, I/P frame only… some of the newer models can also do an AVC Intra (I-Frame only) up at 100-something-Mb/s. That’s on the MkIII, but not the MkII.

-Dave

[Dave Haynie]

I use a Canon 60D, which, like all of the Canons (up until this year, anyway) records in the same format.

Canon shoot AVC at 44Mb/s and store in an Apple Quicktime file, that’s the .mov file you see there. The only weirdness is actually working in most users’ favor — it’s I/P Frame only AVC, not I/P/B, so it’s a bit easier for the camera to decode.

You need the latest Apple Quicktime installed on your PC to read these in programs like Vegas. You didn’t mention the NLE/version he’s using, so that’s my initial guess. Older versions of most any NLE will probably have some issue with some kind of MP4/AVC video file… could be he’s just “lucky”.

And, well… rendering through WMV is the absolute worst possible solution. The WMV file wrapper is very much not intended for editing, and the WMV9 video format is inferior to AVC at the same bitrate. Plus, in re-rendering it, he’s needlessly adding compression artifacts, crapifying the video. And wasting time.

-Dave

[Dave Haynie]

Hmm… other than the obvious interlacing effects on the PC, this plays at full speed, totally worst-case, on my PC. That’s with the preview at Best(Full), and the project set to Best rendering quality and 32-bit. No effects applied, Vegas 11, GPU acceleration on, and with an ATi HD6970, which isn’t all that different than yours. My CPU is an AMD 1090T, six full cores at 3.2GHz… maybe a bit faster than yours (you have six integer cores, but only three FPU cores, in essence), but you should be plenty fast.

Some folks like nVidia. When Vegas 11 came out, I benchmarked the HD6970 versus the nVidia GeForce GTX570… both came in at about $300. The AMD/ATi won every test, sometimes significantly. Both worked fine in Vegas, but the nVidia crashed on a number of OpenCL tests and benchmarks as well. Yeah — that’s going to be totally due to the software, and it’s probably already fixed, but that led me to select the ATi, and I’ve been quite happy with it. So I think you’re ok here.

Assuming you see this ok in Vegas 10 and/or a media player like Windows Media Player (which will play AVC and MPEG-2 with video acceleration on Windows 7, so virtually no load on a good PC).

I know OpenCL doesn’t really handle SLI/Crossfire in the way folks expect it. That should show up in Vegas as the two different OpenCL devices, not one unified double-speed device, when you have acceleration on, but that’s one place I’d look. What happens when the GPU is off, or you disconnect the Crossfire cable?

Could also be that Vegas is just confused about the resources on the Bulldozer chips. Make sure you have all the AMD and Windows patches for Bulldozer. I have a four-core Bulldozer as my lab and multiple audio recording rig, but I have not tried it with video at all. Works just dandy at recording 16 track audio (96kHz/24-bit)… yeah, I know that’s not being helpful. And in fact, that’s all integer. Video decoding and rendering is FPU intensive.

-Dave

[Dave Haynie]

The Vegas CODEC system kind of sucks. Vegas often launches an external program (plug-in) to do the final rendering. Unfortunately, very little information propagates back to Vegas when the UI allows for a setting that the external program doesn’t allow, or worse yet, fails on.

It sure looks that Sony AVC is limited to 1920×1080 or thereabouts. I can’t recall trying anything larger, and sure, enough, in Vegas 11, I get an error.

Moving on to Main Concept AVC, I was able to render to 2048×1080 just dandy. I started with a Blu-ray preset (which probably don’t exist in Vegas 9, but you can use most of the seeings from the Sony Blu-ray preset), set the profile to “High”, and the resolution obviously to 2048×1080. I would be pretty surprised if you couldn’t render likewise to MPEG-2 at this resolution… not certain, though, there may be Vegas 9 issues at work, too.

