[Mark Allen]

I remember that other fix you’re talking about. That was the first thing I found while looking for a solution. The original version had you edit the registry to tell it how to fix the pixels. Yikes. Okay, I have experience editing the registry but still, Yikes. And then someone added the less-hostile version where you edited a text file. And then it got abandoned.

[Mark Allen]

The first thing I learned while writing PixelPatcher was that it wasn’t that difficult to hide hot pixels in moving video except for large clumps of them. But if you freeze-framed it and knew where to look then I wasn’t ever going to get a perfect result without some kind of intelligent algorithm. And yes, I freeze frame and I know where to look. And I tried to see if I could come up with intelligent algorithms but they just didn’t look to be worth the trouble if your real priority was moving video.

The short answer to eliminating the “smear” is that you set either of the radius params to 0 to stop all the blurring. At that point what it does is start from a bad pixel and then search out to find the closest good pixel and then copy it into the bad pixel. Well, pretty much. There’s a third radius parameter which defaults to a very weak 0.5 pixels. That third radius is only visible if you go into the .ini file and set DebugEnable=1. The .ini stuff is at the end of the documentation. The third radius is an internal parameter and you need to enable debug to see it but it’s always actually there. So if you set all three radius params to 0 then it just finds the closest good pixel and copies it to the bad pixel. It might look a little less harsh if you set the two visible radii to 0 and the debug-only radius to about 0.1.

That’s how PixelPatcher started out. There was no blurring. But my 7D has at times had big clumps of hot pixels and filling in bad ones from the closest good ones looked a little off in big clumps with the kind of high constrast edges you’re describing. That was only when I was freeze framing or specifically looking for them. If it was a low-contrast area it looked fine. But if the clump crossed a clear edge then it often looked bad while freeze-framed. So I added the radius params to try to smear it a little bit for the large clumps and then defaulted to a relatively small 2 pixel radius which seemed a good compromise. There’s little point in automatically turning the blurring off in low-contrast areas because it’s not really changing the result very much. It’s those areas with edges in it where the blurring makes a difference.

If you’re actually noticing the problem in moving video and not just while looking at one frame then I’d just try setting the radius params to 0. For single pixels you don’t need the blurring. It’s really only there for bigger clumps. If that works for you and all you have is single hot pixels then you’re done. It you have both single hot pixels and bigger clumps then you could make two different hot pixel image files: one for the single hot pixels and one for the bigger clumps. Then you hook up two copies of PixelPatcher where the single hot pixel copy has the radius params set to 0 and the big clump copy has radius params set to an appropriate value. For my size clump that would be bigger than 2.

What I did was default the radius params to compromise values which worked okay for both individual hot pixels and large clumps. So you end up with a little blurring applied to both kinds. I figured if you really wanted separate blur radii for different hot pixels then you’d run multiple copies of PixelPatcher on the same video, each with a different amount of blurring. I guess I could have adjusted the blur radii automatically based on the size of the hot pixel clump. Tiny ones get nothing and big ones get something. But I figured anyone who wanted that would just control it manually by adding more than one copy of PixelPatcher.

Oh and by the way, if you actually use more than one copy of PixelPatcher on a video it’s a good idea to set the .ini file var VegasBGR=1. It can actually speed things up quite a bit in some setups. Doubly so if you’re running two copies of PixelPatcher. Unless you have a ton of hot pixels there’s very little processing involved compared to your average filter. Almost all of it is overhead Vegas goes through to pass the data to an OpenFX filter. The VegasBGR switch reduces the overhead a lot. I didn’t want to turn it on by default because it’s not strictly OpenFX compatible with it on.

[Mark Allen]

Yea, Vegas 9 had lots of problems with Canon SLRs. Vegas 10 fixed ’em. Cineform NeoScene was actually quite nice but I think it’s $100 if you shop around. Vegas Platinum can be had cheaper than that. Personally, I’d just get the newer version of Vegas. Just try the free Vegas demo first to make sure it doesn’t have any other problems. Vegas 11 is relatively new. The demo is fully functional and works for 30 days.

Trying to use an older editor which predates the video format of a newer camera can often cause problems. It’s best to use an editor which came out after the new video file format was released. At least that’s been my experience. It gives them a chance to work the bugs out. It also means you have to spend some money on upgrades from time to time if you get a camera which uses a newer format. That just seems to go with the territory when they’re coming out with new video file formats from time to time.

[Mark Allen]

I’m not sure if this is what you’re describing but your problem might be a computer RGB – studio RGB problem. If you don’t have experience with the subject, you can do a forum search for “computer studio RGB”. It’s a common problem which shows up when you have a video format which uses the range 16-235 to represent the video levels and you display in the Vegas preview window which displays the range 0-255. The preview window shows less contrast than the original video. The darks don’t look so dark anymore.

You can get Vegas to convert from studio RGB to computer RGB by applying a Color Corrector filter and then using the studio to computer RGB preset. You can just apply it to the preview window if you don’t want it to affect your video.

