Are Vegas's color bars wrong?

GlennChan wrote on 8/14/2005, 11:46 AM
If you go to Adam Wilt's website, he has a set of correct color bars.
http://www.adamwilt.com/ColorBars.html

I've compared the bars to Vegas's, and the colors in the bottom left bars is significantly different (the bars on both sides of the 100% white bar).

This may not be good for broadcasters, as they may view the bars on a vectorscope and see that those bars don't line up.

On Monday I may be able to check Vegas' bars against a vectorscope.



2- If you render out the Vegas color bars, the colors will shift. So don't do that.
If you just drop the bars onto the timeline and view it on a broadcast monitor, that will be right.
Try using the cookie cutter and superimposing vegas' bars over bars.avi from Adam WIlt's site. Mute the bars on top. There's no change.

If you're making pre-made headers with 2-pops and everything, watch out for this issue.

Comments

Jay Gladwell wrote on 8/14/2005, 2:13 PM

You can't judge from those as they were compressed for the web. Besides, computer monitors are not NTSC compatable.

Plus, if you're having problems rendering the bars in Vegas, then you've got a problem with either Vegas and/or your monitor.


GlennChan wrote on 8/14/2005, 4:45 PM
You can't judge from those as they were compressed for the web.
It's full quality DV. It works because it's just one frame.

Besides, computer monitors are not NTSC compatable.
I have a broadcast monitor but it doesn't have much to do with this. You're looking more for a vectorscope.
B_JM wrote on 8/14/2005, 6:08 PM
ive encoded for DVD some anaglyph 3d from vegas - and it didnt work ..
fine in preview on external broadcast monitor , but played back on any dvd player on same monitor - the colors were just enough off that the 3D effect was gone .

this is not only a Vegas issue - but almost any mpeg encoder issue relating to RGB to 4:2:2 YUV conversions -- plus the idiocyncracies of the encoder
Jay Gladwell wrote on 8/14/2005, 6:13 PM

The bottom line is, and the truth is I can only speak for myself and my equipment, since I haven't and can't see yours, is that I'm not experiencing any problems with the Vegas color bars.

So to answer your question, No, there is nothing wrong with the color bars as far as I can tell based on the results I'm getting (yours and Mr. Wilt's may vary).





Grazie wrote on 8/14/2005, 10:19 PM
Put up the Bars. Open your Scopes and see where the colours lie. The last time I did this the colours were spot on in the brackets within the scopes.

Does that mean the scopes are wrong too?

Just repeated the experiment with "Vectorscope" scope. All the colours lie slap in the middle of each of the brackets for: White; Red; Magenta; Blue; Cyan; Green.

Now, if we are "only" talking about the "look" of the colours, well ok. I've got a LCD screen that I really can't satisfactorily "set-up", I use my pro JVC monitor for finalising colours. Once I saw the colour bars sit correctly within the scopes I then re-checked my JVC monitor - I didn't "need" to do anything.

Just repeated this process - still, ok.

Have I missed something here?

Grazie
B_JM wrote on 8/15/2005, 3:31 AM
i think he was saying that he noticed them different only after rendering ... not before ..

StormMarc wrote on 8/15/2005, 3:47 PM
What version of Vegas are you using? Sony admitted that they were wrong in older versions. I think it was fixed in V5.

Marc
GlennChan wrote on 8/15/2005, 6:54 PM
Ok I am talking about 2 different things here, which is probably confusing.

1- The Vegas bars change slightly (neglibly?) if you render them out. This is not the wrong part I'm talking about, although if you're assinine this is kind of wrong.

2- The -I and Q bars (something like that) are significantly off. The +I and Q bars are the two bars on the left and right of the 100% white bar in the lower left.

3- I'm talking about how the color bars show up on an external analog vectorscope. It will have more markings that the Vegas vectorscope, like the ones for -I and Q.

http://www.tek.com/Measurement/App_Notes/NTSC_Video_Msmt/vectorscope.html

4- Using Vegas 6.
GlennChan wrote on 8/16/2005, 9:44 AM
I just tried looking at the bars on a Videotek VTM-420HD/SD, and the Vegas bar's -I and Q bars don't line up. The device takes SDI input so I am not that familiar with how to use it but I'm fairly sure those -I and Q bars don't line up.

If anyone has an analog vectorscope and a deck that does 7.5IRE setup/pedestal they can double check my results.
ForumAdmin wrote on 8/16/2005, 11:44 AM
If you think the bars are in error, please tell us the entire chain of technology used for your test, including codecs and all hardware.
GlennChan wrote on 8/17/2005, 7:38 AM
System specs: in my username.
Print to tape via firewire.
Unchecked use MS DV codec, checked ignore 3rd party codecs
Panasonic NV-GS70 (Japanese) is my home camera.

At work:
Played the tape back on a DSR-1800 DVCAM deck
SDI out to a digibeta deck (sorry, forget model #)
SDI out from digibeta deck to the Videotek scope mentioned earlier. (I don't believe the SDI signal loops through.)

Also on the tape, I put the bars.avi color bars from Adam Wilt's website, generated using Bob Currier's Test Pattern Maker and encoded with the Matrox DV codec.

2- Both these color bars also differ from FCP's color bars, where the -I and Q bars are 'significantly' different.
B_JM wrote on 8/17/2005, 8:24 AM
the same thing has been observed in FCP and premiere also ... plus you are going through several YUV - RGB conversions ...

I also note that the color samples posted on various sites are in PNG format - which ive shown doesnt translate colors correctly anyway ..

I just read this site which mentions much of the same issues - a lot more is found in the adobe and fcp forums, as well as matrox having their own issues ..
http://www.bealecorner.com/trv900/edit.html

It has also been brought up many times at Canopus ..... their various codecs have had issues ... QT DV has changed their setup from v4 to v6 to add to the confusion , plus you can use either rgb or yuv mode which will give you completely different results ...