Hi Sanlyn, apologies for the silence again, been focussed on some other domestic stuff so I don't get much time to spend on this more interesting project :depressed:

Anyways, yes you are right in that I am in the first stage of trying to make a workable lossless capture. I don't want to rush ahead but there have been times where I need to get some content off right away for friends or relatives and I kind of panic as I do not know what / how to get them the content they want in a rushed manner so if this question comes up again I now know what to do.

Just one clarification - you mentioned that I should cut the losssless capture as desired with VirtualDub or avisynth, please can you advise the best way of doing this as every time I have tried doing this, it again takes a long time to cut out the parts I want so in some ways it is less headache for me to simply convert the entire file to AVI rather than go through a laborious process of cutting then converting....unless I am doing something wrong?

Now onto other matters with respect to teletext or CC information...I have a good VCR (Panasonic NV-FS200) and it is managing to give me some good video outputs onto a regular TV set including legacy teletext information (see attached image)...when I previously asked this question on this thread, I was informed that if I use a TBC, all this information would be stripped out but as I am now getting good results I am wondering whether there is any way of preserving this information and reading the data somehow once it is digitally captured ? Is there a suitable decoder available ? The reason this information is useful is because I can timestamp some of the recordings down to the second it was broadcast which could be useful for future references....any help would be appreciated.

Finally, does anyone know where I can get a good TBC from ? I have tried ebay but these seem to be quite rare especially the PAL ones

I don't know what you mean by "convert to AVI". Aren't you already capturing to lossless AVI? Why are you "converting" if you're already in AVI? AVI is a container, not a format. As a container it serves as a wrapper for video encoded in various ways using different codecs, including lossless compressors (huffyuv, Lagarith, UT Video, etc.), Xvid/DivX, Type 1 or Type 2 consumer DV, or MJPEG, etc). So what are you doing when you say "convert" from AVI to AVI?

Avisynth can make cuts using the Trim() command. Or, If you open a lossless AVI in VirtualDuib, you can use the navigation keys in the lower left corner of the window and the commands in the Edit menu to cut out sections of video. Either way, the cut segments are saved out of VirtualDiub using "direct stream copy" mode instead of "full processing" mode. Using direct stream copy, an AVI that goes into VirtualDub as lossless huffyuv will save as the same format. The other way to keep the format as lossless AVI is to set the colorspace and compression you want in VirtualDuib, then save using "fast recompress" mode -- this is the save mode used when applying Avisynth filters and image modification to lossless video. By themselves, the Trim() command or VirtualDub edit commands don't modify the images.

If you are using an NLE to make cuts and you want to save to lossless format, you'll have to recompress the results using the same lossless codec you started with. How you set that up in your NLE I wouldn't know, it depends on the software's interface setup.

External frame tbc's aren't always needed, but I'd say they are most of the time.
I can't answer for teletext or CC information. Why are they needed?

Common External tbc's like the AVT-8710 or TBC-1000 work with NTSC and PAL. Another source of frame sync is a DVD recorder used as a pass-thru device between player and capture card. Pass-thru can't defeat macrovsion, isn't quite as powerful as an external frame tbc, and the machines are usually dedicated to either NTSC or PAL formats. Tbc's are as difficult to find in North America as they are in other countries.

Hi - sorry for not being in contact recently, other priorities mean I rarely get time to spend on this. Could someone kindly confirm whether the For.A 370 is a good TBC compared to the datavideo TBC's like their 1000 or 3000 series?

-- merged --

Hi - please can someone kindly answer my question on TBC's in my last post? I am not sure whether to hold out for a datavideo TBC or go for a For.A one?

Some folks like the For.A TBC, but I do not. I completely dislike those huge clunky TBCs, as the same functionality can be had in a small box from Cypress/AVT or DataVideo.

If you have plenty of space, and can get it dirt cheap, then I'd understand.

If you're in gambling mood, bid $125 on this: http://rover.ebay.com/rover/1/711-53...-/182351935832

Thanks for your advice @lordsmurf. I need a pal TBC, is there an easy way of telling whether the tbc you linked to is PAL compatible? Also is the functionality the same between the For.A and the datavideo leaving aside the physical size issues?