-Dave

[Dave Haynie]

[Stephen Coulson] “One question, considering my source was filmed in 1080 25p would you recommend rendering out to 1080 24p or 1080 50i? Will the fact I will drop a frame every second if I render to 24p cause audio sync issues or will Vegas sort this out?”

Vegas will drop frames or interpolate (depending on what you told it to do), so this will technically work. Not what I would do.

It’s also been common for 25p video to be sped up to 24p and delivered in NTSCfilm format (with 3:2 pulldown added to display at 60i as needed) for NTSC countries, commerically. They do pitch-neutral speedup on the audio to fix the audio issues. This would also work, but it’s not necessary, unless you’re trying to target US players as well.

-Dave

[Dave Haynie]

I would render the 25p as 50i for Blu-ray. That’s generally what’s done for 30p shot on NTSC cameras as well. On a classic TV, it’ll look find, just with perhaps a more film-like cadance. On a modern TV, it’ll look exactly like 25p, since it will be — modern displays are all progressive, and pretty much any DLP, LCD, or Plasma display is going to upconvert any incoming video into 50p for most DVB countries, 60p for ATSC countries (noting here that NTSC and PAL are both effectively just historical footnotes right now).

-Dave

[Dave Haynie]

[Stephen Coulson] “I want to render to a format as close to my source as possible. When I try to render to AVCHD I am unable to select ‘Progressive Scanning’ and I’m forced to accept ‘Interlacing’ (Field Order: Upper Field First). To get progressive scanning with audio included I’m having to use *.mp4. As my source is in 1080p I don’t want to have to render to 1080i, as I understand it this will lower quality and cause those dodgy lines when panning.”

Well… a couple of things. Check the AVC Profile and other settings. If you’re changing settings but keeping Blu-ray compatibility, you don’t have a 25p option.

I suspect that’s your problem. Blu-ray supports only 50i, 60i, or 24p at 1920×1080. This is less of a problem than you might think, since nearly all modern displays de-interlace on playback (LCDs and DLPs don’t support any form of interlacing at the hardware level). So while the film will be stored as 50i, it’ll play back as 25p most of the time.

-Dave

[Dave Haynie]

[John Rofrano] “I keep 2 copies. I would never trust a single hard drive.
“

That’s smart.

For large projects, I have a project drive. I have SATA carrier slots for one 3.5″ drive and two 2.5″ drives. I use these as working project drives during an edit. When I’m done with the project, I make sure all assets, .veg files, etc. are on the project drive. I copy the finished media to my RAID, and the complete project as well (that will not live forever on the RAID, but I have enough space, or too little work, so I have these around for a year or two). I also archive the critical stuff to BD-R.

[John Rofrano] “I plan to get an enclosure that can support RAID 5 like the G-Tech G-SPEED Q so that I only have to copy it to one place and know that it’s safe from a single drive failure but until then, I have two 2TB hard drives that I copy all projects and raw footage onto (twice one for each drive).
“

It’s important to know the difference between RAID and archival. The RAID ensures you have your stuff very readily available. But never assume that it’s immune to loss on the RAID – you still need backups, archival copies, whatever you choose. Your RAID could perish in a storm, get fried by an electrical glitch (lightening strike, bad power supply failure), etc.

And it can get damaged in software. Naturally, they run some filesystem… usually NTFS on Windows systems. That can get soft-corrupted. You can lose data, not though a hardware fault necessarily, but it can get corrupted. I had a problem on my RAID awhile back where Windows couldn’t fix a problem with CHKDSK, and I did occasionally see errors. I actually had to buy a spare 2TB drive to have enough spares to backup and reformat easily — and of course, I lost the protection against single drive failure that day.

Another thing… I use a Drobo. That’s kind of a smarter version of RAID. It can use different drives from different vendors, while many RAIDs really want exactly the same version of the same drive. The way Drobo does it is good for several reasons. For one, while the base unit is pricey, you can start with just two drives. At that point, it behaves much like a RAID1 (Mirror). As you add drives, it behaves more like a RAID5 (or RAID6 in the newer models). In all cases, you get immunity from single drive failure (dual drive failure in the RAID-6-like mode, not supported on my unit).

