Advice sought re: Horizontal black bars across bottom of image (VHS)

[tj7]

I’m converting yet more old VHS tapes, using a Canopus ADVC 100 and a Sony SLV E1000. Between, I am using a Panasonic ES10 in pass through mode which does a wonderful job with everything else I convert.

Sony SLV E1000 > Panasonic ES10 > Canopus ADVC 100 > FCP X

I am using FCPX on a Mac for the process. Here's what I am seeking advice about:

Please see the attached images.

The black bars at the bottom are something I have not seen before.
They do not appear on any other tapes I have converted.
The 2 tapes they do appear on are from the same source.
These tapes appear fairly badly damaged in terms of how the image looks when I bypass the ADVC unit – ie constant picture roll.
Tracking via the Sony E1000 is already adjusted to optimal settings.


Does anyone have any advice as to how I might rectify the problem, or am I stuck with these black bars?

Thanks!

[jagabo]

NTSC tapes captured in PAL mode? Ie, the capture device is expecting 576 lines but only getting 480. So the bottom 96 scan lines are junk.

[smrpix]

--double post

[smrpix]

Based on the head switching being so high in the frame, another possibility: hard matted 1.66:1 (or 14:9) and one of the devices is confusing the matte for horizontal blanking.

[tj7]

NTSC tapes captured in PAL mode? Ie, the capture device is expecting 576 lines but only getting 480. So the bottom 96 scan lines are junk.

Ah! I will investigate this when I get home! Good idea to check. Thanks.

[tj7]

Based on the head switching being so high in the frame, another possibility: hard matted 1.66:1 (or 14:9) and one of the devices is confusing the matte for horizontal blanking.

Ok.. I know about aspect ratios, but have no idea what "hard matted" or "horizontal blanking"... whatever these are, if it is this problem, is there a workaround / solution...? Thanks!

[jagabo]

Hard matte wouldn't have head switching noise at the boundary of the active picture and matte. It's possible the tapes are PAL but incorrectly recorded from an NTSC source -- ie the error is already on the tape so it can't be corrected by setting the capture device to NTSC.

If you can't fix the problem by changing capture settings (try bypassing the DVD recorder too) then you'll have to fix it by cropping away the bottom of the frame. You'll probably still have jerky motion because of the mismatched frame rates.

[smrpix]

Hard matte wouldn't have head switching noise at the boundary of the active picture and matte.

Good point.

[tj7]

Ok, so looking at more footage, it's basically a slightly flattened (horizontally) full frame 4:3 image. Which is fine.
I don't want to artificially stretch out the image in FCP X, I'd actually rather wind up with the slightly flattened image as my full frame as it's somewhat approaching 16:9 which is better for watching on screens these days.

So what is the best way to tell FCP to not simply crop the bottom bars out, but to output the final, cropped video to 16:9 or as close to it as the captured video allows?

All I know how to do is crop and reposition - what else do I need to do in order to output to 16:9?

[smrpix]

In a 16:9 timeline you can use the transform controls to fit the frame as you desire:
https://support.apple.com/kb/PH12604?locale=en_US&viewlocale=en_US (google it if you are region blocked from this US page)

Crudely measured in photoshop from your reduced stills this does look like 14:9 (which is normally presented in 4:3 with letterboxing.) If jagabo is right (he usually is) and you have an NTSC-PAL issue, you'll likely have motion and blending artifacts as well.

[tj7]

In a 16:9 timeline you can use the transform controls to fit the frame as you desire:
https://support.apple.com/kb/PH12604?locale=en_US&viewlocale=en_US (google it if you are region blocked from this US page)

Crudely measured in photoshop from your reduced stills this does look like 14:9 (which is normally presented in 4:3 with letterboxing.) If jagabo is right (he usually is) and you have an NTSC-PAL issue, you'll likely have motion and blending artifacts as well.

Thanks, i'll have a look.
No, there are no obvious motion or blending artifacts....

[tj7]

Someone replied to my thread at FCP X forums stating: "To put it into a 16x9 raster without pillarboxes of some kind REQUIRES scaling. No alternative. You just don't have enough horizontal pixels (or vertical - if that hash at the bottom of the frame is some weird capture of the vertical interval)". If this is the case, is the transform function the way I need to do this still...?

[jagabo]

If you're making a PAL DVD you must have a 720x576, 704x576, 352x576, or 352x288 frame. So if you crop away the junk at the bottom of the frame you have to scale or pad what's left to one of those sizes. If you were making an NTSC DVD you could just crop away the bottom leaving a 480 line image.

But what you have is obviously a 4:3 DAR video. If you're mixing it with 16:9 video you should pillarbox it.

[manono]

I'm with jagabo - I think you have the Canopus box set up incorrectly. If this is a PAL tape and VCR and DVD recorder, make sure your Canopus is also set up for PAL.

[tj7]

I'm with jagabo - I think you have the Canopus box set up incorrectly. If this is a PAL tape and VCR and DVD recorder, make sure your Canopus is also set up for PAL.

Makes no difference when I flip the switch - except that the image falls to the bottom of the screen instead of at the top. So that isn't it, unfortunately....

[tj7]

Ok, this is going to sound really dumb, buuuuut:
When I do *any* kind of scaling (i.e. to enlarge the image to fit 4:3 or 16:9 etc) or simply move the image to the centre of the frame so I at least have a centred image, even at the "master file" level of export, I am getting softening of the image. Expect this with scaling but not the image moving as it is not resized when I do that.
Anyway. When I use Mac OS X'S zoom in function when I play the original raw, unprocessed .mov file (i.e. using control+scroll wheel to literally enlarge the view in finder) there is absolutely no image softening. To create the proportions I want with minimal image distortion, it looks like FCP loses to the finder's zoom in capability. I am now looking to work out how to create a new video file in real time whilst zoomed in, and i'll simply resync with the source audio.
This sounds silly but that's what I'm seeing.
Why is it that FCP isn't as capable of producing as high a quality image when scaling up, as compared with zooming into the image via finder in Mac OS?
Or if I am missing something, what is it?

[smrpix]

In FCPX you can set your viewer preference to better quality or better performance. Make sure you are using the better quality selection to judge accurately. This does not affect the final output, just responsiveness vs clarity while editing.

[tj7]

In FCPX you can set your viewer preference to better quality or better performance. Make sure you are using the better quality selection to judge accurately. This does not affect the final output, just responsiveness vs clarity while editing.

Yep, I'm aware of that but I comparing the final output file, not what is in the viewer.
The output file, even when export settings are at max quality with huge file size, is still softer than I am seeing when I view zoomed into VLC player or Quicktime via OSX finder - and so far a screen capture vid recording app is doing better than FCP in terms of image quality!