VCR/Hi8 capture tests; help evaluate?

[Brad]

Not understanding "correcting the bottom of the image". I have my +tbc & -tbc clips that I made from the test tape, so if I know what to look for...

Take a look at the bottom of the videos. All of them except the Sony+TRV480 have distortion in the bottom 9 lines that obscures any semblance of the picture info contained there. The Sony+TRV480 only has a couple bad lines and then it's back to normal video (but with total garbage for color data).

Just wondering if I should grab one of my crts to use as a monitor on the capture boxes rather than using an lcd like I am now for proc amp tweaks on the next capture go-around.

Only tweak brightness/contrast for capture and you won't need to change your monitor; histograms will provide you with more accurate information than your eyes even if your monitor was $10,000 - because your room, etc. affects the way your visual system perceives the video.

For post-capture color correction you may want to consider it. Even then, I don't see a reason to if you'll be viewing them only on LCD HDTVs and projectors. But I don't know what kinda monitor you have. Mine is a nice IPS panel so if yours is a cheap TN... maybe?

So far, looks like no definitive vcr winner, right? By the same token, no definitive loser either.

Yeah.

[sanlyn]

Been reading another thread, VHS Restoration, that has a ton of worthwhile stuff. Think I'll share with hubby. But in there, there's a mention of monitor calibration. Just wondering if I should grab one of my crts to use as a monitor on the capture boxes rather than using an lcd like I am now for proc amp tweaks on the next capture go-around.

By and large, CRT's are no more "accurate" than LCD's as far as color balance goes. Both require calibration. There's no such thing as display accuracy out-of-the-box. This has to do with manufacturing tolerances, among other things. Because a monitor or graphics adapter is "digital" doesn't mean that it accurately displays your original graphic as-is. If OEMs stopped each monitor in the production line to precisely adjust display accuracy, each monitor would skyrocket in price.

Your statement about digging up the old CRT and the statement about one of your videos looking "faded" prompted me to look up a few references. The video you referred to isn't faded: it has black levels set too high. Also consider that analog-to-DV loses some color info, and you have yet another subtle influence at work.

You'd be surprised to learn that consumer monitors (and many pro monitors, for that matter) have factory settings with little to do with accuracy. Fact is, most consumers don't want accuracy. What they want is bright and shiny pictures. The trend today has it that if red is saturated beyond the roasting point and brightness is set high enough to illuminate a back yard BBQ, the monitor looks "Great!". Trouble is, graphics set up to look "Great" on a "great" monitor will look screwy on a more accurate display. Years ago with my first PC I used a CRT monitor that was actually rather dim. I used Photoshop and brightened the pics to make them look "normal". Later I acquired a more accurate display; those pictures are so bright on it they look washed-out, almost colorless.

The term "color accuracy" has been defined, technically. Details of that definition aside (a mountain of math that we can safely ignore, for we'll never use it), a picture is said to be "accurately" displayed when certain parameters are measured as being within 2% of what they're supposed to be -- that is, the similarity between what is being sent to the screen and what is actually displayed. At 3% or 4% error, the keen eye of a photog or videographer would sense that something isn't quite right, but it's close enough to be acceptable. By 5% to 6% error, folks who are color sensitive (there are plenty of them around) will sense that something is amiss. By 8% to 10%, all but the typical gamer would say that something definitely looks "off". Sadly, the typical digital monitor has display errors between 8% and 20%, depending on which part of the spectrum is being displayed. Brights are usually on the edge of exploding and shadow detail is mucky garbage. Still, the average joe thinks the monitor looks "great", even if Aunt Jennie looks like Mr. Hyde in drag wearing a blue fright wig.

Video and movies are normally created with a color temperature of 6500 Kelvin in mind. That average 6500K comes from the average color temperature at noon in Washington D.C. on a sunny day, with reference to the way humans perceive "white". Out of the box, most consumer monitors are set up for anywhere from 9000K to 12000K. The higher "K" you have, the brighter it looks in your sunlit living room -- and the more bluish it is, such that blacks, grays, and whites never look really black, gray or white. Some makers also set a higher green brightness because we perceive green as "brighter" than colors of equal hue and brightness; ergo, a buyer chooses that monitor because its screen is "brighter". Contrast ratio: someone buys a monitor because it has a 30,000,000:1 contrast ratio, i.e,, the luminance difference between its darkest display point and its brightest. Fat chance. try staring directly at the sun and you'll begin to see what such a contrast ratio would entail. In practice, a contrast ratio of 850:1 is closer to what a decent monitor getd, and it's perfectly adequate for graphics work. Those millions-to-one figures are pure smoke and mirrors.