The TBC-1000 is PAL compatible. That unit doesn't have a power supply, and you'll have to get one. Off-hand, I don't recall the voltage/etc needed, but it's probably in the user manual, available for download on this site.

Again, it's not 100% guaranteed, so consider it gambling. Odds are, however, that's it fine. A guaranteed-working unit is worth 3x that amount.

The For.A has an integrated proc amp. You can get the same on the DataVideo TBC-3000 for about $500 -- if you can find one.

This is mostly about preference. Do you need or want it?

I just checked with the seller and he said that his TBC-1000 is NTSC not PAL. When you say it is PAL compatible, are you speaking generally or about all TBC-1000's?

I need a TBC desperately as I need to get on with the daunting task of transferring my many tapes but TBCs rarely come up for sale on ebay

I've never seen an NTSC-only TBC-1000. The boxes were often marked "NTSC", but would play a PAL signal with zero issues. I've mentioned this many times in the past 10 years.

Hi all, I am really excited as I have finally managed to track down a TBC1000 hopefully in working condition (will find out tomorrow). So after tomorrow, my capture set up will go like this:

VCR: Panasonic NV-FS200
Time base corrector: TBC1000
Capture card: ATI 9600

Do I need anything else or am I now good to go? Do I switch on the TBC on the VCR as well as the TBC1000?

As an aside, I have also managed to pick up a DAC-200 which as far as I can tell is a DV capture box. Do I need this or should I try and move it on bearing in mind all of my content is Hi8/VHS with a small number of Betamax tapes....

Thanks to this lovely forum for getting me to this point - I am really grateful and pleased!

Looks as if you're good to go.

AS explained earlier, you have two types of tbc available. One is a scanline-level ("line" tbc) in the VCR. The other is the full-frame sync tbc ("frame" tbc) in the TBC-1000. You need both.

That's a firm and definite no. Heaven forbid.

Thanks @sanlyn

Is the DAC-200 similar to the canopus advc-300 just for my information?

Also I will enable both TBC's and post results here over the coming days...will need some further assistance post capture to clean up the videos...I am really excited to spend the holidays doing this :)

Is there anything I need to do in readiness for bulk capture in terms of settings etc?

I think I referred to a recent VirtualDub capture settings guide that goes into considerable detail, if you want to check them against your default settings: http://www.digitalfaq.com/forum/vide...-settings.html. You can post short unprocessed samples from AVi captures directly in the forum, up to 99MB each (10 seconds of losslessly compressed YUY2 capture will easily fit within that limit). Cut samples in VirtualDub and save them using "direct stream copy" to maintain the original colorspace, lossless compression, and other properties.

All received- just one thing before I hook it up, do I route the audio through the TBC or bypass the TBC and go direct into the ATI? I am sure the answer is somewhere on this site but couldn't find it....

[EDIT]
VCR Video Out -> TBC -> ATI
VCR Audio Out -> TBC -> ATI
Use the "A" outputs from the TBC-1000
Some external tbc's won't accept audio, but the TBC-1000 takes both.

Thanks Sanlyn, just came across this thread where @msgohan indicates that passing through the audio is better for sync - are there two schools of thought on this?

http://www.digitalfaq.com/forum/vide...ect-audio.html

The purist audiophile school of thought is that the more connections and circuitry you have, the more likely it is that some small amount of noise or degradation is involved. More plugs/jacks + more wires = more chances for less fidelity. If you can hear a difference, you're pretty good and are probably listening through $500 high-end electrostatic phones and a lab-quality head amp(!) for pass-thru. Audio capture works whether you go through the TBC or go directly to the ATI card. To go through the TBC you need two lengths of audio cable, to go around it you need only one. If you fear audio noise (and if you can possibly hear it, which is iffy), bypass the tbc. In my case I don't feed audio through my tbc.

Just done my first capture using the TBC and I can report that it dropped 1 frame out of 78797 total frames - I think that is pretty good but what do the experts think ? Also, for some reason it inserted 3 frames - does anyone know why this could be if it only dropped 1 frame ?

