Missing pixels with s-video vs. firewire capture Sony DCR-TRV460

[Bogilein]

There is no reason to blame the ATI600 for the black edges of your S-Video capture. Nor is the GVUSB2 responsible for the black edges in Alwyn's example. Both capture devices can record the full width of 720 via the S-Video input.


[Attachment 77113 - Click to enlarge]
(The attached screenshot is just to show this and has not been adjusted for direct comparisons between ATI600 & GVUSB).

The S-Video output of your Sony DCR-TRV460 is responsible for adding the black borders.

[lollo]

Thanks Bogilein, excellent post.

BTW, if you can post a comparison between the ATI USB600 and thr IOData GV-USB2 captures using your test tapes it will be very useful...

[Sharc]

There is no reason to blame the ATI600 for the black edges of your S-Video capture. Nor is the GVUSB2 responsible for the black edges in Alwyn's example. Both capture devices can record the full width of 720 via the S-Video input.

@Alwyn: Are you using a DVD recorder in passthrough? Some limit the width of the active picture, adding borders (masking) widths >~704.

[theseeker2]

I captured it with print screen, but I also just tried saving the frame in VLC which is what I use and it looked the same.

Can you share a small DV-AVI clip with this particular shot that you compared above?

Sure, see attached.

[Alwyn]

Are you using a DVD recorder in passthrough?

No.

[Bogilein]

Are you using a DVD recorder in passthrough?

No.

The TBC of the playback device can also have an influence on the result


[Attachment 77126 - Click to enlarge]

[lollo]

The TBC of the playback device can also have an influence on the result

Nice report.

[dellsam34]

The TBC of the playback device can also have an influence on the result


[Attachment 77126 - Click to enlarge]

Where is the blanking area in this screenshot, Was this recorded on tape?

[Bogilein]

Both pictures have the same size 720x576, the right image (TBC On) has the black bars on the right and left side.

The test patterns were recorded from a videotape. If you look closely you can see the head switching noise at the bottom.

[Skiller]

It's absolutely possible to output, record and finally playback a full 53.333 µs (equals 720 pixels) worth of active video without any blanking at the sides with VHS and others, but it's technically slightly off standard. Most DACs blank the sides, as is correct to do.

[dellsam34]

I don't know, never being able to achieve it.

[theseeker2]

I learned so much from all of you on this thread, thank you so much!

The consensus seems to be that the ATI USB 600 is not to blame, but it is indeed the S-Video output from the DCR-TRV460. I do wonder if it's really less of a case of things being "missing" over S-Video or more of a case of the Firewire transfer showing the full 53.333 µs (equals 720 pixels). I did a quick test on an old Sony CCD-TRV29 that only has composite video output (no S-Video or Firewire) with a different tape, and then ran that same tape through the Firewire, and it looked like there were a few extra pixels on the left and right in the Firewire transfer that were not in the composite transfer.

Thanks especially to Bogilein for the comprehensive test, that is much appreciated.

I am going to try to do one last test using an old gaming console as Skiller suggested earlier just to make sure my ATI USB 600 isn't doing something goofy

[dellsam34]

You wouldn't know for sure until you use a different capture device with the same camcorder, We've seen capture devices do this before, tons of youtube videos showcasing the problem. I remember this funny british guy who compared some devices.

[Bogilein]

Since the user dellsam34 doubted the result (he has since deleted the message).

Here is the procedure to make the whole thing as transparent as possible or to tell me where I made mistakes:

Test patterns were played from a DVD with the DVD-recorder Pioneer DVR-630H and recorded on a VHS video cassette (Kodak E240) with a VHS recorder (I have too many and can't remember which one...).
Then the cassette was played once with TBC on and once with TBC off with the SVHS video recorder Philips VR1500 (which based on a JVC HR-S8600) and captured via the analog Y/C output with the capture card Canopus NX.
The captured files were then opened in Virtual Dub and inserted into the image viewer Irfanview using "Copy Source frame to clipboard" function from Virtual Dub. The same with the second image which was added to the side.
The only image manipulation I did, I labeled the images and saved them as jpg. Nothing was changed to the image size.

Today I made 3 more captures, once with the Philips VR1500, once with the JVC HR-S7611 and once with the Blaupunkt RTV950 (which is identical to the Panasonic FS200), where I switched the TBC on and off during recording. Captured via the analog Y/C output of the recorder this time with the Hauppauge USB2 Live capture card with default settings.