IPS monitors have a reputation for better accuracy out of the box. This doesn't mean they're all that great, it just means they're less whacky that the typical TN panel. While some monitors are so designed that they defy accurate adjustment, most consumer displays are capable of excellent performance when adjusted properly. That adjustment is very difficult to get by fiddling with image controls and eyeballing the results. The human eye might know that something is off, but it doesn't know why. The image controls on most monitors are limited (does your monitor have clock and phase adjustments? Mine do). Getting the "right color" is only part of the process; there are also gamma, luminance, and saturation issues. Your eyes say "there's too much red". Does that mean you need more cyan (blue + green), or do you need less red? There's a difference.

I reviewed a website article I hadn't seen in a while. In doing so, I found that I was in error about my concept of gamma, so I stand corrected from re-reading it. The page isn't too long and it's written in layman's language. It has some pics, charts, graphs, and no math (thank goodness), and references to some calibration devices and software (some of it free). One of the freebies is a series of test panels at the lagom website (http://www.lagom.nl/lcd-test/). A handy little tool that is most illuminating (pardon the pun). The website that explains what "calibration" entails is here: http://www.tftcentral.co.uk/articles/calibrating.htm .

My monitors and TVs are calibrated using a colorimeter (it measures colors and stuff on your monitor's screen) and two pieces of software used to interpret the results and help make adjustments. The XRite software is used to calibrate PC's, the HCFR software is used to measure and display the results graphically. A couple years ago I posted the HCFR pictures of what a monitor calibration can do. I accomplished the calibration by starting with free test panels from the lagom website. That got me "in the neigjborhood" of more decent performance. But the later colorimeter + PC software nailed things precisely. The post, description and pictures are in the forum here: https://forum.videohelp.com/threads/335402-VHS-capture-critique?p=2083260&viewfull=1#post2083260.

[dianedebuda]

Take a look at the bottom of the videos.

Oh, I thought you were meaning a specific area of the pic.

In the VHS Restoration thread, the OP was having trouble setting his proc amp controls, so the monitor/calibration was mentioned. Didn't think about situation that he may have been doing it by eye instead of histogram. That's what prompted the monitor question.


So the vcr situation is a coin-toss. And the 550 is better than the AIW. How about the 550 vs the trv480? With the 550 (or another capture card), I can have some control with proc amp and I can't with the DV. Using the 550, I can get a lossless avi and DV is lossy. Guess the question then is how difficult does it make it for post processing with DV? With the types of defects noted in post 25, for example, how many can be fixed without extreme effort (if even fixable)? How do you link to a post like 25 here without the full url? Oh, and can you identify the sources of the two stacked pictures? Found the defects on my captures, but not sure which ones you were talking about. Significance of checked ones?

Better Capture card... In this case I think what you're looking for is one that could possibly exceed the dynamic range provided by the 550.

your JVC S-VHS is blowing out your capture cards' levels while the plain-jane Sony is providing more headroom

Does this mean the 550 is just not handling all of the vcr's signal? I thought "cards" were better than usb devices where there was a choice. LordSmurf has listed the 600 usb as an alternative if agp not available. I'm just a bit confused here.

[dianedebuda]

Read your post, sanlyn, after sending mine. Enjoyed it immensely. All I was worried about was proc amp brighness/contrast settings using a histogram. The OP in the VHS Restoration thread was having problems - maybe he was trying to do it by eye. Hubby's projector is calibrated using a fancy calibration disk. My 65" hdtv is basic, "normal" preset with sharpness turned way down 'cause I use it mostly as a monitor for spreadsheets and trimming tv captures. You would choke if you saw the crt on one of my capture boxes - something has died giving it a purplish cast and it has a burn-in ghost. Only use it once a day to schedule recordings, so as long as it's readable, not inspired to lift the sucker and replace it.