I also connected the audio directly to the ATI and used S-Video connectors from the VCR to the TBC and into the ATI but with the TBC enabled on the VCR (Panasonic NV-FS200), the colours look too bright, they seem to look more natural to me with the TBC switched off. Does anyone know why this could be ?

I will try to upload a video but how do I easily crop 100Mb from a file size of 24.58Gb in VDub ?

Not unusual, but it could be happening because you told VirtualDub to do it. Chek your settings in the "Timing Options" area of VirtualDub as covered in this post from the settings guide" http://www.digitalfaq.com/forum/vide...html#post45242 areas:

Some internal tbc's have that affect, but it also depends on the individual tape. Unfortunately your capture probably has multiple scaline timing errors and jitter, which can't be fixed later. Ultimately the incoming luma levels should be adjusted anyway using your card proc amp and VDub's capture histogram to check for valid y=16-235 luminance levels whether you're using the TBC or not. If your monitor isn't precisely calibrated and/or if you're not using a histogram, and proc amp controls to check incoming levels, how do you know it's too bright? Just saying that it "looks" that way isn't as foolproof as checking with the proper tools, as shown in this section of the recent guide: http://www.digitalfaq.com/forum/vide...html#post45238.

I will try to upload a video but how do I easily crop 100Mb from a file size of 24.58Gb in VDub ?[/quote]There are edit start and stop marks in the lower left-hand corner of VirtualDub's window. Move to the start point and click the start-point icon, then move to the last frame you want to select and click the end-point icon. Then use the "Edit" top menu to delete. Before saving the file set "Video"... -> to "direct stream copy" (DO NOT USE "full processing mode" !!!!!). But rather than make a whole copy of your capture, why not delete what you want during post processing? How do you intend to process this capture?

Please find it attached - for the eagle eyed amongst you, this is the same clip I previously uploaded when I first started capturing before I had a TBC...any comparisons ?

Shall I post this same without the VCR TBC switched on to give you an idea of what I mean about the bright colours or is this clip sufficient ?

Thanks for the tips on the dropped/inserted frames @Sanlyn, I shall play around with the timing options as mentioned in the very useful guide...is there a guide on the histogram feature, I remember reading it somewhere on this site but cannot find it anymore - it went into a lot of detail about cropping the frames then adjusting the levels accordingly - I think it might have even been written by you Sanlyn ?

Levels look about right in the capture sample just posted. I noted what appeare to be more chroma noise (rainbows) than I saw earlier. In a media player if you pause on a frame you can see cyan and magenta blotches in the dark sweater material, and if you look closer you see the same thing in the red sofa background. There is also some line twitter at the top border. There is less posterizing effect in the skin tones this time and highlights don't look like "hot spots", especially on the face. I didn't see scanline "wiggles" or ragged lines.

There are pictures of a corrected and uncorrected capture histogram in the post at http://www.digitalfaq.com/forum/vide...html#post45238 and an image of the crop dialog window in the following post at http://www.digitalfaq.com/forum/vide...html#post45239 . There were two earlier posts in which a little more detail was discussed, but one post was just a copy of the first one, as in this post: http://www.digitalfaq.com/forum/vide...html#post44278.

Thanks Sanlyn - I think there is something wrong with my TBC :depressed: as on very bright images, the image is washed out completely but with the TBC bypassed it seems to be fine...see the two clips in this post and the next one (note: both have identical Proc Amp levels on both clips)....what could be wrong ? The input to the TBC is S-Video and not a macrovision tape (recorded from satellite)

Brightness test without TBC clip attached

Thanks for the With/Without samples. Yes, the TBC does have a problem. The glitch is in the gain circuitry, not brightness. Gain will stretch the histogram from the dark portions outward toward the brights. Higher values of bright pixels are increased more than darker values, so the stretch is more logarithmic. Dark pixels like y=16, which is nearly black, are amplified ess than the midrange around y=128, then brighter pixels are amplified more. Here, pumping of the brighter pixels is going haywire.

