ADVC-3000 compression with BlackMagic Intensity?
 

Greetings Fellow Video Enthusiasts,

My current VHS to Digital workflow includes outputting a VHS tape from S-Video on a JVC SR-VS30, to the S-Video input on the ADVC 3000 and then running that signal out through the ADVC 3000's component output to a BlackMagic Intensity Pro 4K capture card. I'm using the 525i 59.94 NTSC setting for the "Video Project Format", and Quicktime Uncompressed 8-bit YUV for the "Capture File Format." Which is arguably overkill for VHS footage, but I believe I have the hardware workflow to handle it. When I pull up the MediaInfo from the video file created I see that it is showing the file is a 4:2:2 Loseless compression, and the file size is certainly big enough to justify those parameters. However, I'm not sure if the video I'm capturing is uncompressed based off the hardware workflow or if it's just the BM Intensity Pro 4K telling the video to be uncompressed. My concern is, is the ADVC 3000 somehow converting the footage and my Intensity Pro 4K is just making the video file uncompressed because that's what I have the BM IP 4K set to or does my current workflow make sense for uncompressed? I understand the ADVC 3000 can do uncompressed 4:2:2 video via component to SDI, but would it make a difference if I went from S-Video (VCR) to S-Video (ADVC-3000) to IP4K (S-Video). Does component make a difference when capturing uncompressed vs. capturing with S-Video? Does the video signal go through the compression process when outputting anything else than DV?

Thanks.

The way you are using the ADVC-3000 device is just to convert from S-Video to component, Why are you going to the trouble of using such device for a non necessary step? It can capture the video on its own and converts it to SDI digital, from there use a SDI PCI card to stream the digital signal to hard drive.

I'm guessing it's acting as a TBC of sorts in this setup, stabilizing the signal for the blackmagic card. I would presume it would go through an A/D -> D/A step in the AVDC in that case, otherwise one would get a ton of frame drops.

Changing to S-Video would probably hardly make a noticeable difference unless the capture device or AVDC does something unexpected, or if bypasses the internals if it's hooked up that way. Component is a step up from S-Video, though it doesn't make a lot of difference on SD material, and probably less so with NTSC than PAL due to how PAL color is decoded. In either case I don't see how it would give any improvement, only make it worse. Using SDI may be the optimal route though as suggested.

That's my point, If the card is capturing the S-Video signal, applying the TBC and routing the digital frames to the serializer chip for SDI out then why go back to analog?

The reason the ADVC 3000 is in the equation at all is because I've discovered without it, the video and audio get out of sync when trying to capture directly to the BlackMagic Intensity Pro 4K, despite BM's claims of a TBC and being designed for old analog sources.

I understand S-Video is the weakest link in my cable setup, and that going from S-Video out to Component won't necessarily improve the picture any better than S-Video, but I'm trying to figure out if the ADVC 3000 is messing with the signal in any way that would be considered compression and the BM Intensity Pro 4K is just treating the signal coming in as uncompressed simply because that's what I have it set to. I'm mean in theory, by not outputting via the FireWire DV port, and going from S-Video to Component (or any other analog cable for that matter) I'm simply using the ADVC 3000 as a passthrough device, so no compression would need to take place?

We can tell you whether any compression is happening if you attach a short lossless sample.

My preference for lossless compression is Ut Video, for size and decompression speed. In VirtualDub, the appropriate setting is "UtVideo YUV422 BT.601 VCM" [FOURCC code 'uly2']. You can also use Huffyuv, which is more popular on this forum but doesn't have a straightforward modern installer, or Lagarith. You must download and install one before following the steps below.

1. Open one of your uncompressed captures in VirtualDub.
2. Find a scene with saturated colors, ideally with sharp color transitions.
3. Use the In/Out markers to select a short sample. Even 1 second should be more than enough. Definitely not more than 10 seconds.
4. Video -> Fast Recompress
5. Video -> Compression...
6. Select the lossless compressor that you chose.
7. Configure. For Ut Video & Huffyuv, Predict Median. For Ut Video, check the box Assume interlace video (even though it says not recommended). For Lagarith, make sure Mode is YUY2.
8. File -> Save as AVI...
9. On the forum: change reply to Advanced post mode; Manage Attachments button.

Do you have any samples? Wouldn't expect there to be DV compression on anything but firewire, DV compression tends to be quite noticeable on very noisy/grainy video. (There are SDI variants that support compression, used for e.g digital betacam but that's less common.)

Would avoid the "assume interlace video" setting, it's been removed entirely in the latest versions of utvideo and I've found it to cause issues for some versions of the ut video decoder from ffmpeg (used in a variety of applications.)

I wish there was a lossless codec that had a proper setting/flag for interlacing and field order, but that doesn't seem to be a big priority for developers in this day and age.

Really? I didn't think I was far out of date, and it makes my 6-hour interlaced captures considerably smaller than the progressive compression mode.

x264vfw lossless works. And has Sample Aspect Ratio flagging (also missing from typical lossless codecs).

Hm, I must have misremembered, the interlace option is still there in the latest version.

Forgot about x264/x265 lossless, would be great to use if they weren't syrupy slow.

Just did a test and MagicYUV was smallest this time. Never had that result before. It also decodes fastest -- laughably so. CPU used came out mid-2009, so over 10.5 years old. I suspect a new Intel CPU with high single-core performance wouldn't have as big a gap between Ut Video and Huffyuv.

Attachment 11259
Attachment 11260 Attachment 11262 Attachment 11261

The content is the first 2m25s of an NTSC retail VHS capture. With so many progressive frames, the progressive modes are smaller here.

0:00 VCR analog tuner noise
0:09 tape start: blackness
0:23 Columbia logo (29.97p*)
0:44 PG rating screen (still*)
0:49 trailer for "Little Secrets" (23.976p with 3:2 pulldown*)
2:21 Approved for All Audiences screen (still*)

* Playback noise would be considered interlaced since it varies on every field.

Attached is the VHS sample video file and a screengrab from the MediaInfo. This sample is from some footage that was dubbed from a Sony HXR-NX70U running in 480i mode to VHS. I can upload a clip from the video being captured directly from the camera, through the ADVC 3000 into the Intensity Pro 4K if that helps.

Are you stuck on Mac? You should at least ZIP an uncompressed capture before posting it.

That looks kinda crummy to me, but I don't see any evidence of lossy digital compression.

No one said you are capturing in DV, What I meant was that try to capture with the ADVC3000 itself using a SDI/PCIe card, unless you are on the MAC then you are out of luck.

Sorry about the file size, should have thought of that (msgohan). (latreche) The sample clip was captured with the ADVC 3000, ingested into a BM Intensity Pro 4K with the Quicktime Uncompressed 8 bit YUV setting, on a Windows PC.

You're still confused about what capture means, If the ADVC 3000 receiving analog and outputing analog then it is not capturing, the BM is the one capturing. The big point is you are using two capture cards for one job when you can only use the most qualified one. In other words the ADC inside the ADVC is way better than the ADC inside the BM because the former was designed for analog SD video, the latter was designed for analog HD video, Hope the picture is clear now.

I understand the ADVC 3000 isn't capturing, with my setup. I was just wanting to make sure that introducing the ADVC into my workflow wasn't compromising the video's quality (via compression) when capturing with the Intensity Pro 4K. The only reason the ADVC is in the equation at all is because it keeps video and audio in sync. For now, I'm going with the assumption that my uncompressed footage is indeed uncompressed. Thanks for the advice.

Then capture with it and get rid of the Intensity pro, You can sell it and get a SDI/PCIe card for $10 and keep the cash.