VideoHelp Forum
+ Reply to Thread
Page 3 of 4
FirstFirst 1 2 3 4 LastLast
Results 61 to 90 of 119
Thread
  1. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    So basically, what you're saying is that Neat will temporarily convert your clip to YCBCr in order to do it's work, but exports to RGB when it's finished?

    If so, that makes sense, since my exported AVI from V-Dub (after running it through neat) reads as RGB.
    Quote Quote  
  2. Originally Posted by unclescoob View Post
    So basically, what you're saying is that Neat will temporarily convert your clip to YCBCr in order to do it's work, but exports to RGB when it's finished?

    If so, that makes sense, since my exported AVI from V-Dub (after running it through neat) reads as RGB.

    I'm saying the extra conversion isn't occuring (YV12 , I420, or NV12), beyond vdub's RGB conversion

    1) I'm saying YV12 source => Vdub converts to RGB in full processing mode => neat video works in RGB => export RGB

    2) themaster1 is saying YV12 source=> Vdub RGB => Neat video converts to YCbCr => Neat Video exports RGB => export RGB

    If (2) was correct, it would look worse than YV12 source => Vdub RGB => full process but no filter use => export RGB, but it doesn't. If (2) was correct, then using RGB as input source would show deterioration in chroma as it gets converted to YCbCr, but it doesn't

    If (1) was correct it would look the same as YV12 source => Vdub RGB => full process but no filter use => export RGB , and this is what is happening on my mini test
    Quote Quote  
  3. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    [QUOTE=unclescoob;2148662]And as far as this color space stuff is concerned, Sanlyn, on prior posts you have mentioned prepping your clips in Avisynth and then sending them over to Virtualdub for NeatVideo finishing touches. Now those who use Neat know that it converts your video to YCbCr. Here's what the workflow looks like:

    Originally Posted by unclescoob View Post
    #1- Isn't the native video colorspace YV12? Correct me if I'm wrong.
    MPEG and most DV are YV12. I capture from VHS into YUY2 (huffyuv). When working in Avisynth, I don't go back and forth between these (some plugins work in both, some don't). There are a few Avisynth plugins I sometimes use that require YUY2, so I use those first if I'm already in YUY2 -- if I'm not in YUY2 already but have started in YV12, I'd rather avoid an extra conversion and will try to find another way to handle the same problem. Then I move to anything that needs YV12-only. Once I go to VirtualDub/RGB, I stay there until I encode. I don't switch in and out of colorspaces. If I find I have to go back to an earlier colorspace routine, I re-start from scratch - which means I save a lot of work along the way until I get my final clip.

    Originally Posted by unclescoob View Post
    #2 - Sending that to Virtualdub converts it to RGB32.
    Converts to RGB only if you use a VirtualDub or other RGB plugin, or use "Full processing mode". If all I want from VirtualDub is to see the results of Avisynth, then I set "Color Depth" to YV12 (or whatever), "Compression" to huff or lagarith, mode to "Fast recompress", Then "Save as AVI . . .". The fun comes when you forget to reset all the switches, later realize what you did, and have to run the damn thing again .

    Originally Posted by unclescoob View Post
    #3 - Loading NeatVideo then converts you to YCbCr.
    As poisondeathray noted, NeatVideo calls it "YCbCr" by default (which is more like YUY2 4:2:2 -- but in any event it is not YV12). It's a proprietary colorspace and only their hairdresser knows what it really is. Many retail apps do their own thing. NeatVideo also has a special mode they recommend for b&w grayscale video. But it outputs RGB.

    Originally Posted by unclescoob View Post
    #4 - Send it back to Avisynth (for encoding) and you have to re-convert it to YV12
    I don't send it back to Avisynth/YV12 for encoding. If it's in RGB I'm more than likely going to send it to TMPGenc, which like many Adobe and similar apps works in RGB. If I'm going to encode with HCenc or CCE (which I sometimes use), I convert RGB -> YUV, keep it as a separate AVI to check for any conversion problems, then open it with the encoder.