You can get numbers and a histogram that measure the changes by using the script below. First, remove dark borders and head switching noise, then crop the frame so that it's valid for YV12:

The picture pbelow shows the top halves of two frames, frame 8 in "without TBC" on top and frame 14 in "With TBC" on the lower part. In frame 8 without tbc the minimum ("Min") y value is at y=16, and y has about the same value in frame 14 of With TBC. There aren't that many brighter p0ixels in those frames, so the boost in brights isn't so obvious. But you can see that the "Max" value in the boosted frame 14 With TBC is appreciably higher than the "Max" value Without TBC.
http://www.digitalfaq.com/forum/atta...1&d=1483033612

The images below are from the brighter frames, frame 98 in Without TBC, and frame frame n104 in With TBC below it. The y value in the :"Min" darker pixels are about the same in both frames. But brighter pixels in the Max" area of With TBC is off the chart. If you can't spot where the white "luma" graph is located in the bottom image, it's the thin yellow "spike" all the way against the right-hand edge of the histogram.
http://www.digitalfaq.com/forum/atta...1&d=1483033816


This particular also has a problem with chroma flicker. In alternating frames the sky is more blue-green, then more toward magenta, and back again. Below, from Without TBC, are the top halves of frames 100 and 101. The sky and balloons are a different hue in alternating frames. If you open the sample clip in VirtualDub, play it by scrolling through the bright frames one frame at a time, frame by frame, to see the effect. When played at normal speed the frames will flicker rapidly. I didn't see this effect in either the bright or dark pixels of the David Bowie clip.
http://www.digitalfaq.com/forum/atta...1&d=1483034084

This can be fixed in Avisynth using something like the ColorLike() plugin, but it's a pain in the neck.

Wow, thanks for the comprehensive review Sanlyn - I think I have located the TBC issue fault and discovered that it lies with the power adaptor. The adaptor that came with it is 9v wheras the one I am now using is a 12v one and that seems to have improved the situation - I now have 0 dropped frames and 0 inserted frames...will post another file soon of the same clip with the repaired TBC. I think it also fixes the alternating colours but I would be grateful if you could kindly check that for me when I upload it...

One issue I am still having problems with is the in-built VCR TBC. This TBC seems to be interfering with the colours (see attached sample video). The 1st 5 seconds of this clip have the VCR TBC OFF and the latter 5 seconds have it switched ON...any thoughts on this please ? Shall I do all my captures with it off as I think it looks much more natural with it off....

This time the problem affects midtones and brights mostly, but the entire signal is depressed in intensity -- you'll notice, however, that darks and blacks don't change. Hopefully not another power issue, but it could be (lightning striking twice?). Sad to say, the tbc off has line scan errors and more ragged edges. At one point a subtle horizontal line of mild ripple slowly floats downward but disappears as soon as the tbc is turned on. You can see ripple distort the inside edge of the TV.This would indicate tape surface irregularity, which is very common. The color balance doesn't change, but overall intensity and color density does. Trying to correct that problem in post processing would hzve the effect of completley washing out bright detail.

If this line tbc can't be repaired, it might be necessary to obtain a recommended DVD pass-thru device for line tbc activity. The tape you played seems fairly clean and tracks well, but lesser tapes and other problems won't play well at all and will have more obvious scan errors, wiggles, line twitter and other distortion. Those errors can't be corrected after capture. A pass-thru device such as an used Panasonic ES10 or ES15 would very likely be far less expensive than a VCR repair.

Thanks Sanlyn, when you say the entire signal is depressed in intensity, could this be because of my proc amp levels or a VCR fault? Also a minor issue (with or without the VCR TBC) is that I can see a colour mist on blacks when I am expecting to see pure black...could I clean this up in post processing so that the blacks are pure clean black without colour mist if you follow me?

Also could you kindly post a screenshot of what you mean by mild ripple with a histogram of the midtones and brights to get an idea of what it should be and how far off I am ? I would do this myself but don't have a clue how to analyse the information..do you think it is a VCR fault overall with the TBC circuit or something else inside the VCR? I still have my mitsubishi VCR, should I try capturing the same clip with that machine? It doesn't have a line TBC though....

