|
|
| trevlac |
| Posted: Feb 23 2004, 03:37 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
Hi,
I've been trying to add a vectorscope to a waveform monitor and histogram vdub filter I made.
Trouble is, I'm not 100% sure on how to position the color targets on the scale. This is my method so far:
The Horizontal and Vertical are U and V. If I convert 0-255 RGB to YUV, it appears the vectorscope radius is a max of about 118/119. I assume the color targets are 75% saturation of the signal. So that would be 255*0.75=191. I can run 191 color bars. But then it looks like I have to flip (*-1) the V. If I do that, the targets "look" to be in the right spots, but I'm not sure. Also, I'm not sure how big to make the target boxes.
Anyone know what I'm talking about? Any suggestion on how to calculate the scale on a vectorscope?
Thanks |
 |
| Stefan |
| Posted: Feb 24 2004, 12:32 PM |
 |
|
Advanced Member
  
Group: Vdubmod Alpha Testing Team
Posts: 215
Member No.: 142
Joined: 27-August 02

|
This might be useful for you:
HOW TO USE A WAVEFORM & A VECTORSCOPE & what is an illegal level http://www.execulink.com/~impact/scopes.htm
Testing Applications in Uncompressed HDTV Signals (PDF) http://www.synthesysresearch.com/pdf/testing.pdf
Bye Stefan
|
 |
| trevlac |
| Posted: Feb 24 2004, 07:12 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
@Stefan,
Thanks for the links. The testing pdf looks especially good.
I found my problem. I was using YCbCr calculations to get R-Y and B-Y. It looks like I should have been using YUV calculations. U runs from +/-112 and V from +/-157. This seems to match what colorbars would give on a scope.
If you know about vectorscopes, could you take a look at my plugin? It will take me a day or 2 to finish up. I'll edit this post when I have it working.
For now, I do have Histograms and a Waveform monitor in the filter.
http://trevlac.us/colorCorrection/colorTools.html |
 |
| TCmullet |
| Posted: Feb 25 2004, 01:11 AM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
@Trevlac,
I'm drooling and chompin' at the bit to use your WFM filter! I've been agonizingly desiring a WFM equivalent for 2 years. Nobody in digital video land (whether Vdub or Dazzle DVCII where I come from) seems to know what one is. Yet, 10 years ago, as I was in the midst of a stretch in my life where I had invested a great deal in analog video gear (which never became a success), I had gotten hooked on using a WFM all or most of the time.
I strongly wish to download your WFM filter, even if the other components are non-operational. Can you provide a link for download??
Oops... I reviewed your link again, and found the "plugin" link for download. Will give it a shot! |
 |
| TCmullet |
| Posted: Feb 25 2004, 01:41 AM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
@Trevlac,
It's looking great! I uncheck all the colors leaving just luma, and I feel like I've got a WFM in the house again!
It would be real helpful if you could add 2 horizontal lines, 1 for 100 IRE and one for 7.5 IRE. Perhaps making them dashed lines would be good. Maybe a separate switch to enable them would be good. The one for 7.5 IRE is especially critical, but both are important.
I'm going to use this for 2 things, right off the bat. The UPN network is sending to my affiliate a signal with a slight ghost. My inside source told me that the station and the network have gone through spells of each blaming the other for the ghost, with no resolution. This area includes (but is not limited to) the Orlando and Cocoa, FL markets. I use the Exorcist filter to attempt to deghost the image (of Star Trek Enterprise). But it has the artifact of lowering the brightness. Now with this WFM, I may be able to compensate more precisely with a brightness/contrast filter.
2nd thing: Lately, my UPN affiliate seems to be running their black/white levels out of whack (picture seems darker than other stations). It could be the cable company. In any case, my captures are too dark, and I can't trust my eyes. So your WFM tool will allow me to LOOK at the 1-volt-peak-to-peak signal graphicly like we're supposed to as good videographers, then fix it w/a proc-amp filter. (Btw, I probably won't ever mess w/the colors. Only luminance seems to be the biggest issue. At least it was 10 years ago, and seems to be now as well. I never was able to afford a vectorscope, but if black & white levels are correct, then I can eyeball the color hue and saturation.)
Thank you again! |
 |
| trevlac |
| Posted: Feb 25 2004, 02:04 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
@TCmullet
Thanks for the kind words. I found I could not trust my eyes, so I started putting together these tools so I could follow 'instrument flight rules' instead of 'visual flight rules'. Actually, I'm not sure how people adjust by eye.