Your definitions of brightness, contrast and saturation in the VHS Restoration thread gave me an "ah ha" moment. No wonder I've always had trouble figuring out how to adjust monitors.

[Brad]

Grab the link to the post from the post # in the top-right, highlight the text you want to turn into a link, and click the icon above the post editor that looks like a globe with a chainlink on top of it.

I don't know how to fix any of the issues that I circled; the idea is to avoid them of course. The sources are in the filenames of the blow-up images (mouseover them). The checks are noting that the TRV480 seems to somehow be correcting the halos there (or those halos are being introduced by sharpening performed by the 550).

Certain people are dead-set against USB devices. I would not recommend the ATI 600 in this case, as its contrast adjustment isn't useful.

[dianedebuda]

Edited 34 to try out post linking.

I don't know how to fix any of the issues

That's fine. It's good to know what's reasonably fixable and what's the "charm" of vhs. Guess I didn't hover long enough over the pics for the pop-up ids. On the checked spots. I still see the same problems in the trv480 as the 550, just a little fainter. Maybe that's what you're meaning. Later in the clip, I see some fairly substantial color bleed (hopefully correct term) of girl's red dress onto man's pants. Another mostly uncorrectable error?

Certain people are dead-set against USB devices.

Suspect may be an Apple vs. PC, Windows vs. Linux -type soapbox for many. I'm of the camp: if it works for you... Thanks for feedback on 600 for my situation.

Better Capture card... In this case I think what you're looking for is one that could possibly exceed the dynamic range provided by the 550... your JVC S-VHS is blowing out your capture cards' levels while the plain-jane Sony is providing more headroom

Still wish you can explain a bit what you mean by this. Possibly that the Sony signal highs & lows are not clipped off by the 550; the JVC has a larger signal range and the highs and lows are being clipped by the 550? Hubby wondered if there is a gain control on the JVC so as not to overpower the 550. I looked in the manual & didn't see anything. Hmmm, wish orsetto was here to ask.

[Brad]

There are things you can do to mitigate color bleed, but ultimately VHS simply has much lower chroma bandwidth than luma.

I think you explained what I meant. No way is there a gain control. Turning on/off some of the processing features like TBC may have an effect on it but I don't think so.

[jagabo]

You can see the low chroma resolution of VHS here:

https://forum.videohelp.com/threads/319420-Who-uses-a-DVD-recorder-as-a-line-TBC-and-wh...=1#post1980652
https://forum.videohelp.com/threads/319420-Who-uses-a-DVD-recorder-as-a-line-TBC-and-wh...=1#post1981589

[dianedebuda]

Round 2a: Basically same hardware as post 12 + new VC500.

Status:
vcr selected: JVC HR-S7500U
capture devices eliminated: AIW, TRV480 pass-through for vhs tapes
(TRV480) DV downgraded to 2nd tier because lossy.

Purpose:
Select capture device: 550, VC500 Captures look pretty close to us. See same halos & color bloom. VC500 advantage: proc amp setting have greater value range, could use on Win7 (tested on XP); VC500 disadvantage: no YUY2 - does this matter for post-capture processing?.

Select vcr settings: TBC +on/-off, R3 +on/-off, svhs always on (think that only applies to new recordings). Is it better to do sharpening with vcr's R3 or post-capture processing?

Select hi8 TRV480 settings: TBC +on/-off, DNR +on/-off.

all 550 clips used default proc amp settings.
VC500 proc amp settings = default or Bright=7100 + Contrast=3860
Hi8 clips seem to have a lot of color bloom around blue pants & red shirts. Cure available or is that just the way it is?

Relevant clips from Round 2: (post 12)
hi8 TapPractice -tbc, 550 vdub clip1.avi +dnr
vhs guitar jvcs svideo +tbc+r3+svhs, 550 vdub clip1.avi

Hopefully not too much here. I realize some errors are not fixable; what I'm unsure about are those where a tape player's "cure" fixes one thing but creates other problems. Not knowing what can be done post-capture makes it tough.