You didn't change your proc amp settings, you turned your TBC off and on. So it has to be the VCR.

I don't think you refer to the big TV cases in the sample shots shown in this post. The TV cabinets in the images are actually very dark blue around RGB 45. If you mean a kind of floating gray grunge in large shadow areas, that's tape noise and can be filtered out in post processing. A VCR's noise reduction seldom gets rid of all of it.

It's impossible to convey in a still shot what happens during motion. Open your sample clip in VirtualDub and scroll at double speed while observing the left-hand edge of the TV console. You should not see ripples, raggedness, or changes in vertical edges if the tbc is working.

Images attached with VirtualDub ColorTools RGB histograms, TBC OFF (top image), TBC on (lower image).

TBC OFF (frame 127):
http://www.digitalfaq.com/forum/atta...1&d=1483161740

TBC ON (frame 130):
http://www.digitalfaq.com/forum/atta...1&d=1483161867

Learning to work with histograms and similar tools is essential for video work, as you've seen during capture. It's easier to understand articles with examples than it is for you to try to decipher brief notes with no examples. The links below deal with still photo histograms, but the principles for video graphics are exactly the same.
Understanding histograms Part 1 and Part 2
http://www.cambridgeincolour.com/tut...istograms1.htm
http://www.cambridgeincolour.com/tut...istograms2.htm

It's a VCR fault. Many people have more than one VCR, and many of those VCR's don't have TBC's. Rather, they are used with a DVD pass-thru device. The use of a line tbc pass-thru means you're not tied down to one VCR, which means if your primary VCR can't track a tape well (and it happens often) you're up the proverbial creek with no alternative.

Information about pass-thru devices used as line tbc's:

http://forum.videohelp.com/threads/3...=1#post1882662

http://forum.videohelp.com/threads/3...=1#post1983288

http://forum.videohelp.com/threads/3...hat-do-you-use

I have a primary VCR with built-in tbc, has been repaired 3 times, maintenance is costly, and some tapes don't track well. Meanwhile with my ES10 and ES15 pass-thru components I have 3 other mainstream Panasonics at my side, two of which are s-video machines but none with tbc's.

Hi Sanlyn, I have done some investigations on the VCR and discovered a few things that needed addressing in the power supply. I have done a new capture not exactly of the same clip I previously uploaded but something of a similar scene. Please could you take a look and confirm if this has resolved the issue ? It still seems to make the image depressed but the colours are much more vibrant now...so much so that I have to reduce the saturation levels to something like 81 or 82 from 128....once again the first 5 seconds are TBC OFF and the latter 5 seconds are TBC ON...

The strange thing is when I view it on a normal TV using a scart connector and try switching TBC ON and OFF, the image doesn't visibly change in brightness or contrast, it only seems to have this affect via the ATI and datavideo TBC configuration. Should I be connecting to the computer using the yellow video out cable instead of the SVHS cable if you conclude there is a further problem ?

For comparison purposes, I have just done a composite test just to eliminate S-Video specific problems and I think there is still a problem, could you kindly confirm for me please ? It is the same format - first 5 seconds are TBC OFF then TBC ON...

Sorry to keep uploading files but I guess more information is better than limited information.

I have just tried a different tape and it seems that the TBC problems are not present on this one - could it be tape specific ? Is the TBC really sensitive to different tapes ? Same format applies to this clip - first 5 seconds are OFF and next 5 is ON....

Just thinking about the previous tape, the FS200 did have some difficulty tracking it but it does eventually track it, could this be the reason why the TBC is also having problems?

TBC Test v3.avi has exactly the samer problems as the previous test from that same video.
Still Wanted VH1 TBC test.avi doesn't have any of the effects and appears to be operating normally.

Whether or not mechanical tracking makes changes when the tbc is turned on or off, I wouldn't know. But I wouldn't think so. However, it's true that the tbc circuit interprets playback differently when the signal is going through it than when the signal bypasses that tbc circuit card. If the tbc circuit had not been powered correctly, it would obviously not operate correctly. If the power problem was corrected, tyhe cirtcuit should behave.