There are two things the Drobo, after quite some time, buys me over the typical RAID. For one, when I set it up, I could use a mix of drives I had on hand… originally two 1TB and two 512GB drives. But if I was buying new drives, you have a bit of concern over MTBF issues. In short, if I have four identical drives from the same lot, there’s an awfully good chance they’re going to fail at about the same time. A typical RAID may take a few days to rebuild once a drive fails. You really don’t want that second drive failing at the same time. So buying a mix of drives — even different lots of the same drive — dramatically lowers that chance.

Eventually, I upgraded to four 2TB drives, but that was over the course of some years. Since the drives don’t have to be matched in any way, if one dies, I can replace it with a larger drive today… a recent software update now allows drives up to 4TB, a total of 16TB raw storage.

-Dave

[Dave Haynie]

[Matthew Jeschke] “Oh interesting. I never thought of it that way but that makes sence :)”

Basically, a larger drive will pack more data per track, and this has been the primary reason drives have increased in speed. There have been 7200RPM drives for decades… I had a 2.1GB Seagate back in the early 90s (cost about $1200) that could manage about 10MB/s peak. Today, any old 5600RPM drive is likely to approach 50MB/s, maybe 100MB/s reads. It’s not spinning any faster, but the amount of data read per spin is dramatically increased. This means more performance per revolution, and fewer seeks per unit of storage.

Plus, you need a ton of storage for video anyway. I would generally recommend speed over size if your main goal is large multi-track audio projects (seeks become the gating factor), but you’ll get more speed for the usual demands of video on a larger drive, even if it’s slightly slower.

And as you’ve correctly noticed, if you can keep your CPUs near 100% during a render, you have fully optimized you system. Computation is the only thing that SHOULD be a bottleneck these days, at least on most projects (I’ll admit to having done a few animated projects with 50+ tracks, including some raw HD tracks… this can clobber CPU and HDD alike).

-Dave

[Dave Haynie]

[John Rofrano] “Yes, it is probably softer. Think about it. You’re using the same sensor to record more resolution. How could that possibly work? Since the sensor doesn’t magically grow new pixels, the same number of pixels are being used to represent more data so the information still must be packed into the same number of pixes and therefore the data gets “fuzzy” and so does the picture. “

Well.. 16:9 mode isn’t necessarily “more” resolution. It’s actually just a different interpretation of the same resolution: it’s 720×480, no matter what. If you show this on a 4:3 display with letterboxing, the horizontal resolution is identical, the vertical resolution will be cut down (except for a very few analog sets that could adjust and deliver full resolution in 16:9 letterbox mode — actually fairly common in PAL countries, but not here). If you display on a 16:9 screen, you’re covering more horizontal area, but less vertical area, than a TV of the same diagonal size in 4:3.

But actually, the real problem here is that the PD-150 had a crappy 16:9 mode. It doesn’t actually deliver full resolution in 16:9 mode, it crops the 4:3 image to 16:9, then scales up the vertical resolution to 480 pixels to go on tape.

This wasn’t uncommon in the early days of digital video. My first DV camcorder did exactly the same thing: it would either crop and zoom or letterbox, my choice, but it didn’t have a real 16:9 mode. I had a slightly less old Digital8 Prosumer class camcorder that addressed this: it expanded the horizontal by using the spare imager pixels normally used for digital image stabilization when in 16:9 mode. Once we all (well, most of us) moved to HD, we got real 16:9 resolution. And actually, the sensors are usually still 4:3, it’s just that they’re designed to be cropped to 16:9 in normal operation (and of course, some of the earlier HD camcorders used lower resolution chips with pixel offset to approach to full HD without killing light sensitivity as much as other camcorders did, but that’s another tale).

-Dave