Vegas for Broadcast? Very frustrated

Back to post

Comments

Coursedesign wrote on 1/31/2009, 3:57 PM
Broadcast clamps have given many video editors a bad reputation.

Why?

Because beautiful video is shot and then edited with a sharp eye on the monitor to get a really nice-looking image. Because of time stress, the editor then simply adds a broadcast clamp before rendering for final delivery, to be sure he' can't be accused of shipping video with illegal values.

So?

Well, the beautiful images he saw on his carefully calibrated monitor looked very different at the station. The beautiful super-white highlights aren't super anymore, they now just meld in with the background (or the foreground as the case may be), and the nice superblack shadow details are now crushed to just black across the board.

The proper way to deal with illegal video is as part of the color correction (before and/or after effects processing). This means manual work on a scene basis, but this allows you to make an intelligent choice about what to do with the illegal values if any, based on what subjectively looks best.

This is very different from just crushing the superblacks and the superwhites, which is what the broadcast clamp does.

A curve tweak can allow you to keep detail in hot highlights for example, with a bit of compression instead of the "brickwall limiter" nature of the broadcast clamp.

(Video is more complicated than audio, so the above analogy is not perfect. For example, you may have excessive levels of one color channel even though the overall "IRE" level is legal.)

kplo wrote on 2/1/2009, 10:14 AM
Glenn,
Thanks for the clarification. I generally only use the broadcast colors filter (lenient, no 7.5 setup) as the last step to assure that any pinpoint light sources or supersaturated colors (tomatoes, etc...I do a lot of food shoots on location) are within legal limits after I have tweaked all elements with CC, secondary CC, curves, etc.
So far, the scopes appear to have accurately reflectled black levels (video and superblack) as viewed on my calibratedCRT production monitor.
Fortunately, any "super saturated" colors are clamped enough to look good on-air and stay within spec.
I agree that it's really not a big deal in practice.
Worry mode off.
By the way, your site has a lot of good info...well done.
Thanks again,
Ken
Coursedesign wrote on 2/1/2009, 10:27 AM
Glenn's page on Vegas 8 and 8/32-bit is an outstanding resource.

...accurately reflectled black levels (video and superblack) as viewed on my calibratedCRT production monitor.

What are you using to feed that calibrated monitor? If it is a Firewire 4:1:1 DV25 output, this is not up to broadcast standards, which specify 10-bit 4:2:2, which can look quite different.

Of course not everyone cares about that part of the broadcast spec, and if the customer is happy, you've made your sale, so you're right to turn off the "worry mode."

kplo wrote on 2/1/2009, 10:25 PM
Coursedesign,
My monitor is fed the output from my Vegas timeline over firewire, through my Sony DSR-30 deck over an S-VHS connector to my monitor.
The image played from my 15mbs broadcast master MPEG stream and the flip tests derived from it (all local cable network spots in the L.A. area and elsewhere, are "flipped" or bandwidth reduced to 3.75mbs for insertion into the broadcast stream) are a very close match to the on-air results, indicating that the tools in Vegas are indeed, pretty accurate.
All my spots are up to broadcast standards or they would be rejected in this or any other major market.
Fortunately, I no longer have to submit spots on Beta SP tape, but can upload a high bitrate mpeg file to the cable company's server for encoding.
As an old curmudgeon, I'm more finicky than most, and can assure you that properly shot and processed material originated on DV can look great on-air. Actually, by the time your spot is flipped, you can forget about any differences between 4:1:1 8 bit and 4:2:2 10 bit color. :)
Ken


Coursedesign wrote on 2/1/2009, 11:12 PM
He-he about the many turns footage goes through in broadcast (and far worse by the time it comes out of a cable set top box). Already at the station, it has gotten much worse with digital TV.

I've shot extensively in both 4:1:1 8-bit and 4:2:2 10-bit, but I don't want to deal with NTSC DV for DVD ever again (other than for my unique archival footage). For broadcast at least you won't have the 4:1:1 chroma cut in half...

GlennChan wrote on 2/2/2009, 1:14 AM
1a- 4:1:1 gives you a quarter chroma resolution horizontally. It's possible to reconstruct the chroma with a box filter (can't remember if that's the right term) and can cause chroma to look terrible- try rendering red text to the DV codec. Only NLEs do this (depends on the codec)... you can usually work around it and fix it. The other methods for chroma reconstruction don't give such terrible looking results. If implemented reasonably well and if you aren't dealing with highly saturated colors, then 4:1:1 works reasonably well.

1b- Unfortunately, in practice, there is little adherence to where the chroma is centered. (And apparently the DV standard is really weird in that it calls for Cb and Cr to be centered 2 pixels apart; I've never seen a DV codec do this.) Though nobody really notices this (even in greenscreen).

2- MPEG-2 at 4:2:0 (not the 4:2:2 variant) can get messy if the footage is interlaced. If it is interlaced 4:2:0, then vertical chroma performance is roughly a quarter*.
*You won't run into any issues with boxy-looking chroma, unlike DV. However, you'll have the problem of chroma appearing in a comb-like way.
see the "interlaced chroma problem" at http://www.hometheaterhifi.com/volume_8_2/dvd-benchmark-special-report-chroma-bug-4-2001.html

---
So the bottom line is, 4:1:1 ain't so bad if you are going to 4:2:0 interlaced anyways. Both will roughly cut the effective chroma performance to a quarter. It would just make it consistently bad in both directions. In most cases we don't notice this.

