VHS playback video AGC? Bright flashes = waveform monitor loses sync

[Brad]

This all started from trying to inspect some Macrovision tapes on my waveform monitor and other devices. I've gotten increasingly confused by inconsistencies, and measuring MV-free pre-recorded Hollywood movie tapes hasn't clarified anything yet.

I've fallen back to one of the most basic VCR models I own, a JVC HR-J693U, because I'm hoping to eliminate the variable of possible signal tampering by the VCR. The tape used for this post was the US release of The Fifth Element.

The attached capture was done with a SAA713x card on Windows XP with Vertical Sync recovery mode set to Free Running, Automatic Gain Control disabled, manual gain set to 0% ("-3dB") and Brightness boosted to make the Vertical Blanking Interval more visible. I encoded the Huffyuv capture to 4:2:2 x264 for posting.

The other attachment is a smartphone recording showing the Leader 5870 waveform monitor. It's a different playback of the tape than the capture, but I trimmed them to match. The first flash takes place on frame 414 (WFM) and field 414 (bottom of capture).

Note: I haven't bought any 75 ohm terminators for the WFM. Instead I'm feeding the loopthrough output into the input of a DMR-ES15. This should work just as well, no?

Waveform monitor display
The WFM recording starts with the VCR's blue back, showing that -40 IRE and 0 IRE are where they should be on this generated screen. On the WFM, the first two flashes (at 0:08 and 0:10) cause it to lose sync. There's another blip at the 0:40, but it's much more mild. The COUNT overlay seems constant, at just under 90 IRE (except it's slightly lower on blue back vs tape playback).

SAA713x capture
Sync level is always clipped, so there's no way to see its actual value.

At the first flash on the capture:

1. Blanking level drops, both horizontal & vertical.
2. COUNT overlay drops.

[Attachment 50956 - Click to enlarge]


[Attachment 50957 - Click to enlarge]

Over the next several fields, the overall level of the flash itself, the COUNT overlay, and vertical blanking all gradually rise. Here is field 420, where they're nearly back to the level of 414. Horizontal blanking is still clipped, and it doesn't return to normal until it jumps up at field 430.


[Attachment 50958 - Click to enlarge]

At the end of the next flash (fields 591 & 592) there's a corresponding effect in reverse. When the bright field flashes immediately back to dark:

1. Blanking level abruptly rises (H & V).
2. COUNT overlay abruptly rises.

[Attachment 50959 - Click to enlarge]

[Attachment 50960 - Click to enlarge]

Followed by a gradual drop toward normal. Here is field 599.


[Attachment 50961 - Click to enlarge]

Other flashes later in the video produce similar AGC effects.

The Avisynth script for the screenshots is just:

[source filter]
AssumeTFF().SeparateFields()
VideoScope("both",false,"Y","Y","Y")

Questions:

1. What's so special about those two or three flashes that hiccup the WFM, while the rest of them don't?
2. Is this expected behavior for a waveform monitor fed unstabilized VHS?
3. Why is the capture card showing an AGC effect when it's in "manual gain" mode?

[lordsmurf]

Nothing to add, just reading with interest.

[Eric-jan]

@vaporeon800: what's the goal of your post ?

[petet]

First, thank you for posting a nice picture and video of a wave form monitor (I don't have one) and helped me solidify in my head that the horizontal blanking should in fact be below 7.5 IRE, at 0 IRE, for NTSC (I kept wondering how far the 'setup' went into the blanking areas and whether I could trust my pedestal option). Also, finally I have a picture of the thing I've been wondering about but can't google quite right - what is that text in the VBI called? Makes me want to stitch together multiple captures to save that (in addition to normal captions, teletext, etc, that my capture card can already decode).

What's so special about those two or three flashes that hiccup the WFM, while the rest of them don't?
My guess is they might be 'special' in that they are a particular type of transition used in splicing these cuts together? A hard pure max white flash seems to throw off WFMs ability to keep sync? Versus normal explosion or less pure flashes.

Is this expected behavior for a waveform monitor fed unstabilized VHS?
Probably depends on the model and material... I would think they should be more robust, but I'm not much use here.

