I am in the process of trying to transfer some of my videotapes to DVDs, and was wondering what filter(s) I should use to remove the 'halo' effect from videos? In other words, after I remove the 'noise' from my videos, I notice what looks like a sort of 'halo' effect in areas where there is supposed to be a gradual difference in color shading/brightness, or whenever a bright colored object is near, or right next to, a dark colored object. I had asked this before but got no responses.
Any practical suggestions will be greatly appreciated. Thanks in advance.
Hi it's me again. I think the only way would be to cut the borders with the "crop" filter. Please post a picture, but this will be very difficult to achieve I'm afraid.
Is this from capturing analog video?
Attached is a picture so you can see what I mean, and yes, this is from capturing analog video.
It would be good to have a picture BEFORE noise reduction, too.
Just thinking that this looks a bit like chromashift.
Try capturing with virtualdub and use a plugin filter to correct for the chromashift.
Or maybe use avisynth to see if the already captured video can be corrected with a chromashift plugin filter.
Have a look here for some info that might help.
http://www.doom9.org/index.html?/capture/chroma_artefacts.html
Quote from: zakk on January 26, 2014, 04:10:21 AM
It would be good to have a picture BEFORE noise reduction, too.
Quote from: AQUAR on January 26, 2014, 11:12:44 AM
Just thinking that this looks a bit like chromashift.
Try capturing with virtualdub and use a plugin filter to correct for the chromashift.
Or maybe use avisynth to see if the already captured video can be corrected with a chromashift plugin filter.
Have a look here for some info that might help.
http://www.doom9.org/index.html?/capture/chroma_artefacts.html
I tried usinng chromashift, and the effect is still the same (I don't think that chromashift explains the "halo" effect anyway).
The only thing I see are the stripes on the left suit collar of the guy. Is that what you are talking about?
I don't see any halo's either, just some colour bleeding that looked a bit like chromashift.
Its a bit harder to get at the problem when there is some difficulty presenting/understanding the issue.
Since its analog video its probably interlaced.
So if we are talking about the stripes on the edges:
Presumably the guy is moving and they are some combing artifacts from a simple (weave!) deinterlacer.
Try a better deinterlacer like YADIF or store as interlaced and let the media player do the hard work..
Apologies for the completely-tangential derailment, but: Is that Tom Baker?!? Color me inappropriately curious about the source of those sample images. ;D (Heck, some backstory might generate more attention to your problem, Gandalf!)
More usefully, though, I'm still trying to understand the nature of the "halo" complaint, exactly. I see two things of immediate interest, comparing the "before" and "after" images provided:
- It's clear from the second (before) image that the input video is interlaced ââ,¬â€ see the "serration" at the edges of the shirt collar, ear, nose, and hairline. It's even more clear that it's being deinterlaced either poorly or not at all. Proper deinterlacing should be able to produce an output video without those alternating-scanline artifacts, but trying to postprocess them away by treating them as "noise" will have much less satisfying results (as you've discovered). I'd recommend experimenting with all of the deinterlacing options available to you, to see if you can find a good fit. Interlacing has extremely predictable effects on the video image. It can usually be mitigated quite effectively without degrading overall quality, and without having to resort to heavy-handed general noise-reduction for "cleanup". Combating it that way tends to just smear the distortions around, instead of repairing them.
- The other possible source I see for your "halo" complaint, is some slight blocking and lack of gradient smoothness in the fine details of the image. For instance, that area of brightness from light reflecting off the back wall (second quarter of the image, in the background between whatever-that-is on the far left and the foreground subject on the right), which appears as a fairly uniform light-to-dark gradient in the "before", but in the "after" it's been smoothed to more of a solid-color gray stripe, with a harder edge where it transitions to the darker background. The noise-reduction being applied could very well be the source of that effect. Or, it could be a sign of overly aggressive compression in the output encoder. Perhaps the output bitrate/quality is set too low, as that can often manifest as blocking and loss of quality in detail areas.
Yes, it's Tom Baker. I'm trying to transfer a Doctor Who video to DVD (so I don't wear out the videotape). Anyway, he "halo" effect is more noticeable when the video is n playback mode (after I have done the editing) but how do I go about editing it so that I don't get that "halo" effect?
Maybe in playback mode the de-interlacing is by blending.
This is diffusing the edges where there is motion and might look a bit like halo's.
I support FeRD_NYC in that this seems to be a capture of interlaced video (as opposed to film!).
Best to avoid the 'halo's' by properly processing the analog video.
Use something like virtualdub with a fast/lossless codec as per my suggestion in your audio thread.
Apply a good deinterlacer, maybe try YADIF or the bob deinterlacer with debobbing to avoid jitter.