Anyway, there are 2 things you should know about the WFM. 1st is that it does not really draw a trace, like an analog device. It just draws the dots. So colorbars will not be a 'full' picture, but a set of lines. It seems to me that the trace is more of a distraction than something that provides info. 2nd, the 0-255 vs 16-235 input option, changes the luma display. I've stuck with a 0-255 scale across the screens. The 16-235 input option will give you the fuller luma range that matches the RGB range, but this can get confusing.
On the IRE7.5 and IRE100. I will gladly add more detail to the scale grid. But why do you want IRE 7.5? As far as I understand, digital has no setup. So black is 16 for YCbCr regardless if you are talking PAL or NTSC or J-NTSC.
So, if you mark your input as 0-255, then on the WFM luma display, black will be 16 (same as NTSC analog 7.5IRE) and white will be 235 (same as NTSC analog 100IRE).
I was thinking about adding dotted lines between the current solid lines. That would break the 256 scale in 16 point sections.
What do you think? I really have no idea what commercial PC based scopes do. If you (or anyone) thinks other scales or displays would be better, please let me know.
|
 |
| TCmullet |
| Posted: Feb 25 2004, 08:43 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
| QUOTE | | ..it does not really draw a trace, like an analog device. It just draws the dots. So colorbars will not be a 'full' picture, but a set of lines. | Really? So what? I have a video file captured from my old Video Toaster's smpte color bars. I loaded it just now. Looks just fine to me. Stairsteps are square, everything is almost as I expected. When a true WFM paints the line across from left to right, it is indeed a solid line, but so what? Now it's a slightly dotted line; I'd expect that now that we're in the digital domain. So what if the adjacent bars don't have a vertical "line" connecting their horizontal platforms. The important thing is the platforms and their being at the correct level. The only quirk I see different is in the black bar area where the middle one is subdivided into the 3 sections, below true black, at true black, and above true black. It looks like you can't represent anything below true black at all--it's simply truncated on the bottom. But I guess I should have expected this. At least it's nice to see the latter 2 showing correctly.
| QUOTE | | Actually, I'm not sure how people adjust by eye. | I think it's very error prone, and I think a lot of people don't really know how imperfect their adjustments really are. They think they've made it look "better", but w/no WFM, the goodness of their results will be limited to their own viewing environments. Therefore there's a lot of bad video results floating around out there. Last year, I downloaded a few episodes of Star Trek Enterprise off of Kazaa (of different formats: mpeg1, divx, etc.), and even w/o a WFM, I could tell the black/white levels were screwed up royally on many of them.
| QUOTE | | ..the 0-255 vs 16-235 input option, changes the luma display. I've stuck with a 0-255 scale across the screens. The 16-235 input option will give you the fuller luma range that matches the RGB range, but this can get confusing. | I'm sticking w/the 0-255. I see what doing the other does, and 0-255 (which makes the displayed data limited to 16-235, seems the most logical. I'm only barely starting to understand what any of this means. I think in terms of IRE.
| QUOTE | | But why do you want IRE 7.5? As far as I understand, digital has no setup. So black is 16 for YCbCr regardless if you are talking PAL or NTSC or J-NTSC. So, if you mark your input as 0-255, then on the WFM luma display, black will be 16 (same as NTSC analog 7.5IRE) and white will be 235 (same as NTSC analog 100IRE). | I want 7.5 as that's where the correct black level is supposed to be. What if my captured video has black set too high? I need to lower it with a brightness filter, and I lower it to the 7.5 mark. But I guess you're saying that 7.5IRE equals 16RGB; so be it. "Digital has no setup?" What does THAT mean? As I got into analog video in '92, I learned that the 4 common procamp controls of brightness, contrast, tint (or hue), and color (or saturation), are more accurately called (the 1st 2 anyway), black level (or setup) and white level. They called black level "setup" because you start your "set up" process by "setting up" the black level first. Then once that's done, you adjust the white level appropriately. This would never have made sense to me until I had a working old beat up tube-based Tektrontic WFM to watch as I adjusted the procamp on a TBC feeding it. Then I could see that changing the black level (brightness) shifts the ENTIRE video signal range upward and downward, but changing the white level (contrast) only STRETCHES the range bigger and smaller. So it makes sense that you do black first, then white. Of course to the neophyte, "black" is a misnomer as it doesn't adjust only the blacks. It adjusts the whole range, but we call it "black level" because we adjust this control to make the darkest part sit at the correct bottom of the scale as our reference for everything else. Then we do "everything else", which is simply adjusting the contrast (white level) to make the brightest parts come just shy of the top. Which I guess you'll tell me is 235 in RGB terms, or 100IRE in analog terms. So why you say there's "no setup in digital", I still don't know. We still should "setup" the black level first, then adjust white level to get an appropriate spread between darkest and brightest. At least I plan to keep doing that (or maybe I should say, "resume" doing it after years of not having a WFM).
