Eliminating head switching noise?
 

So there's a feature on my JVC HR-S5900U named "Video Stabilizer". By activating it, it essentially allowed for me to finally get rid of the head switching noise so that it doesn't stick out like a sore thumb, especially in dark scenes. There's one problem however. The picture seems to jitter between fields? It has nothing to with Virtualdub, the Startech USB3HDCAP, or my Leitch x75, as the jitter still occurs when I bypass all of it and go directly to a TV. How might one overcome this issue. Am I just going to have to resort to masking the last four lines? What's the best way to eliminate Head Switching Noise?

This is the screen to enable the feature on the JVC HR-S5900U
Attachment 11315

This is what it looks like with "Video Stabilizer" OFF
Attachment 11318

This is what it looks like with "Video Stabilizer" ON
Attachment 11319

PS: Please forgive me, this is my first post, so hopefully everything is comprehensive and I didn't screw up on anything. Also one more thing, upon further inspection you can notice where the two white dots at the bottom seem to have the same behavior as the two lines that rapidly change length when Video Stabilizer is on. Might I be able to pull off the same Video Stabilizer effect if I just write some software to mask the rest of the line where the white dots are?

From the samples the noise bar on the bottom is not that bad with video stabilizer off, I would just leave it off, The jitter could be due to a faulty ADC in the frame buffer or just low quality one, Get a VCR with line TBC for better results.

Thank you for your reply. I will keep it off from now on.

Avoiding head-switching noise is one of my crazy perfectionist pursuits, but even I'm not much bothered by the 4 lines in your sample.

If you want to truly eliminate it at the source, you can buy a Panasonic S-VHS and adjust it. But doing this can potentially introduce vertical jitter and (I think) introduce 30Hz head-switching noise into Hi-Fi audio. https://forum.videohelp.com/threads/...ertain-VCRs%29

The Video Stabilizer function itself just masks it, so I don't see why that's an issue. Aren't you going to do some post-processing of the capture anyway? Most of the time with your "On" capture you're getting ~85% of the first HSW line and only ~30% of the one below it, with the bottom two completely masked.

I'm surprised that latreche34 didn't suggest dropping the StarTech in favour of using the Leitch's SDI output to an SDI capture device like he uses. This would allow you to capture 720x486. The number of active lines will vary. This has always been extremely inconsistent with NTSC sources. But it does give you more room to play around with cropping/masking.

Note: if these are newer TV recordings, this won't help. If my own Canadian analog cable TV tapes are anything to go by, many broadcasts from the 2000s onward will be limited to 480 active lines.

These captures were Mitsubishi HS-HD2000U <YC> Canopus NHX-E2 (aka Canopus NX Express PCIe). Note that I cropped a different number of full black lines from the top for each of these shots, demonstrating the varying start of active video. Also note some color distortion at the top of the frame, and inconsistent luma levels (AGC issues at frame top?).
Attachment 11320
Attachment 11321
Attachment 11322

Home Alone (Canadian VHS) FernGully trailer. The Mitsu plays this tape with a lower switching point, so some frames can be cropped to noiseless full frame 720x480. But in this trailer, the number of active lines literally changes from shot-to-shot. Here is one frame and then the very next frame. The first has a full black line & a half-line. The second is full-frame.
Attachment 11323
Attachment 11324

Your JVC's Video Stabilizer doesn't cause jitter?

Interesting, I wasn't aware of the existence of "486p" capturing. I just need to find a USB friendly lossless or uncompressed SDI capture device since my captures are done on a Thinkpad T410 Laptop.

The majority of my tapes are either broadcast masters or home recordings from a video camera.

Using the Video Stabilizer on the JVC causes the jitter.

PS: Seeing the last two attachments are absolutely amazing, with no head switching noise to be seen. It really gets me interested in SDI recorders even more.

486i is what we want.

http://www.digitalfaq.com/forum/vide...html#post42279
http://www.digitalfaq.com/forum/vide...html#post59574
Previously he used a Magewell: https://forum.videohelp.com/threads/...Capture-Device

There are probably cheaper SDI-to-USB devices that work just as well. But I encountered an Osprey SDI capture device that altered levels instead of just passing through the data, so I'm wary of recommending anything that hasn't been tested and verified by advanced users.