If we go with the VC500, looks like we'll match oldretiredguy for the vhs tapes

access the controls from Video -> Levels (hotkey: L)

Like jagabo, proc amp controls not available this way for me with the 550; ok with the VC500.

Looking around, seeing other newbies almost as paranoid about tape damage as I am

Edit: skip 5th file: vhs guitar jvcs svideo +tbc+r3+svhs, aiw vdub clip1.avi; uploaded in error

[Brad]

The VC500 uses UYVY instead of YUY2 (aka YUYV). The only difference is the order that the data is stored internally, but codec developers have to specifically author their software to support the different ordering. Huffyuv internally reorders UYVY to YUY2 so that downstream software doesn't have to worry about supporting it.

(v1.2.2) This version adds UYVY compression support. Since UYVY is just YUY2 with a different sample ordering, I simply reorder the bytes during compression, effectively converting the image to YUY2 on the fly. Decompression to UYVY is not supported, but decompression to YUY2 will work. Huffyuv will also now function as a UYVY codec if you don't have one installed on your system already.

Not sure regarding R3 edge enhancement; perhaps someone else can advise. You're right that the S-VHS On/Off function is irrelevant.

I'll look at the samples tomorrow I guess.

[dianedebuda]

I kind of thought that the UYVY -> YUY2 was happening from what I saw on oldretiredguy's thread, but it was just suspected at one point. A little later, you'all were changing him from huffyuv to an avisynth for reasons I didn't understand, so I kind of got lost there. Think there's supposed to be info in avisynth that lists this, but haven't got avisynth loaded yet.

I should have been saying color bleed instead of color bloom. Bloom kind of makes it sound like a good thing.

On the vdub proc amp controls: is there a way to move from field to field, e.g. brightness to contrast, with the keyboard? I would have guessed Tab, but it doesn't work. Everytime I touch the slider to select the control field, it jumps a bit. Right/left arrows fine for adjustments once there. Also, is there a hidden reset?

I'll look at the samples tomorrow I guess.

Am grateful for any time you care to spend. Helps so much. I thought about adding samples with proc amp adjustments for the 550, but thought it'd be too much. Hopefully the default 550 vs the default VC500 is enough since I didn't really see a lot of difference. But I may be missing things more experienced eyes would see.

Actually fairly pleased with the tapes so far. The vhs guitar was shot hand held in really bad lighting and tight quarters tethered to a vcr & a/c outlet. I think hubby did a reasonably good job there. And I shot the hi8 tap practice without much regard to lights, etc. Surprised it doesn't jerk all over the place since it also was hand held. It wasn't meant to be kept, but I'm glad I have it now. Got a few frames of my daughter practicing at home that magically appeared when I changed the TRV480 from auto to hi8 format. Wish I had that. Don't think the analog -> digital went too badly. How much these can or should be improved, dunno - suspect not much. Once I finish with the capture setup, I can turn my attention more to what's available as post processing and the logical order of fixes.

[sanlyn]

Why is there so much aliasing in all these captures?

[Brad]

I should have been saying color bleed instead of color bloom. Bloom kind of makes it sound like a good thing.

Bleed describes colors extending beyond the objects' boundaries, while by bloom I thought you meant oversaturated.

On the vdub proc amp controls: is there a way to move from field to field, e.g. brightness to contrast, with the keyboard? I would have guessed Tab, but it doesn't work. Everytime I touch the slider to select the control field, it jumps a bit. Right/left arrows fine for adjustments once there. Also, is there a hidden reset?

I don't believe so. If you access via the Video -> Capture filter... method you can reset with the Default button though.

Hopefully the default 550 vs the default VC500 is enough since I didn't really see a lot of difference.

Have you been able to get the Histogram to display during capture preview? Ideally your samples would be made with Proc Amps set to keep everything in the blue area rather than defaults. Also, with the VC500 you should set the Sharpness control to 0 at all times. Probably the 550 too.

[sanlyn]

Hmmm. Well, okay, I'll ask again: where's all the aliasing and motion noise coming from? I gave up on that problem. Seems to be an interlace problem in your setup somewhere.