These old VCRs are all getting long in the tooth and exhibiting dimented behavior at times. They are extremely difficult to maintain and rebuild. I finally grew so impatient with my AG-1980 I had to stop using it, and the current one is my second copy. Every rebuild produced a different problem after a few short weeks. I finally went back to upscale mainstream non-tbc 1996 Panasonics and a remade 1991 SONY classic, coupled with an external tbc pass-thru device and my 2004 AVT-8710. Except for the noise levels, performance was approximately equal and tracking was slightly more stable. It means extra cleanup steps in post processing, but cleanup was always needed anyway. Since you have VirtualDub and various histograms and 'scopes at hand, you should be able to test-fire your other tapes to make sure things are working with your FS200.

Hi Sanlyn, well I think I have finally identified the source of the interference and it WASN'T the VCR but the capture card itself ! I can't believe it but I think it may be true. To prove this point, I have uploaded another TBC test of the same clip I previously uploaded (the one you did the helpful RGB histograms on). Please could you kindly analyse this clip again and confirm if everything is in check - first 5 seconds are TBC OFF on and next 5 seconds are TBC ON. Overall, I think the picture looks much better and the blacks look much more black...could you also confirm how I could clean this up ? Could I make the black pure black without the random colour patterning ?

Thanks for your help - I think I am almost ready to capture in bulk now

Thank you for the new sample.

The TBC has the obvious effect of curtailing highlights and constricting the histogram leftward toward the middle range. You can see this in a histogram as well as I can, but I've attached a video sample with a histogram in the image. Highlights are blown out and gamma is too high, so that the white jacket "glows" and dominates the picture. The TBC does make a smoother image during motion, apparently by cleaning up some line skew. I also see that you have a slender, light gray vertical bar of noise along the left border.

Your capture has illegal luma and chroma levels and a cyan color cast. Frames 10 to 19 are missing, replaced by 10 duplicates of frame 10.

Black is a color that contains equal proportions of red, green, and blue. If a black object has a predominance of any one color, it will not be perceived as clean black. The dark TV consoles in the capture have a blue color cast. But as far as that goes, you don't have a clean white, either (white here has a cyan cast -- which is blue-green). Because of the color imbalance and high gamma, facial tones look washed out.

A "pure black", as you put it, has equal proportions of 0 RGB value. That is, pure black is Red-000 Green-000, Blue-000 (or "RGB 000-000-000"). A very dark gray "TV black" is RGB 16-16-16. The dark grays in the darkest shadows of a white garment would be RGB 64-64-64. Middle gray lies in the middle of the spectrum: RGB 128-128-=128.

Beyond the midtones, lighter grays up to the color white also have equal RGB proportions. A light gray is about RGB 180-180-180. White starts looking "white" at about RGB 200-200-200. "TV white" is RGB 235-235-235. Super-white is RGB 255-255-255 and usually looks rather unrealistic when displayed.

An RGB-0 black has no detail. If you adjust colors so that the TV consoles in the images were really pure black, they would have no detail at all. They appear to be a very dark gray at about RGB 35-35-35 or so, but the consoles won't look that way at all points because they reflect studio lighting in varying degree at various locations.

I achieved more realistic and valid levels by using Avisynth's Levels filter to tame highlights down to y=235 and to lower gamma by a few points with the command "Levels(16,0.9,255,16,235, coring=false, dither=true)". I then worked up a more accurate and pleasing color balance and dimensional quality by using VirtualDub's ColorMill to do the following:

- Lower the midpoint from RGB 128 to about RGB 100 and adjust the midpoint rolloff at the lower ranges to prevent an effect that appeared to be "glare" in the lower regions below the midtones.

- Partially repair the cyan over balance by moving Hue toward red about 5 points.

- Cleaned whites by adding bright red and reducing bright green and bright blue.

- Made dark grays look more black by reducing blue and green in the dark region.

- Achieved a more natural skin tone by raising midtone saturation slightly.

- Tamed the glare of the white coat by reducing highlights slightly with ColorMill's "Bright" levels setting.