Then compress/remux with your favorite codec/container using avidemux.
I tried hat, it doesn't do a thing to remove the halo effect. But yes, it's a capture of a videotape, as I clearly stated before (a previously unreleased episode of Doctor who called "Shada"), so how do I go about removing what I call a "halo" effect?
We just don't see the halo effect you are trying to describe.
Hence we have made some educated guesses that might lead to what you are trying to point out.
Maybe we should have asked first if this halo is present on the analog video.
Can you be more specific in what you have tried.
- did you compress with a at high bit rate.
- did you capture using a fast/lossless codec.
- is the halo present on the raw captured video.
- is your pc fast enough to keep up with the capture.
- what deinterlacing methods have you used.
- has anything you tried made any impact on this halo.
The spatial quality of the video capture is poor, probably because it is like that on the analog video tape.
You cannot change that with filters, and some deinterlacing methods will in fact further degrade the vertical resolution.
Only thing I can suggest is to post a sample of this video capture and clearly point out where you see this halo.
Quote from: AQUAR on January 26, 2014, 11:12:44 AM
Just thinking that this looks a bit like chromashift.
Try capturing with virtualdub and use a plugin filter to correct for the chromashift.
Or maybe use avisynth to see if the already captured video can be corrected with a chromashift plugin filter.
Have a look here for some info that might help.
http://www.doom9.org/index.html?/capture/chroma_artefacts.html
After taking the time to actually read the link that you provided (unfortunately my brain is sometimes slow on the "uptake" ) I owe you an apology (I have come to the conclusion that, yes, the issue is, in fact, chromashift) and I am including two sceenshots of filters (showing their default values),the first is of the Cnr2 filter, and the second is of the YUV filter. My question is What settings would you recommend for these filters?
@ Gandalf
Glad you are making some headway in solving this halo issue.
Chromashift is not really subject to a best filter setting.
The type of analog signal presented to the capture device, the quality of that capture device to separate out the video components, the characteristic impendance of the cable etc, all have a causal impact on issues like chromashift.
Typically, low end capture device's have a S-Video input, ie not the best.
Just play with the settings for trial and error correction.
Or visit the avisynth forum and see if you can talk to some of the "older" video buffs that know how to deal with such analog video peculiarities.
Try someone like Donald Graft with has much expertise in analog and digital media.
I don't know if this would be of any help or not. but my VCR uses the RED-YELLOW-WHITE cable hookup (just so you know).
That means composite video plus L & R audio.
All of the video signal components (luminance, chroma, sync) are combined so that they can be carried in a single channel (one wire!).
The decoding/transmission of this video signal isn't perfect and so the various components interfere with each other. Degradation like chroma shift is one problem caused by trying to send everything over one wire.
Initially a broadcaster performs RF modulation on composite video for over the air broadcasting (antenna input!).
Not that good quality and subject to multipath ghosting.
Then you play back the recording by pushing it all back into a composite signal (not so good).
Next you digitise that with a capture device, but good ones are very expensive.
Consumer grade ones do work okayish but don't try to compress the ouput on the fly (hence use fast - lossless codec).
Next best thing to composite video is S-video, a little round multipin connector.
Here Chroma and Luminance are send on their own channel.
Best of all is component video where 3 wires are used to minimise interference between the video bits.
Thats what you are dealing with.
Thank you.
Quote from: Gandalf on February 01, 2014, 05:06:24 AM
I don't know if this would be of any help or not. but my VCR uses the RED-YELLOW-WHITE cable hookup (just so you know).
Actually, what I was asking about was the Cnr2 filter, NOT the Chromashift filter.
We actually still don't know what problem you are trying to deal with.
Only thing we know is that you observe some 'halo effect'.
From our questions we deduced that you are doing some basic capturing of analog video over a composite connection.
From your pictures we see that its interlaced video.
The quality of your original analog interlaced video, video playback device, capture device, video connection are all at basic consumer level.
Hence, chroma artifacts, deinterlacing artifacts, dropped frames etc are quite normal for that kind of work flow.
Halo = my guess remains with chromashift
Other chroma artifact = very likely.
Cnr2 filter is a chroma noise reduction filter for rainbow and colour "blotching" artifacts.
Are you seeing these?
Anyway, the same answer applies for setting the paramaters of Cnr2 as it is for any chroma artifact filter.
Its a process of trial and error around a baseline set of figures.
Baseline figures for all these filters are usually found with the filter descriptions.
Just note that all these filter processes are not going to change the basic character of your video.
They just ameliorate various static issues and some are great at giving perceptual visual improvements.
With trialling, what looks best to you is best for you.
