Temporarily set the project properties to widescreen SD 720 x 480 and set your preview screen to full size best resolution. Now you can see exactly how the sharpening is going to affect your final downrez. Adjust it until it looks the best. You'll find that you can add a whole lot more sharpening than you would expect.

I don't really believe in adding artificial sharpening usually, but downrezzing is a whole different story. A liberal amount of sharpening during a downrez can work wonders.

"liberal" = A lot, generous, without limits

"conservative" = A little, sparingly, small

Did you mean "liberal"?

Grazie

Surprisingly, yes I did. Again, you have to let your own eyes be the judge, but my experience is that the sharpening filter settings are in a range that is meant for sharpening an image where the input and output resolutions are the same. When you are outputting to the same resolution, even the tiniest bit of sharpening is often too much.

When you are downrezzing, the whole situation seems to change and it takes a generous amount of sharpening get the look you are after, which is basically as little damage to the image as possible during the downrez. Try it and you'll see what I mean right away. Don't overdo it, but be aware that the look of the sharpen changes quite dramatically when you downrez. That's why I recommend to set the sharpen while you are previewing at the end resolution.

I don't know if Farss is referring to the same thing, but a common practice is to turn down sharpening in the HDV camera (they're all oversharped to begin with), and then add minor sharpening back in during post.

That's pretty much what I did although I could have probably turned the V1P down more. I wound it down to 5 and used the Light preset in the Unsharpen Mask while I'm downconverting to 16:9 SD PAL.
I didn't want to push the sharpening too much in either the camera or in Vegas due to the amount of noise but as Laurence says you probably should evaluate the amount of sharpening on a case by case basis.

There seems to be some good deals to be had down here on a V1P + DR60 combo. I know, compared to the EX1 the V1P mightn't seem all that sexy but for events like this it's still a pretty good, relatively cheap combination. As I had to shoot for nearly 3 hours without a break if I was using the EX1 given space constraints offloading cards during the shoot would not have been practical so I'd have needed 3 16GB cards and that's a lot of capital investment.

Bob.

Shouldn't you apply sharpening after downconversion?

Applying before will make aliasing worse, whereas applying it after doesn't.

The exact same thought had crossed my mind as well.
When I get a chance I'd be interesting to do a comparative test, especially with this noisy footage.
Back around the V6 days there was some advantages to doing everything in HD and then downscaling but I think there was a subtle change in how some of the filters worked so I'm not so certain now.

Bob.

I add the sharpening during the downconversion. To my eyes that seems to look best. If you take my suggestion and preview the video at best quality at the target resolution and add sharpening as you step through frames, you can get the right amount of sharpening pretty easily.

Hmm it might be that sharpening usually defaults to after pan/crop (therefore, after downconversion).

You are talking about a pre/post thing. I don't know how Vegas does it by default.

Vegas defaults to post. Pan/crop applied first, then filters. You have to click the little triangles in the eventFX window to get pre (before pan/crop).

Thanks for the sharpening tip. I just tried adding the sharpen filter to the video output effects and what I find strange is that even at the "0" setting the filter sharpens the video. This is confirmed by enabling and disabling the effect. Anyone know why zero is not zero?

Marc

"Anyone know why zero is not zero?"

I seem to recall that being noticed as a bug a some time ago.

Bob.

"Shouldn't you apply sharpening after downconversion?"

No, as the HDV is 420, when downscaled ends up being closer to 422.. in turn, you want that master to be as pristine as possible so sharpening DURING the downconvert will yield a better 422 source
it will come VERY close to DV50 in aliasiing and edge contrast
IMO, its actually nicer tean DV50 simply due to the aspect ratio being "rounder"

The unsharp mask is the better alternative as it allows for radial pixel manipulation as opposed to a straight cut contrast differentail which normal sharpening provides.

DJ: Huh?

I don't think we're on the same page here.

I'm talking about
footage --> pan/crop / resizing (Vegas does this automatically / render or project properties affects this) --> sharprening --> render.

So depending on how you look at it, that sharpening is happening ""during"" downconversion.

were on tha same page mate.. but the question was specific whether or nto sharpening should occur during the conversion or once its converted (ie sharpen the lower res final)

I suspect you're correct but haven't personally tested this.
It used to be that for example GB was applied at the project resolution and then the resizing was done but that was way back in V5 or V6 days. For certain now GB is applied at the destination resolution. I haven't tested this to see if the yellow triangle can switch this back, I had thought that only applied to pan/crop etc but you could well be correct that resizing fits into the same part of the flow. I only mention GB as that was one FX I used a lot in this regard, I assume any FX would work the same.

What PJ is talking about is another issue and certainly one worth keeping in mind, on another page though.

The overall issue I think is kind of interesting in relation to the V1P/E, which is subtly different to the V1U. In 25p the V1P/E has arguably too much vertical resolution and depending on the footage one might be looking at using GB or the Median FX to prevent problems of line twitter. Just to further confound the issue the V1s uses level based DNR and the in-camera sharpness setting can cause problems with that system.

All that aside, assuming by one means or the other one was to apply the sharpening at the source resolution before downscaling specifically what sort of aliasing are you concerned about?

Bob.

To me, you can't apply sharpening at the same moment you downsample. They are separate processes. You do one and then the other (or vice versa, since we're discussing the order).

*you might mean that you pick a different algorithm that gives you sharper results. But in Vegas your choices are good (bilinear) and best (bicubic something)... best makes more sense for downsampling / scaling down.



The normal kind of aliasing / the aliasing that occurs when downsampling. You can throw a zone plate test image to see aliasing happen.

http://www.worldserver.com/turk/opensource/#ZonePlate

1-Turn down sharpening on HDV cam. IMO, they're all too sharp. Like adding high frequencies to crappy MP3 encodes, which nicely increases detail, but makes some details brittle and harsh. Sharpening is much the same.

2- Capture as HDV.

3-Edit, and apply sharpening in post. Convolution Kernal can be tricked out to give you exactly what you want, save the preset.
While editing, view on an SD external monitor if SDMPEG is your final destination.

Ah yes, THAT one!
The bane of my existance over several projects!
You really need to watch out for this in Vegas, the way Vegas does the downscaling keeps as much resolution as possible with a decent execution speed but enough HF detail can sneak through to cause some interesting effects, like 70% of the frame blinking at around 1Hz.
I did ask for an optional better filter but I think that's no small ask, the better it gets the slower it runs and in the end it wasn't that much of a chore to fix manually. I think our landscapes are rife with things that really bring this problem on.

Bob.