Also gave up on the Tap practice scene. The ringing and ghosts were a bit much for one day after fighting it in the guitar clips. Got weary of the interlace problem, too, so the attached samples are progressive. Some of the clips have crushed darks. The s-video cable seems to be softening everything. I have crummy results from most s-video wire, but I found a couple that outdo most of them. You might try this one: http://www.bluejeanscable.com/store/shopbycable/1808A.htm. I've discarded most my other s-videos.

The samples are just quickies to demonstrate possibilities. You could tweak forever. Contrast range of the originals were out of spec for RGB display. The third guitar clip was too soft to work with. As you probably know, that clip changes brightness levels halfway thru. I used Avisynth, VirtualDub, and TMPGenc Video Mastering Works v5. I meant to ask: what does "svhs" means in these and other titles: ?

vhs guitar jvcs svideo +tbc+r3+svhs, aiw vdub clip1.avi
vhs guitar jvcs svideo +tbc+r3+svhs, vc500 vdub clip1.avi
vhs guitar jvcs svideo +tbc-r3+svhs, vc500 vdub clip1.avi

[dianedebuda]

Bleed describes colors extending beyond the objects' boundaries, while by bloom I thought you meant oversaturated.

Bleed is what I meant about the blues. Saw some "too red" in a couple places, so I guess that'd be oversaturation.

Histogram to display during capture preview

Got that working for the VC500. Adjusted Brightness and Contrast so everything in blue area 98% of the time. But from what sanlyn says, my happiness is short lived. Saw what needed to be done with 550, but haven't loaded GraphEdit or GraphStudio for real-time tweaking yet. Didn't touch anything else, including sharpness. Next time...

where's all the aliasing and motion noise coming from? I gave up on that problem. Seems to be an interlace problem in your setup somewhere.

Aliasing is jagged edges often caused by inadequate sample rate - is that even close? Only thing I can think of is that I captured these with Preview on; earlier captures were with Overlay. Probably not relevant, but no dropped frames & cpu usage not particularly high. I was wondering about computer, but you looked at the aiw clip that I accidently uploaded & that was from a different box. I do have the vcrs stacked on top each other. Possible interference problem? I'll seach around the forum for examples of "aliasing", "motion noise" and "interlace problem". I've sure I've seen them, just info overload. If I can "see" them, I can try some changes while waiting for new cable. Wonder if bad svideo cable was why hubby remembered some of the composite captures looking better than the svideo?

The ringing and ghosts

I think that was I was trying to describe by bloom/bleed. Dang, it's hard to use these terms correctly.

Some of the clips have crushed darks.

Hopefully those were the ones without the B7100 C3860 in the names. I tried to use the histogram to set the Brightness and Contrast in those.

Contrast range of the originals were out of spec for RGB display. The third guitar clip was too soft to work with.

Haven't a clue what I'm doing wrong with contrast then.

what does "svhs" mean

I was attempting to document the vcr's settings. Didn't think vhs/svhs mode meant anything except for new recordings and jagabo confirmed it earlier.


Not trying to give you a headache, sanlyn. Was targeting variations in vcr and camcorder settings plus rudimentary proc amp settings. I do appreciate everyone's feedback. Earlier I was pleased with captures so far. Bubble burst. Aunt Jennie IS Mr. Hyde in drag wearing a blue fright wig.

[sanlyn]

The last round of captures is easily better than earlier ones. Keeping levels within the boundaries of the capture histogram aren't that difficult. Except for black borders (usually a little "spike" at the far left side), keep the histogram data out of the "red" side borders.

SVHS: If there's an "svhs" mode in the capture software, then I'm not certain why it's there. The point is: "SVHS" is a tape format. SVHS stores luma and color in a similar way that VHS does, but with a tad more resolution. "S-video" is a type of signal transmission (transmits luma and chroma along separate wires, usually gives a cleaner pic than composite does). s-video can be used to send the signal from VHS, SVHS, DVD, cable box, etc. Can't use it for anything HD, though -- s-video circuits are limited to SD.