I happened to find some information on the internet that I think will be of help to others trying to restore analog video tapes at the following link:
http://www.digitalfaq.com/guides/video/introduction-restore-video.htm
Article is a good read with a nice glossary.
Basically backs up the items we touched on in this thread.
It's not bad. The author gets a bit caught up in gear wankery, IMHO, and anyone who endorses Monster cable by name is presumed shady. I recommend looking past that stuff or taking the recommendations with a big handful of salt. Which isn't to say that the basic advice of "you need decent equipment to get good video captures" is wrong, because it's not.
In terms of the software advice, though, it gives some good helpful hints. It could use some more illustrations, especially since not all of the terminology is entirely accurate (as is basically admitted at the top of the section) and some of it is downright wrong.
Most egregiously, the issue described as "Top-Screen Tearing" is not in any way an example of tearing, which is actually a hard-edged display issue caused by sync issues with the screen refresh rate. (Wikipedia has a decent-enough introduction to the phenomenon of Screen tearing (https://en.wikipedia.org/wiki/Screen_tearing).) The distortion illustrated in the article is commonly referred to as "bending", or even more precisely "flagging".
But I agree with AQUAR, there's some useful advice in there, and overall it's a worthwhile read.
@ FeRD_NYC
Anyone who endorses Monster cable by name is not presumed shady BUT presumed greedy!
Audiophiles will argue that they can hear an improvement when using Monster cable.
Since all cables have resistance, a characteristic impedance and some conductor configuration, then different cables do (in a pure sense) cause a difference in signal received. Issues like attenuation, reflection at mismatched connection points and skin effect are real.
How real and what impact depends on what is being transmitted.
IMHO, for applications targetted by Monster cables, the ratio of Price increase / Performance change = Astronomical.
Not good, as I like getting the biggest bang for the buck.
Money would be much better diverted into improving critical hardware components.
Regarding the article - I've seen much worse and still felt these had merit for educating the reader.
Sometimes little white lies are good. Big lies never (like monster cables!).
This forum needs a few more members to help the casual avidemux user - you seem like a good candidate.
Quote from: AQUAR on February 13, 2014, 10:12:09 AM
Anyone who endorses Monster cable by name is not presumed shady BUT presumed greedy!
Audiophiles will argue that they can hear an improvement when using Monster cable.
Since all cables have resistance, a characteristic impedance and some conductor configuration, then different cables do (in a pure sense) cause a difference in signal received. Issues like attenuation, reflection at mismatched connection points and skin effect are real.
How real and what impact depends on what is being transmitted.
Ugh, well,don't get me started on "
audiophiles" ââ,¬â€ I've seen knock-down, drag-out flamewars with Those Typesâ„¢ who've sworn up, down, and sideways that they can hear a difference based on the quality (==$PRICE) of the
DIGITAL cables used! The notion of "crisper 1's and sharper 0's" invariably sails right over their heads.
There was even one guy, i
swear this is the absolute truth, who claimed that it for anyone who
really cares about sound and picture quality, it's absolutely imperative that they buy a multi-$100 (US) gold-plated, antimatter-shielded, double-dutch-woven, genuine unicorn intestine HDMI cable.
Especially important with HDMI, you see, because if you use an inferior interconnect,
the audio and video will "interfere" with each other being carried over the same cable! :o
(As a hint for any novices out there: (a) That is one of the stupidest statements anyone could ever make about a signal like HDMI's digital stream, for reasons that hopefully were instantly obvious but otherwise you should make an effort to learn, as it's important to your understanding of how digital works; and (b) Empirical testing has shown that the
only time the construction of an HDMI interconnect can have
any bearing at all on the signal quality is when the cable length is
greater than twenty-five feet (8 meters), and even then it's highly unlikely until the lengths go beyond
fifty feet (15 meters). Any shorter than that, and the requirements of the HDMI standard alone are high enough to ensure that
any cable built to meet them will give you the same performance.)
...Yet Monster's built a massive empire by preying on exactly that kind of stupid. My favorite take on Monster is still this Penny Arcade strip (http://www.penny-arcade.com/comic/2002/11/25). (Which is about Monster Power, specifically, but IMHO applies to Monster Cable as well.) Long story short, they're awful... just awful.
Quote from: AQUAR on February 13, 2014, 10:12:09 AM
IMHO, for applications targetted by Monster cables, the ratio of Price increase / Performance change = Astronomical.
Not good, as I like getting the biggest bang for the buck.
Money would be much better diverted into improving critical hardware components.
That's exactly it. In the analog world,
can the interconnects contribute to signal degradation, and cause a worsened result? Absolutely, it's
possible! No question. But generally, the only time that's going to make any
practical difference is when you're starting with a signal that's so pristine and pure as to be "better than" the cables, coming from top-of-the-line equipment. For the average consumer, as long as they're not using the absolute worst, cheapest, came-free-in-the-box cables they can find, they'll be more than "good enough".