The results of the levels and color re-work plus some light denoising are attached.

If you are unable to view video frame by frame or to analyze with histograms and pixel sampler devices, video processing will be more frustrating than it should be.

Wow thanks for the very useful and in-depth analysis Sanlyn :salute:. You are right in that I am unable to view video frame by frame or analyse with histograms and pixel sampler devices, so this is perhaps a point of frustration for me right now. Where do I start with this ? You kindly provided some references in post #147 but how do I get to the tools to be able to produce the cool histograms you have just uploaded ? Also, if I have a 3 hour tape, how do I quickly check for dropped or duplicate frames without having to manually go through frame by frame on VirtualDub ?

Now onto the upload itself - I am concerned about the dropped frames you mention, indeed I have just uploaded into VirtualDub and can see exactly what you have mentioned...but how could this happen if I am going through a frame TBC (TBC1000) ? I am concerned about this as if this can happen on this short clip then it could happen on a lengthy 3 hour tape then I will have to re-capture the whole tape again.

In terms of the illegal luma and chroma, I set the proc amp levels as per the attached screenshot. When I checked this in the VirtualDub histogram before I started capture, everything looked ok and the histogram showed blue peaks largely without any reds. What could have caused these illegal levels ?

Thanks for the advice on pure blacks, I guess the picture looks a little noisy to me the noise shows up in places I expect to see as pure black if that makes sense. This clip is probably not a good one for the reasons you mention in terms of studio lighting but there are others I could upload showing better what I mean. I will try to do this in the next post.

How do I get a real time analysis of the RGB values while I am viewing a captured AVI or during capture ? This could make my task easier to spot these problems during capture. What should I be doing pre capture to identify and resolve these issues ? I really want to prepare myself better to avoid any surprises post capture...most of the information you have pointed out below have really come as a surprise to me...

By the way, the slender, light gray vertical bar of noise along the left border is actually part of the analogue scambling they used on satellite back in the 90s (the corresponding scrambling information can be seen on the extreme right of the frame). This is how they used to scamble and unscramble analogue video - I'm glad that some things have moved on for the better :congrats:

Thanks for all the cleanup work you have done, it looks amazing ! I wish I knew even 10% of what you knew, that would make it much easier for me but I am still at stage 1. Are there any other tutorials you can point me to ?

Finally, after your analysis, do you think the VCR TBC and VCR itself are now producing quality captures or do you think there is a still a problem with my VCR and or capture card ? Overall, I think the image quality seems to be better than post #147 and I don't seem to be getting the crushed RGB values when I switch on the VCR TBC, is that right ?

Thanks once again for all your help and advice

This is what I mean by colour patterning on a greyscale clip :question:

The proc amp values shown in your jpg are almost exactly the same values as the ATI defaults, which is not much of an adjustment -- that is, if one was required. Adjustments are made in Virtualdub capture with the borders cropped off (temporarily) and the histogram on display. Different segments of a video can change levels, so I usually play a minute or so, perhaps check another segment of a tape that I know has outlaw values (and might have to capture that segment separately with different settings, although that's unusual). I check briefly for overruns or flashing patterns that exceed the desired y=16-235 range, then adjust for the worst case scenario and tweak later where needed. It would take forever to capture a tape using start-and-stop every time levels undergo a minor change.

Levels at the start of your capture appear to fall off the right edge beyond the scope of the graph, then get "pulled in" when the TBC goes on.

Frames might be missing/inserted if you changed a setting in the VCR that affected playback speed factors. The missing bleeps look like the very start of capture, where it can happen until other components get up to speed for a few frames. I start Virtualdub capture a few seconds before I start the VCR, then trim off the lead-in excess.

The captures look better, but you need to adjust levels with the TBC turned on if you're going to use it. It does change levels at the high end, possibly there's an autogain limiter involved with it.