Indeed; for mine too. The question was directed at latreche34.

It causes jitter on mine as well.

I think what it does is use something like this chip function (that's the syscon chip the newer JVCs use) and inserts a vertical sync signal at an approximate position based on the head switch point if it thinks it's missing or noisy. (A bit like how VCRs generally achieve a picture in ffwd/rwd and pause) As it has to guess a bit on where the vsync signal is it will inevitably not be perfectly in sync with the video signal which will result in the picture jumping a bit if it's a bit too early or late compared to where it should be.

On an older VCR with manual head switch adjustment you can move the head switch point down to the blanking area but that can easily mess up the vertical sync signals or cause trouble for a TBC or capture device and cause jitter as well.

This is not a line TBC, it is some sort of frame buffer used on low end JVC VCR's, I don't have a VCR with video stabilizer, I have a LTBC instead.

Indeed the 6 lines comes in handy for getting rid of that noise, I always wondered why capture devices don't capture extra pixels on all sides so when cropped you end up with a legal resolution of 720x480 for NTSC and 720x576 for PAL/SECAM that contains pure video. I think I brought this up over at videohelp too.

I haven't had luck with Magewell, I've tried 3 devices, 2 silvers and one black and got black screen on all of them, The only one worked is the BM UltraStudio and they are not cheap, had a chance to snatch one from craigslist for around $100. Keep in mind if you are going the SDI lossless route a USB 3.0 is a must. If you have a desktop you can get a SDI/PCIe for like $10 and you don't have to worry about the USB 3.0 requirement.

We're talking about vertical jitter / bounce / field hopping / etc. terminology. I see that indeed your HR-S7600AM doesn't list Video Stabilizer in the manual. The HR-S7600U does, but I'm not familiar with whether that model is even the equivalent. I think it isn't?

It certainly has an extra board for PAL, The US model states:


The other difference I can notice is B.E.S.T (Biconditional Equalised Signal Tracking) is called Active Video Calibration in the US model.
Both seem to have the Digital R3 feature for luma edge correction to enhance details, I turn it off for certain tapes.

So I spent some time trying to find some info on exactly what recorder can recognize and emit a 720x486 59.94i data stream for a piece of software, hopefully for VirtualDub, to record.

According this FAQ PDF for the Video Editing Software EDIUS 7 on Blackmagic Support, absolutely all of the recorders listed here have support for 720x486 59.94i support.
Attachment 11342

Awesome. My question now is, will one of these work in VirtualDub, or at least any of the variants, such as VirtualDubMod, or what I always use for capturing, VirtualDub2?

Upon further research... It was confirmed on the VirtualDub2 forum, that a Blackmagic Decklink was in fact able to communicate with VirtualDub2, which by presumption means all of the Blackmagic interfaces listed in the EDIUS FAQ PDF would be able to interface with the program. You may or may not have to hunt for a specific Blackmagic Driver version if you plan to get 10-bit recordings working. Though since this was from a post made in 2018, It may have already been resolved.
https://sourceforge.net/p/vdfiltermod/tickets/84/

In summary, I have not tested any of these devices myself to 100% confirm they work in VirtualDub. However, through the resources listed above, So long as you stick with the recorders listed in the PDF, you can get 720x486 59.94i 8-bit YUV recordings through VirtualDub, giving you the ability to potentially eliminate head switching noise when cropping to 720x480. Now to find one at a reasonable price, now that's going to be tricky, lol.

The capture device is the one that generates the 486 lines, So if you are doing this as analog>SDI>USB3 or analog>HDMI>USB3 (stay away from this HDMI workflow) or any similar PCIe setup the USB or PCIe device or card has nothing to do with generating the 486 lines, it's the first capture device.

You're absolutely right and something I overlooked. The process from analog to sdi does need to maintain the full 486 lines. With that said I was told a LEITCH X75 would in fact be capable of this 486 lines, so the goal for me was to find a device that would accept the signal from it.