    Originally Posted by unclescoob View Post
    Here's my question: What do you do in Avisynth concerning colorspace conversion before you send it to V-dub for Neat?
    I don't always use NeatVideo or RGB, but when I do it's after I've completed all YUV processing. Then:

    Code:
    AssumeTFF()       # <-or AssumeBFF(), whichever applies.
    
    ConvertToRGB32(matrix="Rec601", interlaced=true)  # <- interlaced
    ConvertToRGB32(matrix="Rec601", interlaced=false) # <- not interlaced
    I read where AssumeTFF/BFF isn't necessary, but I do it anyway -- just in case. The conversion routine cares more about the interlace factor than about BFF/TFF, so I've been told.

    If I've gone thru all this rigmarole to get my MPEG and find that I really needed to do fix something that must be done in YUV, then I go back to my original YUV starting clip, or back to the YUV step where I screwed up, and re-start from there.
    Last edited by sanlyn; 23rd Mar 2014 at 12:43.
    Quote Quote  
  4. Member
    Join Date
    Jul 2009
    Location
    Spain
    Search Comp PM
    Originally Posted by sanlyn View Post
    I read where AssumeTFF/BFF isn't necessary, but I do it anyway -- just in case. The conversion routine cares more about the interlace factor than about BFF/TFF, so I've been told.
    The conversion routines don't give a damn about BFF/TFF (what has the temporal order got to do with it?) - they only need to know if the clip is interlaced or not.

    Just thought I'd clear that up, even if it only goes to illustrate scoob's point:
    Originally Posted by unclescoob View Post
    Because, as usual, people come here with one issue and then everyone with their cutesy little "HO's" (hehe) drift off into the land of confusion. You all start disagreeing with each other on jargon and crap that has nothing to do with the poster's question, and the poster is left with his hands in the air saying "errr....guys? uh....my question?"
    Quote Quote  
  5. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Originally Posted by Gavino View Post
    Just thought I'd clear that up, even if it only goes to illustrate scoob's point:
    Originally Posted by unclescoob View Post
    Because, as usual, people come here with one issue and then everyone with their cutesy little "HO's" (hehe) drift off into the land of confusion. You all start disagreeing with each other on jargon and crap that has nothing to do with the poster's question, and the poster is left with his hands in the air saying "errr....guys? uh....my question?"
    I'm with you, Gavino. Sometimes I'm overly cautious -- especially when I feel I'm not so sure what I'm doing.

    unclescoob, the jargon does stumble now and then. Thankfully the likes of Gavino, jagabo, et al, are around to catch me making a boo-boo (they always do that!). But I'd rather be corrected than merrily keep making a fool of myself .
    Last edited by sanlyn; 23rd Mar 2014 at 12:44.
    Quote Quote  
  6. Member 2Bdecided's Avatar
    Join Date
    Nov 2007
    Location
    United Kingdom
    Search Comp PM
    Originally Posted by unclescoob View Post
    I've tried it and the results look great. But it's indistinguishable from HcEnc, and I'm sure that TMPEGenc's results are no different. The difference? Mainconcept is a transcoder, so it can convert your files to other formats (Blu-Ray included). HcEnc and TMP only do MPEG-2.

    So what exactly does the uh..."real stuff" do otherwise? Just curious.
    I don't own any pay solutions. But I've seen videos that I just can't encode without blocking, encoded in a pristine manner using things I can't afford.

    If you haven't seen any differences, then you're probably not pushing the encoders that hard.

    HD home movies, downconverted to pin-sharp interlaced SD, are very difficult to encode. 25Mbps DV is fine, but 8-9Mbps MPEG-2 can struggle. Depends on how much movement, noise and detail there is.

    Cheers,
    David.
    Quote Quote  
  7. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    Sanlyn, why do you self-deprecate so much?
    Quote Quote  
  8. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    In the process of learning this crazy process, I've made some real boo-boo's ! No doubt, I'll make more.
    Last edited by sanlyn; 23rd Mar 2014 at 12:44.
    Quote Quote  
  9. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    Ok folks, here it is as promised. Sorry it took a while. Attached are two dark scene clip samples from my previous example earlier in this thread: The unfiltered clip and the filtered clip.