Some other chroma information (it gets a little hardcore...): http://www.glennchan.info/articles/technical/chroma/chroma1.htm
Coursedesign wrote on 2/2/2009, 9:47 AM
Glenn,

I don't quite understand this.

For 4:1:1, "the vertical chroma performance is roughly a quarter" as you state.

Ditto with 4:2:0, "the vertical chroma performance is roughly a quarter."

Where things really fall apart in my experience is going from 4:1:1 to 4:2:0 for DVD.

Looking simplistically, that drops every other chroma pixel, so we get 4:1:0, which has a chroma performance that is roughly an eighth. This means we are now proud owners of 12.5% of the subject chroma information, having tossed 87.5% of the chroma info in the trash.

In practice, chroma smoothing can synthesize some fake chroma pixels to make it looks better, and I'm all for it (in Hollywood that's even done to fake 4:4:4 from 4:2:2 footage, which has already dropped half the chroma info).

But "good" hasn't been my experience in going from NTSC DV to DVD, especially not compared to going from 4:2:2 or 4:2:0 (although PAL 4:2:0 DV doesn't have exactly the same chroma sampling as MPEG-2, the chroma sites are defined differently).

Cliff Etzel wrote on 2/2/2009, 11:33 AM
To me, this discussion has broken down into esoteric technical lingo fest that only confuses and takes away from the job of shooting, editing and delivering projects.

I just read and watched a video piece where NBC is now using Digital Correspondents (read - Solo or backpack video journalists), shooting with SONY V1U's and editing on a laptop for NBC's nightly news broadcast.

This discussion now appearss to be a nitpicking contest around the pure technical aspects of delivering content for broadcast? As far as I'm concerned, ff the V1U can capture content and then be edited on a MacBook Pro for NBC Nightly News national broadcast specifications, I think the topic has run it's course.

Vegas is capable of the very same thing as FCP - if not more IMO.

I'm going to be getting Glenn Chan's DVD on color correction from VASST to get the straight scoop at a user friendly level for working with Vegas Pro. All this discussion did was create confusion for me.

Cliff Etzel - Solo Video Journalist
bluprojekt | solo video journalism blog
Coursedesign wrote on 2/2/2009, 11:52 AM
One of the benefits of acquiring a deeper understanding of something, like say color sampling, is that it allows you to note during the preproduction stage what might be a problem in production or post production.


V1 and other HDV cameras, as well as XDCAM EX1/EX3, etc. shoot 4:2:0, so you lose nothing going to DVD, and we know that the output looks very good in broadcast (within certain limits; you won't confuse the output of a V1 with that of a Sony F35 other than for well lit closeups with good post).

That may be all you need to know for solo journalism, and if you're going to be the one doing the post, I certainly would recommend Vegas over FCP 100 to 1.

For other people, with other needs, the answer may be different. To find the best choice for specific needs, it takes either trial & error or education.
farss wrote on 2/2/2009, 1:05 PM
Very well put, if more people made the effort to grasp the technical aspects of this art form there'd be a lot less angst and blow outs in budgets.
Too often statements are taken at face value and interpolated incorrectly because the underlying technology is not understood. Simple example:

"V1 and other HDV cameras, as well as XDCAM EX1/EX3, etc. shoot 4:2:0, so you lose nothing going to DVD,..."

Which is correct however more than once I've had people think this will also hold true if they capture HDV as DV.

Another example of a major oops. One local had been shooting with a Z1 and decided to upgrade to an EX camera, the images do indeed look better. What wasn't considered was that their shooting style involved a lot of heavy image stabilisation in post. The better looking images from the EX ended up as mush.

For anyone who wants to get a handle the technical stuff and how it impacts the creative side without delving into a lot of technical mumbo jumbo and maths, I would again strongly recommend "The DV Rebels Guide". Stu is very good at explaining things in a way that creative people can understand.

Bob.

Jay Gladwell wrote on 2/2/2009, 1:18 PM

Cliff, check your e-mail.

Cliff Etzel wrote on 2/2/2009, 1:24 PM
Sent a response but it was bounced back as undeliverable
FrigidNDEditing wrote on 2/2/2009, 1:56 PM
Hey Bluprojekt -

I'm a fan of the Color Correction Volume 4 DVD from VASST, but it did come out before all this stuff with Vegas Pro 8's 32bit stuff etc... so it's not going to go over that specifically. It's a great DVD for learning the filters in Vegas Pro and reading the scopes, and some other stuff. I did a review for it a while back, but sadly I can't even remember where it was published :(.

You should be pretty happy with the DVD though, and you'll get to put a face on the name of one of your fellow forum members :).

Dave
Cliff Etzel wrote on 2/2/2009, 2:20 PM
Thanks Dave - Since I've read from others that the 32bit float mode is leaving them wanting, I've stuck with 8 bit for the immediate future unless someone can explain otherwise. I've posted a new forum topic specifically asking about 8 bit versus 32 bit editing.

Cliff Etzel - Solo Video Journalist
bluprojekt | solo video journalism blog
FrigidNDEditing wrote on 2/2/2009, 3:46 PM
no real reason IMO unless you're working with bad keying, sometimes you can get better results using 32 bit in those situations.

Dave
GlennChan wrote on 2/2/2009, 6:35 PM
but it did come out before all this stuff with Vegas Pro 8's 32bit stuff etc... so it's not going to go over that specifically.

That is why there are free articles on my site that goes over that stuff. :)

http://www.glennchan.info/articles/articles.html