PhilipsVR1500-USB2Live.avi
JVC7611-USB2Live.avi
RTV950-USB2Live.avi

Enclosed also my test patterns as zip file, unzip the file and you can burn the DVD Folders (Video_TS/Audio_TS) e.g. with imgBurn on DVD to perform own tests.

Capture TestFile DVD.zip

Enclosed also the test patterns as mpg file in case someone wants to compare something.

PALCapture Test File.mpg

[dellsam34]

What message I have deleted?

[Bogilein]

the part where you ask where I got the screenshots from, because they must have been manipulated

[dellsam34]

In this thread? I really don't remember. Anyway as you can see from the link I posted capture devices do crop a little bit, which leads me to believe it's his capture device, anything is possible.

[Bogilein]

Anyway as you can see from the link I posted capture devices do crop a little bit, which leads me to believe it's his capture device, anything is possible.

This is exactly where the mistake lies.

I can confirm that the ATI600 does not cut anything, but very few people think that it happens in the playback device. A TBC or just the analog output can be responsible for this. An external device such as a TBC or DVD recorder can also be responsible.
I think everyone should know what their own capture hardware can or cannot do and not draw the wrong conclusions.

With the linked test it would be interesting to know if he has considered this.

Why don't you do a test yourself with your Pinnacle 500 USB?
Play the test patterns with a DVD player and see if the Pinnacle displays the full width of 720. If not, it is due to your player. The Pinnacle can record the full width of 720 without cutting anything off. If your player can do this, use your JVC 7600 as a passthrough device and see if it outputs the full width without black edges and finally check the whole thing with your Brighteye 75 and Snell & Wilcox units.

[Sharc]

Here is what I am getting for a DVD recorder in passthrough:

Original Picture = SMPTE colorbar 720x576. 7 bars ~103 pixels wide each, all with same width, burnt on DVD

The DVD is then inserted into the DVD recorder Panasonic E55 and played back.

The Oscillogram shows the Y (luma+sync wire) coming out of the S-Video port of the E55. The scanline duration of all the bars is 52us.
The first bar (grey on the left) and the last bar (blue on the right) are sligtly cropped, means shorter than the center bars which have a duration of ~7.7us each: The grey bar is ~1us cropped, the blue bar is ~0.6us cropped.

Hence the S-video out of the DVD recorder delivers a slightly cropped picture to the capture device.

The dongle (GV USB2) captures the reduced width (as it gets it from the E55 S-video port) and pads it to 720 (53.333us)


[Attachment 77207 - Click to enlarge]

[lollo]

Hence the S-video out of the DVD recorder delivers a slightly cropped picture to the capture device.

Clear with your experiment, thanks!

[Bogilein]

The result of the Panasonic DVD recorder does not surprise me.

All Panasonics that I have tested output a width of 704 + black borders via the Y/C output.
Regardless of whether the DVD recorder works internally with 704x576 or 720x576.
With the Panasonic recorders that work internally with 720x576, however, you can capture the full width of 720 via the HDMI output without black borders.

A general statement that DVD recorders only deliver a width of 704+black borders via the Y/C output should not be derived from this.

Here is a small selection of what various devices reproduce via the Y/C output.


[Attachment 77208 - Click to enlarge]

[Attachment 77209 - Click to enlarge]

[Attachment 77210 - Click to enlarge]

No one should jump to conclusions without knowing what the used devices can and cannot do.

[Skiller]

Hence the S-video out of the DVD recorder delivers a slightly cropped picture to the capture device.

Indeed, but with a recording made in the analog days, there isn't anything there that would be lost by this. Unless the picture is off center, which is actually one of the main reasons the ITU standard settled for the longer 53.333 µs window – it includes some of the front and back porch of the analog video signal, rather than just the active picture area. This allows for some headroom.

As demonstrated by Bogilein, it is only with some DVD-players (and some DV camcorders) that we can even output such a digital-originated frame whose headroom isn't black to an analog video signal (and being slightly off-standard with it). Many DACs simply blank these areas when converting the digital 720 px frame to analog.

[Sharc]

More (or similar) important IMO than a few pixel cropping is that all your examples came with the correct Rec.601 Sampling Aspect Ratio (approximated by 12/11) for sampled 4:3 PAL video signals, independent of the DVD recorder model.

[Skiller]

Yes, this is because the ITU standard (more precisely: sampling rate) is respected by all devices Bogilein used for conversion from digital to analog and then analog to digital. I would not have expected anything else. In fact I have never ever seen it happening otherwise (ignoring user error when capturing with certain hardware).