Oh, also, if "digital has no setup", then why are there Vdub filters that affect the "setup" level (brightness)? But after all this discussion so far, I think we can safely let that remain a rhetorical question. It's obvious that if there's a way to adjust brightness and contrast, then there needs to be a way to monitor those adjustments with a tool such as yours and not just with our subjective eyeballs.
| QUOTE | | Actually, I'm not sure how people adjust by eye. (quoted again.) |
As expensive as a WFM was, a vectorscope was way further out of my economic reach. But I was advised that if I set black/white levels correctly (via my WFM), then color hue and saturation could reasonably be adjusted by eyeball. Once I got used to what my monitor looked like with a good, well-adjusted video sequence, then it's not too hard to watch out for those bleeding reds and overall comfortable color levels. And for hue, I could carefully look at flesh tones, and/or use the VT (Video Toaster 2) to split the screen with color bars and fine adjust the hue coming in from taped bars that way.
It was interesting to note that if you adjust color before black/white levels, you can really screw things up. I noticed that you could "correct" wrong color level by boosting the black level (or was it white level?). I do remember that a grossly low white level will make the colors look too intense. If I lowered the color, then raised the white, the color would greatly disappear. (This is just an example.) So I learned to defer all color issues until after black & white levels were properly adjusted. Then eyeballing the color matters wasn't too hard. Of course I used the same NTSC monitor for all testing, with its proc amp controls set to their detents. (A possible exception might be brightness, but I'd still leave it unchanged most of the time.)
| QUOTE | | On the IRE7.5 and IRE100. I will gladly add more detail to the scale grid. I was thinking about adding dotted lines between the current solid lines. That would break the 256 scale in 16 point sections. | On the WFMs I've seen, 7.5 and 100 should be the only dotted ones. There may have been others, but I know that any dotted ones had specific function; not simply that they were every-other grid line. Definitely make 7.5 & 100 dotted or dashed in some way--we need to be able to see most of the signal as it approaches the limits. A solid line could be a hinderance. But I don't think having dotted lines scattered through the midst would be of any help. Actually, now that you've divulged that the 7.5IRE black is equal to 16 RGB, that does help as 16 is half of the 32 that you already have, and one can eyeball-divide that somewhat. But it'd still be helpful to have a dotted line just to make it that much easier and quicker to use. Same at the top end (100IRE or 235 RGB).
| QUOTE | | I really have no idea what commercial PC based scopes do. | Me neither (no money). I only know what I remember when this old beatup boat anchor worked years ago, and how much I've missed it.
| QUOTE | | What do you think? (about scale displays) | I've told you all my thoughts. But I would like to suggest some other enhancements, if I may. 
Add a preview mode. This may sound superfluous, but there's another idea that would require it; add a slider (and a switch) that would allow us to monitor the waveform of a single line anywhere in a given frame. I want to be able to find a frame, then slide this slider up and down the frame watching the WFM as I go. Then I might shift to some other frame in the file and do the same survey of various lines on that 2nd single frame. My old WFM had this single line feature, and it was darn handy at times (rarely, but I definitely did use it).
I realize that what you've done is replaced all the video with our WFM chart. So obviously, the filter has to be deleted once all proc-amp tweaking is done and before Vdub "save" is started. Having a preview mode (showing the WFM chart) would allow you to invent the slider so that one could monitor the single lines as described.
Now one more thing... I don't know if this is possible--can you also show the video image itself (in a separate window) at the same time as the WFM chart? The preview window in Vdub has the frame-number-changing slider at the bottom to let us move to any frame in the file. But it would be hard to find the frame we want, if we can't see the image as well.
If you can't do this, then maybe a feasible alternative would be to do what some PC-based WFMs have done. (I know I said I'm not familiar with them--true, but I have seen pictures of one in action a few years back.) That is, show the video image in the same window as the WFM chart, that is overlay the chart on top of the video image. It's not a perfect solution, but it does have a certain elegance about it--only one window to mess with and my eyes don't have to switch back and forth between the image window and the WFM window. Perhaps this could be an option controlled by a checkbox switch.