[dianedebuda]

sanlyn, I'm glad you think these later captures are better than earlier ones. Like I said, I used the histogram to set proc amp settings, but it was only for the VC500 captures with the B7100 C3860 in the names. For those clips, the signal was within bounds (blue). The last one you said was too soft, but that was one that I'd tweaked & the signal looked fine (to me) as it was recording. Dunno.

I think you're misinterpreting my naming scheme.
vhs guitar jvcs svideo +tbc-r3+svhs, vc500 vdub clip1.avi
tape type = vhs
tapename = guitar
tapeplayer= jvcs = JVC HR-S7500U
tapeplayer cable port = svideo
tapeplayer (internal) settings= tbc on,R3 off,svhs mode on svhs mode is only for new recordings, but I wasn't sure when I orignally started naming the clips. Is useless piece of info here. Sorry if it caused confusion.
capture device=VC500
capture software = vdub

last ones added
proc amp settings = B7100 C3860 = Brightness=7100 & Contrast=3860

Looks like I've got some studying to do. Sigh. Pass the valium.

[Brad]

I'll seach around the forum for examples of "aliasing", "motion noise" and "interlace problem".

If it makes you feel any better, I have no idea what sanlyn is describing either. Could you post a picture, sanlyn?

The s-video cable seems to be softening everything.

How can you tell that? The only composite captures available for comparison are from a different VCR altogether.

[jagabo]

I'll seach around the forum for examples of "aliasing", "motion noise" and "interlace problem".

If it makes you feel any better, I have no idea what sanlyn is describing either.

I don't know what he's talking about either. But I've only looked at the guitar clips. The only significant aliasing I saw was temporal -- the thin line between the frame and back panel of the folding chair. But it's interlaced video so that's expected.

[Brad]

I think you're going to have to decide for yourself whether you prefer the look of R3 on or off. The edges are more sharply defined, but it's definitely either adding halos or exaggerating halos that are baked in. What you should be comparing are vertical columns and diagonal lines; horizontal rows are always going to have the same resolution.

Some of the clips have crushed darks.

Hopefully those were the ones without the B7100 C3860 in the names. I tried to use the histogram to set the Brightness and Contrast in those.

The ones without those in the filenames are the problematic ones. I could post my analysis of them, but it'll be easier for both of us if you pick an R3 setting and then recapture with higher brightness and lower contrast than default (for both 550 and VC500), with Sharpness=0 on the Proc Amp settings.

For our purposes of picking the capture card, it's better to overcompensate than trying to ride the levels as close as possible (i.e. if you think C3860 is a good value, doing C3700 is going to make things easier).

Should I look at the Hi8 stuff now, or wait until after new VHS clips?

[sanlyn]

I'll seach around the forum for examples of "aliasing", "motion noise" and "interlace problem".

If it makes you feel any better, I have no idea what sanlyn is describing either. Could you post a picture, sanlyn?

All frames shown here are fro0m interlaced video. All frames have motion. Apparently the O.P.'s camera doeswn't handle motion well, or maybe it's the cap devices. Probably poor interlace work in the cameras. I don't think it would be both cards, though. Left-click an image for full size display; these are full screen captures of playback with MPC-BE and VLC Player (latter has deint turned on).

VLC Player (deinterlace turned on):

[Attachment 23172 - Click to enlarge]

MPC-BE:

[Attachment 23173 - Click to enlarge]

VLC Player - Live Telecast, SD-DVD cap - has motion on both sides of frame.

[Attachment 23174 - Click to enlarge]

MPC-BE - interlaced newsfilm - figures and camera in motion- HD PVR

[Attachment 23175 - Click to enlarge]

The s-video cable seems to be softening everything.

How can you tell that? The only composite captures available for comparison are from a different VCR altogether.

Looks sharper to me. Doesn't matter. There's so much noise every time something moves, it'll all look alike after cleanup. And if the owner sees no need for cleanup, then again it doesn't matter.

I'll seach around the forum for examples of "aliasing", "motion noise" and "interlace problem".

If it makes you feel any better, I have no idea what sanlyn is describing either.

I don't know what he's talking about either. But I've only looked at the guitar clips. The only significant aliasing I saw was temporal -- the thin line between the frame and back panel of the folding chair. But it's interlaced video so that's expected.

