When you say at random points in the tape, I hope that is not for the same tape on the same player with the same tracking setting. If that is occurring, then there may be something more serious happening with the player. With a healthy player, you should have fairly repeatable behavior when you reach a bad spot on a tape with the same player and same tracking setting. Not an absolute rule, but should happen most of the time. Naturally, you can get different audio/video performance from the same tape on a different player. That's what you may have to fall back on to get enough good bits to splice together. Of course, video will be unavoidably different looking in color or contrast somewhat when switching players, irregardless of what other adjustments you make to any proc amp controls. So, ideally, try to get the video to all be from the same player.
Not sure I can add anything you haven't already read, but I'll share my experience and what I've done. Capture first with the same tracking setting throughout (either auto tracking, or manual adjustment that works best for majority), then playback capture and note what time marks/frames (in
Virtualdub) have tracking issues (video and/or audio). You can also make notes while capturing. Go back to each on the tape and fiddle with various tracking settings until you get the best you can, and capture that for "coverage". Using
Virtualdub (and probably avisynth as well), splice together the best combination you can. When splicing, you need good attention to detail, because you have to match up the exact frames from each clip. Having multiple instances of Virtualdub open is necessary here to find the frame numbers you need to match up between clips. This can become a lot of effort quickly, depending how many inserts you have to do. Avisynth is easier to splice multiple fragments together (using a.trim(b,c)++d.trim(e,f)++g.trim(h,e)). Virtualdub only allows appending video to the end, so it's not as good for complex jobs, but it can be done. You just have to save each fragment to a separate file first before appending them all together in Virtualdub.
Regarding audio, you can have various scenarios. You could have the audio good throughout the original capture, but with some video issues in spots. You could have audio less than ideal in spots of original capture, but good video. Or you could have both bad audio and video in the same places. I have had some success with getting good audio out of a spot by adjusting tracking that went beyond the good video limit. If the HiFi cannot be salvaged and you have a player that allows control of this, you can force the player to output the linear audio track only. Perhaps that won't have the glitches found on the HiFi. Quality suffers, but it may be superior in some cases. If you have bad video but good audio, you could splice together a good audio version and a good video version from different clips, then make a new hybrid from those two that have both good audio and video. More on how you can do that later.
When doing this kind of thing, it assumes you've got great audio/video sync. So a frame TBC is likely a requirement, so you can maintain audio and video sync throughout these rough spots. If you don't have synced audio between each of the insert clips with respect to the main capture, I don't know if there's any good solution there. You could try splicing for video only, and use the complete audio track from another capture. But that still may not work out.
I've had the experience of another VCR with inferior video quality (i.e. a non-SVHS machine) playing the HiFi audio better, i.e. without crackles or falling back to linear audio, and so I capture some or all of it with that VCR. Often it was by using the VCR that made the recording (because it's alignment was different). Then get your audio capture and video capture videos trimmed up so the start and end frames line up and they have the same number of frames. Load the audio capture in Virtualdub and save audio portion as .wav. Now load the .avi to be used for video, and set the Audio menu settings to use the .wav file you just saved for the audio stream, while using Direct Stream Copy for the video setting. This will mux these two separate audio and video portions into a new .avi. You would do this only after you've got all your splicing done for the best looking video. Use Direct Stream Copy in Virtualdub whenever splicing video to save time and potentially quality of video from re-encoding unnecessarily.
You referenced a post from 10 years ago, that also got into the topic of fixing issues like dropouts. There are more solutions for those kind of issues than there was back then. If you weren't already aware, search the forum for RemoveDirtMC, RemoveSpotsMC, ReplaceFramesMC and you should find some examples and links to the .avsi files and relevant Avisynth plugins needed. I am quite impressed with how much can be cleaned up with these "Replace MC" type functions. It won't resolve significant tracking issues. But those single frame dropouts and glitches can be reduced/removed quite surprisingly in many cases.
Best of luck to you!