    Notice how dirty the unfiltered clip is. The filtered clip looks alot better, however there is still some garbage floating around that I cannot get rid of for the life of me. Here is the script I used:

    Code:
    Import("C:\Program Files\AviSynth 2.5\plugins\FixChromaBleeding.avs")
    Import("C:\Program Files\AviSynth 2.5\plugins\TemporalDegrain.avs")
    MPEG2Source("yaddayadda.d2v", cpu=3)
    TFM("yaddayadda.d2v", cpu=3")
    TDecimate(mode=1)
    Tweak(hue=0.0, sat=0.93)
    FixChromaBleeding()
    TemporalDegrain()
    TTempSmooth(maxr=5, lthresh=4, cthresh=3,interlaced=false)
    TTempSmooth(maxr=5, lthresh=4, cthresh=3,interlaced=false)
    I was thinking of adding one more TTempSmooth BEFORE temporal degrain, and leave the other two stacked at the end of my script as-is. But I don't know. Anyway, any advice is appreciated
    Image Attached Files
    Quote Quote  
  10. The problem is that the noise has the same amplitude as some of the detail. If you use enough noise filtering to remove the noise you remove the detail too.
    Quote Quote  
  11. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    So considering that fact, how would you rate my filtered result?

    By the way, is that just regular analog VHS noise, or does it look like a bad transfer to you? Or bad encoding, perhaps?
    Quote Quote  
  12. Originally Posted by unclescoob View Post
    So considering that fact, how would you rate my filtered result?
    It depends on whether you'd like to have less noise and loss of detail, or more noise and retain more detail. MCTD at high (plus enhance=true) removes most the noise (enhance=true adds a little grain to reduce posterization artifacts) but removes some of the details -- the black lines in the blue blanket, for example.
    Quote Quote  
  13. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Interesting source. I've seen worse, but it has several problems. My PC is busy with another project right now, but will try a few things later this morning.
    Last edited by sanlyn; 23rd Mar 2014 at 12:44.
    Quote Quote  
  14. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    thanks Sanlyn.

    Jagabo, I wouldn't mind keeping the details and leaving some grain (which is "healthy" for video, anyway). Am I at least somewhat on track with my filtered result?

    MCTemporalDenoise is good, but I notice it causes TONS of banding in dark areas (even at low settings with enhance=true, and deband activated)
    Quote Quote  
  15. Originally Posted by unclescoob View Post
    MCTemporalDenoise is good, but I notice it causes TONS of banding in dark areas (even at low settings with enhance=true, and deband activated)
    I wasn't seeing much of that.
    Image Attached Files
    Quote Quote  
  16. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    I'll check out your sample when I get home from work (not allowed to here).

    Thanks.

    Sanlyn, still waiting on ya, buddy.

    Thanks.
    Quote Quote  
  17. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Pretty good work, jagabo. I had some slight banding from TemporalDegrain in a couple of scenes. Fixed that, but theer are level and edge ghost issues (the edge ghosts are in the source, not in the plugins). Haven't figured out the edge issue yet, the only really nice ghost filter I've found is in TMPGEnc Plus. Working on it . . .
    Last edited by sanlyn; 23rd Mar 2014 at 12:44.
    Quote Quote  
  18. I didn't bother with anything but IVTC and MCTD. And I didn't fine tune anything beyond:

    McTemporalDenoise(settings="high", stabilize=true, deblock=false, sharp=false, enhance=true)

    As I mentioned earlier, the black lines in the dark blanket have disappeared.