Thank goodness none of my stuff plays like that every time something moves.

[Brad]

Your other examples are video formats that your graphics card's hardware deinterlacer can handle, not YUY2 AVIs.

Here is VLC deinterlacing it:

[jagabo]

Thank goodness none of my stuff plays like that every time something moves.

The OP's videos are interlaced video in a container/codec that's not flagging it as interlaced. Just turn on the player's deinterlacer if you don't want to see comb artifacts. If the OP's goal is to make DVDs she can simply encode them interlaced and let the DVD player or TV handle the deinterlacing. If she wants to upload to Youtube she can deinterlace before encoding.

[sanlyn]

If you set VLC Player's interlace dialog windows to "on", I believe that means it's "on". I don't know that MPC-BE has an on/off button for it, but I'm not going to fish out 30 interlaced and telecined videos to make the point that now and then we encounter crap from consumer cameras that makes me glad I never gave in to the temptation to buy one. The two "OK" captures (one from a DVD in my DVD drive, the other unaltered from my HD PVR video) were made with the same player settings used for the O.P.'s clips, and have no display problems. One of the noisy caps was from the original YUY2 capture with ffdshow's huff enabled (I believe huffyuv was in use there? Why don't other huffyuv videos have ragged edges on my players or PC's?). I have plenty of YUY2 Lagarith and huffyuv videos around that don't play that way. Anyway, I just checked the clips on one of my other PC's that has a different card, and a laptop that has yet another card. I made the progressive mkv's because I got tired of looking at the noise.

The image above in post #52 has ragged edges, just not as many in that particular frame. But VLC player won't play YUV huffyuv or Lagarith YUV videos, so I don't know how that image was made with VLC Player. The VLC player image I posted had been converted to RGB so I could play it in VLC. The two noisy images were made thus:
- VLC Player: Clip converted to Lagarith RGB.
- MPC-BE: Original YUY2 clip with ffdshow's huffyuv enabled -- otherwise, it was scrambled with whatever version of huffyuv the O.P. is using.

When I work on these captures I have to fast-recompress them with ffdfshow's huff to Lagarith YUY2. None of them will otherwise display properly on any of my PC's using the O.P.'s version of huff. I think it was mentioned that the "CCE Patch" version of huff was used. There was more than one version of that patched deal which were issued to handle a compatibility problem with huff and CCE encoders. I have a CCE patched version myself for my captures, which works OK on the capture PC but not on other machines where I have the same version installed (I have to open them in Avisynth and compress to Lagarith to avoid problems). What causes that is a mystery to me. But none of my captures have ragged edges on playback and I've been capturing that way since 2002.

There are a couple of other threads with similar caps that have even worse problems on playback.

[Brad]

Why don't other huffyuv videos have ragged edges on my players or PC's?

Have you played them using ffdshow's Huffyuv as you did this one?

The image above in post #52 has ragged edges, just not as many in that particular frame.

Yadif isn't perfect. None of the frames look like yours; yours looks interlaced.

But VLC player won't play YUV huffyuv or Lagarith YUV videos

The screenshot that I posted is from the original file, using VLC 2.0.8. What it won't do is display BFF videos in the correct order when bobbing is used.

[sanlyn]

My VLC player won't play as you suggested. I used MPC-BE anyway which seems to have fewer problems.

All of the caps I posted from the two media players were from interlaced video. The O.P.'s clips after deinterlacing/processing/reinterlace played the same way. Anti-alias filters had no effect.

[Brad]

We've already explained why your other two examples are irrelevant.

[sanlyn]

Why are they irrelevant? They demonstrate that the players are deinterlacing properly. I'm not talking about combing effects. I'm talking about jaggies.

[jagabo]

When working with interlaced video in containers/codecs that don't flag the interlacing it's up to you to make sure it's handled properly. That's all there is to it.

[sanlyn]

Well, let's see:

deinterlace with QTGMC (also used yadif and bob)
cleanup
reinterlace (Separate fields, weave, etc.)

That's not correct? Plays back with jaggies and twittering lines in all players on all my PC's, just as in the originals. Plays that way as lossless AVI and encoded with MPEG or x264. That doesn't happen with other videos.