But the subject can get more complicated because there’s the issue of whether you should edit your video in 0-255 or 16-235. I have a project which involves two cameras: one is 16-235 and the other is 0-255. That’s why I got stuck dealing with the situation.

But if you’re just doing basic edits from one camera then working in the camera format is probably easiest and you can make the preview window look better by applying the effect to the preview window rather than on your video clips (uhh – events).

[Mark Allen]

I used to run platinum 9 many years ago and it’s good with HD. It’s not so good with the kind of video format you’re dealing with. I upgraded to Vegas 9 Pro which worked great in HD with my HDV camcorder. But then I got a Canon 7D and the stutters and red screens began. A 7D and T2i use the same video format and should behave about the same in Vegas.

People suggested trying Cineform which required a translation step and much larger files. It solved my problems in Vegas 9. But then Vegas 10 came out and it made my 7D problems go away. Vegas 9 predates the camera you’re using and in my experience was unusable on a large project with the 7D. Cineform fixed the problem but costs more than the platinum version of Vegas. Vegas 10 just made the 7D useable and previewed fairly smoothly to boot. Plus there were no translation steps. It just works directly on the video files from the camera. Granted, you still need a powerful machine to get smooth results from the 7D. Their codec requires a lot of CPU or GPU power to handle. But Vegas 10 fixed it for me.

If your time has value then I’d just try the demo of Platinum 11. It should make your T2i very useable. Assuming none of the unusual 11 problems I’ve been reading about rear their ugly head. That’s why I’d try the demo first. But Vegas 11 will be way better at handling a T2i video format then Vegas 9. I still have bad memories about the stuttering and red preview windows that I had with Vegas 9 and my 7D. Vegas 10 fixed all those problems.

[Mark Allen]

I’ve often found that the built-in computer RGB to Studio RGB and vice-versa filter presets just don’t get the job done and I have to resort to brute force. I generally bring up the waveform video scope and then fiddle with levels to try to match up the minimum and maximum luminance. If that’s not enough then I break out the color curves filter and adjust the curves for RGB all at once. With the help of the video scope I can usually match the two shots up. It sucks but after enough cycles I can always get them to match.

I wrote a stuck pixel OpenFX filter a while ago to fix a large purple blob which showed up in my video one afternoon. Since that worked so well I’m about to write a color remapper filter to try to get an accurate match for shots with seriously mismatched colors. Now that you’ve reminded me of how many times I had to struggle through matching up mismatched luminance, I’ll be sure to remember to include a luminance-only mode and see if it can’t simplify dealing with this kind of problem. I’ve spent enough time trying to match up videos from different cameras that it would probably be worth the small extra trouble. I probably won’t have it ready for a few weeks though.

[Mark Allen]

You may have gone from Computer RGB [0..255] to Studio RGB [16..235]. The preview window is [0..255]. If you output to [16..235] used by lots of video formats then the contrast is increased. The darks get darker and the lights get lighter. There’s a bunch of threads on creative cow on the subject.

If that’s your problem try searching for threads with “computer studio rgb”. It’s more complicated then you might think. I’m working on a large project with video from a Canon HV30 and a 7D. The HV30 is studio RGB and the 7D is computer RGB. I decided to work in computer RGB and then add filters at render time to convert to studio RGB if needed. But that’s just one way to get it done.

[Mark Allen]

“My question is: what is the bottleneck? If neither disk nor CPU nor memory is maxed out, what am I waiting for?”

You need to include the GPU on your list of resources because it can also be your bottleneck. Especially with Vegas 11. There are lots of GPU monitor utilities you can use to see how busy it is. GPU-Z is one. There are plenty of others which will display GPU loads.

[Mark Allen]

“With some projects, Vegas takes a long pause before responding to the mouse. This happens when I switch away from the application and come back. It also happens when I first load the project. I didn’t mind having to wait around 10 seconds sometimes for the GUI to come back, but now I’ve been waiting for about 15 mins and still nothing.”

I’m never seen 15 minutes but I used to have to wait a couple minutes on one computer after switching back to Vegas from another program. There’s something you can try to improve that for program switches. Go to the Options menu, click Preferences, and then go to the General tab. Then uncheck “close media files when not the active application.” After that it behaves like you haven’t switched away from Vegas.

[Mark Allen]

Just a wild guess here but I’ve seen programs give out of space errors if there’s too much space available on a hard drive. They get an overflow while doing the calculation and get a negative result instead of a positive one and think they’re out of space.

I have an old version of PhotoShop which thinks it’s out of space when there’s more than 1 TB of empty space on the drive. My solution was to make a “Filler” directory and copy a bunch of video files in there and bring the available space below 1 TB. After that, everything’s fine. Then I delete part of the filler over time as I add more files to the drive.

I’d try putting some filler on your drive and take it below 2 TB of empty space. 3 TB drives are fairly new. Their installer might have some outdated code in it.