Far more benefit would come from spending money on upgrading the
equipment, rather than buying expensive cables that ultimately won't make a damn bit of difference to their setup.
Quote from: AQUAR on February 13, 2014, 10:12:09 AM
Regarding the article - I've seen much worse and still felt these had merit for educating the reader.
I absolutely agree! I only mentioned the tearing thing because I think it's important to at least
try to get the terminology right, especially for less experienced folks. It greatly facilitates discussing any issues they may be having, or asking for help on forums like this one, if everyone's on the same page with what they're referring to when they use certain words. Something I think the author of that guide agrees with, even, since the "Glossary" section opens with a note asking for terminology corrections. If I were a more conscientious reader, I'd have forwarded along my tearing/flagging note as requested. (I may still, in fact.)
Quote from: AQUAR on February 13, 2014, 10:12:09 AM
This forum needs a few more members to help the casual avidemux user - you seem like a good candidate.
Oh, I think y'all do fairly well. I've never asked a question or brought up a problem here and
not gotten a satisfactory answer, which is a pretty good track record! But I'm happy to help out where I can, in sharing my own experiences and insights.
The biggest hurdle for me is that I'm a purely casual user of avidemux, and sometimes weeks or months will go by where I just don't find myself using it. So, I'll undoubtedly vanish from time to time. But when I'm around, I try to give a little back. :)
I didn't even know "monster cables" were available for the digital world.
Better correct my equation - ratio of Price increase / Performance change = Infinity.
I wonder how these people reconcile their beliefs when dealing with wireless transmissions!
There are some very clever people here that from time to time help out with good advice.
Most of them are casual as well so, you would fit in nicely.
Oh, you have
No. Idea! In fact, I wish I could say I learned about that sort of suff reading this piece, but the truth is I'd encountered such beasties previously:
Apparently This Matters: $1,300 HDMI cables (http://www.cnn.com/2012/06/01/tech/social-media/bellini-hdmi-cables/)
...I think the title really says it all. :D
That discussion in question, the one I mentioned with the guy swearing that audio and video would "interfere" with each other in HDMI, occurred in the comments there. Specifically, it's in this comment thread (http://www.cnn.com/2012/06/01/tech/social-media/bellini-hdmi-cables/#comment-545811128), and "siranthony" is the genius in question. (It gets a bit confusing to follow, because Disqus is sorting newest-first, so you have to read the thread "upside-down" after the initial comment.)
But it doesn't stop there! One classic from a few years back, until it got dropped by the manufacturer after everybody and their mother realized it was a total scam, was Denon's Premium "Denon Link" Cable (http://denon'sPremium"DenonLink"Cable). Here's how they described it in the product lit:
QuoteDenon's 1.5 meter (59 in.) proprietary ultra premium Denon Link cable was designed for the audio enthusiast. Made from high purity copper wire and high performance connection parts, the AK-DL1 will bring out all the nuances in digital audio reproduction from any of our Denon DVD players with the Denon Link feature connected to a Denon Link enabled Denon A/V receiver. The AK-DL1 employs high level tin-bearing alloy shielding not typically available in commercial cabling, to eliminate data loss caused by noise. Additionally, signal directional markings are provided for optimum signal transfer. Attention to detail when building this cable was used by employing high quality insulation and woven jacketing to reduce vibration and to add durability.
Here's the thing. Sweep aside all that bullshit, and turns out
it's just a goddamn ethernet cable! A
short ethernet cable! And they really were charging $500US for it! (http://news.cnet.com/8301-17938_105-9967991-1.html) >:(
It's sad as hell when even reputable companies like Denon will succumb to the temptation to separate fools from their money. Fortunately,
those companies can usually be shamed into cleaning up their act. Unlike Monster and their ilk... there's just no hope for them.
so I take it that once it says [solved] in the title, topic is free to go about whatever ?
One thing to silence most audiophiles though, is to ask them if they ever cared to have their hearing metered. The very most never bothered.
Given that a large portion of them audiophiles are my age (47) or older, ââ,¬Â¦
well, anyway, thanks for the laugh with the 1300$-hdmi-cable. 8)
good night ââ,¬Â¦
@ snork - solved topics are not free to go on about whatever.
The banter about monster cables is a derivative of this topic.
And it has lightened the usual activity, of members that participate in this forum, just a little bit.
It made you laugh, and maybe we saved some money for those easily persuaded by technical sounding nonsense.
But if really deemed inappropriate in some way, its fine for the administrators to remove the "monster cable" trailing end of this thread.