One more thing-- could you somewhere somehow save the last used settings of your entire filter? I would like not to have to pick WFM and deselect red, green and blue channels every time I want to "turn on" my WFM. Where do other Vdub filters save any of their settings from invocation to invocation? The registry?
Well, my fingers are tired, but fortunately I've said everything I think I wanted to say.
I'm hoping you can release a version with those 2 dotted lines as quickly as possible (and before implementing anything else). That's my wish-- I hope you don't mind me expressing it. |
 |
| trevlac |
| Posted: Feb 26 2004, 02:39 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
| QUOTE | | QUOTE | | ..it does not really draw a trace, like an analog device. It just draws the dots. So colorbars will not be a 'full' picture, but a set of lines. | Really? So what? |
That's what I was thinking. More of a limit of the analog device.
| QUOTE | | The only quirk I see different is in the black bar area where the middle one is subdivided into the 3 sections, below true black, at true black, and above true black. It looks like you can't represent anything below true black at all--it's simply truncated on the bottom. |
The blacker than black thing is an issue of capture and conversion to PC RGB. When you capture in say YUY2, the card (and driver) will map 7.5IRE to 16 (for YCbCr). The card may allow for signals below 7.5 to be mapped in the range of 1-15 (zero is reserved). IF that works, there is another problem with keeping it. When most codecs convert from YUY2 (YCbCr 4:2:2) to RGB, they expand the B/W range from 16-235 to 0-255. This is appropriate for PC viewing, but it truncates anything out of the IRE7.5 to IRE100 range. Good reason to adjust a signal before you capture (if you can).
| QUOTE | | "Digital has no setup?" What does THAT mean? | Poor choice of words on my part. I have only done Digital video. I'm still quite a novice. I mean, In digital RGB or YCbCr (which are the 2 color spaces used) Luma = 16 is black and Luma = 235 is white. This is regardless of the analog source (PAL , NTSC, J-NTSC) range 7.5IRE - 100IRE or 0IRE - 100IRE. So 16 luma is black.
| QUOTE | | Then I could see that changing the black level (brightness) shifts the ENTIRE video signal range upward and downward, but changing the white level (contrast) only STRETCHES the range bigger and smaller. |
This is a big reason for the tools I put together. So I (and others) could see what those damb controls do. 
Your suggestions ..
| QUOTE | | Add a preview mode. |
I'm not clear on all of your suggestions. Is preview mode an xtra window used when you are setting the filter options? I really just learned how to program this stuff. I'd have to figure out how that xtra window works. Is the single line option only in preview mode?
I believe you can save/reload filter setting in VDub, using some options under the file menu. My filter should work this way.
I thought displaying all of the charts over the actual video was interesting. I assume the area that the chart covers is made a bit darker, because a bright picture comming thru would make it hard to read the chart. Only value in preview ? Maybe on regular output display ? ..................... I have made a few changes, so here is my working copy http://trevlac.us/colorCorrection/clrtools_1_3b.vdf
I added the 16 and 235 dotted lines. You can remove the other grid lines and keep just those. Also added a passthru option that just passes the video thru, but still saves the other options (don't need to delete the filter to do final save).
I played with the chart over the video thing, but there are a few problems. Seeing the chart thru bright video, and speed. This is not implemented.
The Vectorscope is there and working ok. I have not created the grid and scale. A bit of Trig in determining the points to plot.
After I get the VScope ok, I will look at the chart on video thing. Then I'd have to learn about preview windows to do the other stuff.
Trev |
 |
| TCmullet |
| Posted: Feb 27 2004, 07:44 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
| QUOTE | Your suggestions ..| QUOTE | | Add a preview mode. | I'm not clear on all of your suggestions. Is preview mode an xtra window used when you are setting the filter options? |
Many if not most of Vdub filters have a button something like "show preview". When you press it, a window gets thrown on the screen like another little Vdub, only there's no play button-- only the other controls, like single-frame, slide-to-anywhere-in-the-file, next-key-frame, etc. The image area shows the video image as processed by all the filters, including the one you are currently setting up. This window allows you to tweak whatever filter you're setting up and see the results immediately on whatever frames you want. I was suggesting that you add this preview window, and make the WFM show on it only. However, now that you have a bypass option, the preview window is less important from that standpoint.
However, ...
| QUOTE | | Is the single line option only in preview mode? |
Yes, I'm suggesting that as the other reason to have a preview window. In your original version, you have to exit the filter-parameter-window AND exit the filter-add-window, before you can observe the WFM output on the main Vdub window. At that point, none of the WFM features can be changed.