    The source (with gamma stretch to make the blanket more visible):
    Click image for larger version

Name:	src+.jpg
Views:	100
Size:	84.8 KB
ID:	11729

    After MCTD (also with gamma stretch):
    Click image for larger version

Name:	mctd+.jpg
Views:	146
Size:	52.1 KB
ID:	11730
    Last edited by jagabo; 2nd Apr 2012 at 13:02.
    Quote Quote  
  19. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Got the banding licked, finally. But that green edge in the source.....What a problem! I'll run off a copy in a bit and see how it flies.
    Last edited by sanlyn; 23rd Mar 2014 at 12:45.
    Quote Quote  
  20. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    How do you think my results look?
    Quote Quote  
  21. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    Wait, Sanlyn...green edge on the source?? Where?? You're not talking about his shoulder are you?
    Quote Quote  
  22. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    I'm talking about the green big-mouthed blob. I forget what it's called.
    Your results: okay, but one problem: TemporalDegrain didn't "cause" the artifact pattern you referred to. It just didn't clean it. Wasn't designed for that; it's not grain.

    Running a final AVI now. You'll love this; the combo of MCTD and NeatVideo gave the best results. Slow as hell. Back in a just a bit.
    Last edited by sanlyn; 23rd Mar 2014 at 12:45.
    Quote Quote  
  23. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    Originally Posted by sanlyn View Post
    the combo of MCTD and NeatVideo gave the best results
    Hope it kept the details at least. I haven't played Jagabo's sample yet, but that screencap makes me nervous. Most details are gone!
    Quote Quote  
  24. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    jagabo's looks pretty good to me. Can't go by the screencap, I think it's "enhanced" to show some midtone (gamma) effects.

    I'm trying to get up two versions, one with NeatVideo, one without. Gotta stop now and get dinner ready for the wife. Back later.
    Last edited by sanlyn; 23rd Mar 2014 at 12:45.
    Quote Quote  
  25. Yes, the images I posted were gamma stretched to bring out dark details. Here's the same frame of the source at the original brightness:
    Click image for larger version

Name:	org.png
Views:	96
Size:	567.1 KB
ID:	11738
    Quote Quote  
  26. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Running MCTD, I get pretty much the same thing jagabo did. This video looks the same, no matter what you run, within reason. The longer you work on it, the more all vfersions just look the same. Where did this video come fom, anyway? It's in worse shape than the toon you submitted some time back. Was this recorded to VHS? To DVD on a recorder? The grunge looks like compression artifacts - which means you have to either live with some grain, or live with some banding effects (the grain actually covers some banding in the source. Cut all the grain, you'll see it). You can smooth the heck out of it, but there won't be much left if you do.
    Last edited by sanlyn; 23rd Mar 2014 at 12:45.
    Quote Quote  
  27. The noise looks to me like an off air recording from a weak signal. And something else:

    The reason this video is particularly difficult is that the vertical line noise doesn't vary from frame to frame. So temporal filters won't get rid of it. Sixteen frames averaged together:

    Click image for larger version

Name:	avg16.png
Views:	108
Size:	501.8 KB
ID:	11742

    You need a strong spacial filter to get rid of the vertical lines. But that means you will lose details with similar amplitude.
    Last edited by jagabo; 2nd Apr 2012 at 20:32.
    Quote Quote  
  28. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Odd, the more noise you remove the more you see the damage, however it happened. Not the first time I've seen this effect, though. I'll post some results later tonight, if just to demonstrate that various smoothers and cleaners get the same results with this particular clip.
    Last edited by sanlyn; 23rd Mar 2014 at 12:45.
    Quote Quote  
  29. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    Believe it or not, to answer your question on the source: The source is from the official, retail DVD set from the Time-Life series. Sad how studios don't bother with decent transfers. I own tapes of these episodes (copies from a guy who taped them off cable tv). I'm almost willing to work with those and I'm starting to think I'll have better looking results! Is there any way deals can be worked out with these studios to obtain TAPE copies of their masters instead of transfering to DVD?
    Quote Quote  
  30. Banned
    Join Date
    Dec 2010
    Location
    New York
    Search Comp PM
    What is a good spatial filter for this type of stuff, btw?
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!