Why is the capture card showing an AGC effect when it's in "manual gain" mode?
VHS has internal AGCs for recording and playback right? Also, I know from tuning the ADV7842, there's two 'AGC' sort of registers. One based on the sync that you can control and make manual or automatic and another based on the black level: 'Peak black overrides and increases the gain of the luma AGC if the input signal becomes lower than a set threshold. Peak black is not user controllable.'

[Brad]

Also, finally I have a picture of the thing I've been wondering about but can't google quite right - what is that text in the VBI called?

In the past I showed a more interesting example: CKVU-TV added their call sign and city. There's no name for this, to my knowledge. It's just exploiting the flexibility of VBI to place low-resolution human-readable text into the VBI, along with Vertical Interval Test Signals like colorbars on the lines below.

Why is the capture card showing an AGC effect when it's in "manual gain" mode?
VHS has internal AGCs for recording and playback right?

Definitely for recording, but I can't find any references to suggest that any VCR has AGC for playback. Poking around in service manuals hasn't suggested the presence of PB AGCs, though I can barely make sense of a lot of the service manuals anyway.

In any event, it's a moot point because I managed to trigger the same effect (WFM sync loss + card AGC effect) by playing a certain bright flashing scene on a DVD. But the player used is DMR-ES25, which has "Lighter" (NTSC USA, 7.5 IRE) and "Darker" (NTSC Japan, 0 IRE) black output levels. Only "Darker" causes the WFM to glitch out.

I know from tuning the ADV7842, there's two 'AGC' sort of registers. One based on the sync that you can control and make manual or automatic and another based on the black level: 'Peak black overrides and increases the gain of the luma AGC if the input signal becomes lower than a set threshold. Peak black is not user controllable.'

I think this probably explains it. Thank you. When AGC is enabled on this card, it offers two additional on/off options: White Peak Control and Color Peak Control. Nowhere is there a Peak Black Control shown, so it's probably always operating in the background, as you suggest.

[petet]

The only reason I considered a playback AGC is from Dave Ronan's "Practical VCR Repair" page 450, which showed up in Google Books while I was looking for best practices for the use of AGCs with tapes. If they thought it worthwhile to have one for VCR recording, potentially from another VHS/VCR, would it not be worthwhile to use one on in a capture card? Which is more stable or faithful to the signal if the syncs and color bursts aren't perfect? Right now I'm going with Analog devices recommendation for a fixed gain on tape captures and automatic gain for other more stable sources.

Imagine if all tapes had these test signals embedded in their VBI - then you could correct for tape and player particulars.

Right now I'm fiddling with with the ADV7842 to get it to get it to display the VBI - it has tons of fun automatic decoders for all the various data that could be in there and even more options to hide/mask/sanitize it so it doesn't make it into the displayed picture. Going to have to stitch as both it and the display chip it is paired with is designed around 480 display lines. Very much want to jump into the raw RF capture game, but this chip already does 8x oversampling at 12 bit... if I can just finish tweaking it.

[petet]

There we go! Got my 21 lines I was looking for. Turns out there is something called "Vertical interval reference" or "vertical interval test signals" which does exactly what I was thinking about last post. EDIT: sorry Brad, you literally just told me this.

Some good broadcast EP capture. Outside of some of the encoded material in there, looks like maybe some VIR/VIT? EDIT: yup, looks like the VIR/VIT is roughly at 70-50-7.5 IRE. Not quite, but its trying!


[Attachment 52683 - Click to enlarge]

Some macrovision just for fun and some text 'BHCP9'


[Attachment 52684 - Click to enlarge]

Now time to experiment with stitching and potentially leveraging VIR/VIT...

[Pinto007]

SAA713x capture

https://forum.videohelp.com/threads/388472-Does-Grex-really-work-on-VHS-MV#post2515487

[lordsmurf]

[Attachment 52683 - Click to enlarge]

[Attachment 52684 - Click to enlarge]

I actually find the soup commercial more interesting than a SW movie for the 1000th time.