Waveform monitoring different single lines in a particular frame is by definition something that would have to be done while your filter's window is open. And if it's open, we'd have to have a preview window to see anything of WFM output. (Vdub's window is behind it all, and locked out.) So while the WFM filter is open and it's preview window is open, I'd slide to my frame in question (let's say frame 2,341). Then when I'm at frame 2,341 I'd switch to single line mode. I'd scroll to look at each of lines 0, 1, and 2 (in turn) to see how the closed-captioning looks in waveform (just for my edification). Then I'd scroll down to some other line in the visible area and look at that line. I'd scroll around to many different lines and observe the waveform change to match each line that the slider is "pointing" to at the moment. In normal mode, we are seeing 480 superimposed lines at any one time. In single-line mode, we are looking at a single line of video, therefore there should be a graph of a single continuous function line; the graph of the voltage of whatever line we're looking at (between .075 and 1.00 volts peak-to-peak -- 7.5 to 100IRE). Of course, your filter would need to display that line number somewhere so we can know what line we are monitoring.
Does this clarify it enough for you?
| QUOTE | | I believe you can save/reload filter setting in VDub, using some options under the file menu. My filter should work this way. | No, I'm asking for something additional. When I add WFM the first time to a Vdub session, I have to pick "WFM", and deselect red, green, blue. Then let's say that later on I delete the WFM filter. Then if and when I add it again, I still have to pick WFM and deselect the colors. I'm asking you to save those options so that the next time I add the filter, it will default to the last options I had set in a previous "adding" of the filter.
Your bypass switch somewhat allevates the need for this, but not entirely.
| QUOTE | | I thought displaying all of the charts over the actual video was interesting. I assume the area that the chart covers is made a bit darker, because a bright picture comming thru would make it hard to read the chart. Only value in preview ? Maybe on regular output display ? | I hadn't thought of it, but perhaps maybe the picture should be darkened a bit. If you have a preview window at all, then I think your 2 questions go away. It'd be "yes, both". Chart on video might be too difficult to make look acceptable. One could always invoke 2 instances of Vdub, a bit of a hassle but workable. What we REALLY need is a separate piece of WFM software that somehow ties into the current Vdub filtered output chain, but has it's own chart window independent of Vdub.
I think that chart-on-video should be lower priority than preview window or single line selectability.
(A little off point.) I don't understand what use there is for WFM windows for each of the colors. The waveform of a video signal is a luminance issue, and not relevant to color. But maybe it becomes relevant in the digital world. (I'm still thinking NTSC analog.) Any thoughts on why you have R,G,B WFM windows? |
 |
| TCmullet |
| Posted: Feb 27 2004, 09:16 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
Hey, I just thought of something else a true WFM ought to have. (It's been too many years.) If the video is interlaced, we should be able to select even, odd, or both fields for showing. A set of 3 radio buttons would be appropriate, with "both" being the default. (Unless of course, it's been set to something else in a previous invocation.) |
 |
| trevlac |
| Posted: Mar 2 2004, 02:58 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
@TCM
I haven't had a chance to wrap this up. I will try to finish the Vector scope tonight.
I appreciate all of your suggestions. I does seem that the preview (or better yet, a seperate window while in processing mode) would be best.
I'm not much of a programmer. I'll have to learn a bit to even do the preview. Actually, I'm not sure that is useful since you have to switch to a different filter to actually make adjustments. So, we've got a poor mans scopes for now. 
| QUOTE | I don't understand what use there is for WFM windows for each of the colors. The waveform of a video signal is a luminance issue, and not relevant to color. But maybe it becomes relevant in the digital world. (I'm still thinking NTSC analog.) Any thoughts on why you have R,G,B WFM windows? |
Well, RGB is really the form of the original source (from the camera) and is also the form of the data in VDub. As you set the black level, you may notice that you can crush one of the color channels past zero(such as green) as you move the luma black down. I believe you can also get an idea if a frame has a certain color cast to it. The vector scope would be best for this, but a histo and WFM (combo) will tell you if you have a lot of bright red (compared to G&B) and where the red is in the frame. You can also use a R wfm to see if something is wrong with your camera (or source).