The basic YUV histogram is from Avisynth. I write a text script to open the file and display the histogram. The "Histogram" command has several parameters that determine the type of graph. The graph used in the sample is the "Levels" version. The script is opened and runs in VirtualDub. The three commands I used to open the file are shown in the code below. Note that the path to your video is its location in my setup -- you would have to change that path to match your own system. Also note the conversion to YV12 (this particular histogram works only in YV12, Others can work in YUY2):

A video file or running script can be viewed frame by frame in VirtualDub using the scroll controls and icons in the lower left corner of VirtualDub's window. But any NLE with a histogram control can be used to run the video.

I took pixel value samples using a handy desktop tool called CSamp, a small standalone icon that reads pixel values where I place the mouse cursor. Csamp is a bit fiddly and makes you hold down the left mouse button during readings. Something more friendly but larger onscreen would be the free ColorPic, which is resizeable and like CSamp it places a runtime icon in your taskbar's right corner. CSamp with instructions was posted here a while back (download link): http://www.digitalfaq.com/forum/atta...-time-csampzip. Below is an image of CSamp in use with another filter known as the Photoshop-style gradation curves filter in a Virtualdub window (the CSamp control is the small window in the middle of the image):
http://www.digitalfaq.com/forum/atta...ame34_curvejpg

The ColorPic filter is larger but less fiddly and "reads" at all times when running: https://www.iconico.com/colorpic/. It has several panels shown on the website, but most of them can be closed to make the form smaller.

The ColorMill Virtualdub plugin is a popular color tool that works in separate color ranges like those filters in bigger guys like Vegas or Premiere Pro. But it uses sliders instead of wheels and is less complex. Free. Zip attached, with source code and a .txt help file. You won't need the c-language source code, just the filter.

Learn this stuff one step at a time. It helps doing it that way, because just when you think you've got it all down you run into a new glitch you never saw before.

Thanks for the encouragement Sanlyn and the advice. I will investigate the points you mention above.

By the way, I cropped the video first then adjusted the proc amp levels...and the result was the settings I provided. Do you chance the hue levels by eye or is there a tool for hue levels like the histogram? I thought the histogram only deals in brightness and contrast?

Regarding dropped frames, if I have a 3 hour tape, how do I quickly check for dropped or duplicate frames without having to manually go through frame by frame on VirtualDub ?

Also does the VCR and Capture card setup look good to you now? Is it better than post #147?

As noted earlier, I think that if you checked levels with the histogram while the TBC was turned off, the reading appears to have exceeded he bright range that the histogram could read. It's also possible that the particular scene pumped levels higher than in other segments.

Per notes in the previous post, hue was adjusted with the ColorMill control in VirtualDub while thje Avisynth script was running. Corrections were made with the YUV histogram histogram turned off. ColorMill is an RGB-only VirtualDuib filter. Hue was not a major correction. I wouldn't do it during caspture -- Color balance will change in VHS by the minute. If you capture with color corrections and change your mind later, it's very difficult to undo in post processing and can be a real pain in the neck when the corrections don't look so great in other segments of the total capture.

10 duplicate frames frozen in time is not something you're likely to miss when viewing a capture. 1 or perhaps 2 frames is often unnoticeable. But if you find this happens often, something is amiss. In VDub capture I turn drop/insert frames off. If you turn it on, you'll see a few random drops with bad tape and it will be reported in the log window at the right of he capture window. If you see lots of dropped/inserted frames one after another or in groups and pairs, something's wrong.

I've already noted that your captures look better, but that TBC definitely continues to alter the image levels when turned on but not as destructively as earlier. Take histogram readings with it turned on, not off.

Hi Sanlyn, thanks for all the tips and advice, it is really useful for a novice like me to understand the details of how to analyse these clips so that I don't need to ask for help everytime.

While I get up to speed with this information, could you kindly take a look at the clip I have uploaded with the TBC in the off position and in the next clip I will upload another video of the same segment with TBC on for comparison purposes. Does this look any better or worse compared to my previous upload ? I changed some components on the ATI board so hopefully this has improved things..in the meantime I will try to analyse myself using the plugins you recommended but it might be quicker if you can help one last time.

I hope there is no more illegal chroma/luma....

Just one question, where do I download Avisynth from ?

2nd TBC test with TBC switched on