Here is a link to an Avid guide on color correction. It's a worth while read about digital color. I wish I found it earlier. Their DV package is $695. They have a free one, but it does not have the color correction tools.
ftp://KCAvid:KC951348@support02.avid.com/...d/AvidCC_UG.pdf |
 |
| TCmullet |
| Posted: Mar 9 2004, 03:53 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
@trevlac: Have you made a new version since 1.3b? |
 |
| trevlac |
| Posted: Mar 9 2004, 09:23 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
Yes. The link on my site gives you 1.3 (no beta).
http://trevlac.us/colorCorrection/clrtools.vdf
I simply completed the vectorscope (some optimizations and the scale).
I think the next thing is to firgure out how to get the workflow to work better. I've been working on a batch of old stuff that I've had sitting around. I'm trying to get a feel myself.
I have noticed that if the filters that I use to make the adjustments have a preview, I can adjust to the scope readings reasonably well. This is the real preview that is needed, because the scope doesn't really adjust anything.
I also see at least one bug. I do not repaint the scale on the WFM between every frame (because it really does not change). However, depending on the other filters I use, this area sometimes gets messed up.
----------------------------- On a side note:
I find it interesting that few people seem to comment on the use of scopes. This is going to be a big part of my capture process.
|
 |
| TCmullet |
| Posted: Mar 11 2004, 03:25 AM |
 |
|
Advanced Member
  
Group: Members
Posts: 312
Member No.: 3970
Joined: 2-May 03

|
| QUOTE | | I have noticed that if the filters that I use to make the adjustments have a preview, I can adjust to the scope readings reasonably well. This is the real preview that is needed, because the scope doesn't really adjust anything. | Sounds like you're giving up on the preview option. But if you do, then you can't implement the ability to scroll up & down the image displaying the waveform of any single line. Or switch back & forth between 2 fields. I'm hoping you'll still do these things. Just yesterday, I had a need for the single line monitoring.
| QUOTE | | I also see at least one bug. I do not repaint the scale on the WFM between every frame (because it really does not change). However, depending on the other filters I use, this area sometimes gets messed up. | I don't see the scale get messed up, but I do see strange garbled stuff in the background. Fwiw.
| QUOTE | | I find it interesting that few people seem to comment on the use of scopes. This is going to be a big part of my capture process. | Absolutely. In fact, I've been thinking for days now, that I need to do the community a service and make a thread just to get on the soapbox to explain why this is important for everybody. I guess my dabbling in analog a decade ago was worth something after all. (Will write it when I get some time soon hopefully. Hate to keep this great tool to ourselves. It needs to be the last filter in every set of Vdub filters! (Passthru checked of course) Thanks for a great job so far. Will try out that vectorscope (not that it will do me much good--as I've said, WFM is the big thing). |
 |
| trevlac |
| Posted: Mar 11 2004, 04:10 PM |
 |
|
Advanced Member
  
Group: Members
Posts: 176
Member No.: 8501
Joined: 26-January 04

|
| QUOTE | | Sounds like you're giving up on the preview option. But if you do, then you can't implement the ability to scroll up & down the image displaying the waveform of any single line. Or switch back & forth between 2 fields. I'm hoping you'll still do these things. |
I've not given up yet. It's just that I'm not really some wizbang programmer. I'm looking for the easy way. I am experimenting with my own work flow to get to know the problems. I think I am going to try to adjust things so you can call the filter from avisynth. My idea (you mentioned this also) is to run 2 copies of VDub. 1 for the WFM and 1 for the video. A common avs script might be a good way to tie the windows together. Down side is you'd have to use avs filters to change the levels.
| QUOTE | | I don't see the scale get messed up, but I do see strange garbled stuff in the background. Fwiw. | That's what I'm talking about.
| QUOTE | | Will try out that vectorscope ..... |
Here are some things to do with the vectorscope.
1) Remove color cast: Find a picture with a nice amount of white. Look at the vscope. The white should be a blob in the middle. If it pushes out toward a given color, use grafts hue/sat/intensity filter to lower the saturation of the offending channels.
2) Adjust skin tones: This is a bit more tricky because I'm not sure a PC monitor is a good guide to what is correct. Find close up head shot. The skin tones should be a little above the line that is at about 11 o'clock. Adjust the hue of the red and yellow. Maybe even the saturation. You'd have to preview to see the effect, but this a way to tweek skin tones.
3) Make things more vivid: For something like cartoons, this is nice. Adjust the saturation up for the colors you want to effect. Make sure they do not go past the target boxes. Yellow and cyan should be lower than the others, so it is good to look for these colors. You can check with the 'hot pixel' option to see if you've really gone too far.
Trev |
 |
|