Video8/Hi8 playback/capture options

[HDClown]

What's the best option for me to use in terms of playback/capture for Videp8 and Hi8 content? I have a Sony DCR-TRV120 with built in DNR and TBC. I can use in 1 of two pays:

1) Connected via FireWire, which will end up capturing the source into a DV file
2) Connected via RCA L/R and S-Video connected to my ATI TV Wonder 750 USB capturing HuffYUV

I've done some capture using #1 from a Hi8 source and have an interest result. Here's the MediaInfo for the video section

Video
ID                               : 0
Format                           : Digital Video
Codec ID                         : dvsd
Codec ID/Hint                    : Sony
Duration                         : 8mn 22s
Bit rate mode                    : Constant
Bit rate                         : 24.4 Mbps
Width                            : 720 pixels
Height                           : 480 pixels
Display aspect ratio             : 3:2
Original display aspect ratio    : 4:3
Frame rate mode                  : Constant
Frame rate                       : 29.970 fps
Standard                         : NTSC
Resolution                       : 8 bits
Colorimetry                      : 4:1:1
Scan type                        : Interlaced
Bits/(Pixel*Frame)               : 2.360
Stream size                      : 1.68 GiB (95%)

The DV file is 720x480 with a 3:2 display ratio, but it's tagged as a 4:3 original aspect ratio, and WMP/QuickTime will size the window as 640x480 for playback. VLC will size as 720x480. QuickTime also shows "720x480 (640x480)" in the video inspector.

I capture my VHS using method #1 (but from a S-VHS deck source instead of the camcorder). The MediaInfo on those files doesn't have 2 different aspect ratios, and the file always shows as 720x480 sized, regardless of what player I use.

I want to capture in whatever method means the best quality capture. I would think DV would be better in this case, be cause it eliminations extra A/D conversions, but I need someone to confirm for me.

I'm also a little confused on the different aspect ratio/resolution things that I am seeing when I playback the .AVI creaetd on DV import. Could someone explain this? My assumption here is that since the original content was Hi8, which is "full NTSC resolution", or essentialliy 640x480i, but the import through a D8 camcorder and D8 uses 720x480i, this is why there is the difference in the AVI file with display ratios and resolutions.

[lordsmurf]

There's actually a similar conversation here: http://www.digitalFAQ.com/forum/showthread.php/advice-preserve-svhs-2254.html
This may be good reading for you, too: http://www.digitalFAQ.com/forum/showthread.php/aspect-ratios-editing-2128.html

For best quality, I prefer analog capture of 8mm and Hi8, to HuffYUV or even straight to MPEG-2 (15Mbps for editing, or less than 9Mbps for DVD-Video).

[HDClown]

If you find MPEG-2 acceptable at 15Mbps, what do you have against DV at 24Mbps, aside from the colorspace difference? What do you recommend as the MPEG-2 encoder for import?

I will have to do a side-by-side comparison of my equipment options, but if there is no visible difference between bringing this 8mm/Hi8 in as DV vs HuffYUV, I'm not sure I see the benefit to all the extra HDD space. I've been passing my W808 S-VHS VCR through a Panasonic DMR-ES10. Now that I know the TRV120 has built in TBC, I can try pass-through with that, but the device only has 1 s-video port, so I can input s-video/output composite, or vice-versa. I suspect it would be ideal to stick with s-video in/out like I can with the W808/ES10. Based on the source material I've looked at currently, none of it is in bad need of time based correction.

I'm not willing to purchase any additional equipment at this point, so this is what I have to work with. I was originally fine with HuffYUV for all my VHS, but this is going to entail a minimum of 6TB to capture everything I have, and that's space I simply do not have, and nor am I in a financial situation to acquire it right now. So I need to find a good alternative.

Where I'm confused on PAR/DAR/AR and SAR is with this DV material I've been working with an re-encoding it. To simplify my life, I've been using Handbrake and the recommend high profile preset. This defaults to anamorphic loose sizing and ends up spitting out 712x480. I can make it none or strict and keep it at 720x480 like the DV. The MP4's made from Handbrake only show as 720x480 and display this way by WMP/QT/VLC, which is different than the DV file which shows in WMP/QT at 640x80 but in VLC as 720x480. All I ultimately want to know is if, in the process of encoding, I'm doing something to distort the actual image as it was originally shot and intended to be viewed.

[spapakons]

Hello!

I know it's an old thread, but I want to help whoever reads it. I would capture in analog using S-Video connection for best image quality and either a lossless codec or very little compression. Of course I would then edit the video and save as DVD or video file with more compression to keep file size reasonable. I would capture at the maximum resolution supported by the card, in my case 768x576 25fps for PAL, so I have less waste when cropping the black borders. I would then deinterlace it to improve quality and resize to 720x576 for DVD. Today's graphics cards offer hardware acceleration, so I can capture to HEVC directly using a very high bitrate to appear almost lossless without wasting disk space. For standard definition video I use NVEnc HEVC at 2000K bitrate or more. This is better than 10000K MPEG 2, to get an idea.

[lollo]

I would capture in analog using S-Video connection for best image quality and either a lossless codec or very little compression.

Capture a Y/C time base corrected signal 4:2:2, lossless, interlaced, 720x576.

Of course I would then edit the video and save as DVD or video file with more compression to keep file size reasonable. I would capture at the maximum resolution supported by the card, in my case 768x576 25fps for PAL

Analog PAL SD signal is framed to 704(702)x576, so capturing at 768x576 is a useless degradation of quality (and probably a distortion) by the internal resizing performed by the card.

so I have less waste when cropping the black borders.

Black borders have nothing to do with the resolution, and if you have less black borders is because distortion of the frame.

I would then deinterlace it to improve quality and resize to 720x576 for DVD.

A resizing, even in software, is another potential loss of quality. In addition to the one present because capturing at 768x576. Capture at 720x576 since the beginning.

Today's graphics cards offer hardware acceleration, so I can capture to HEVC directly using a very high bitrate to appear almost lossless without wasting disk space. For standard definition video I use NVEnc HEVC at 2000K bitrate or more. This is better than 10000K MPEG 2, to get an idea.

There are zero reasons today to do not capture lossless, specially if you plan a restoration.

[Alwyn]

Video 8 tape PAL DV capture verses Analogue Lossless capture:

https://www.youtube.com/watch?v=Ap59r45OmOw

[aramkolt]

That's an interesting comparison - I really can't see much difference outside of the levels that they were recorded at, but it could be that YouTube is blurring some of the detail away. Keep in mind that's also PAL footage which has a fair bit more vertical resolution than NTSC, so those captures tend to look better than NTSC if that's what you're used to looking at.

I do still think there's a place for higher bitrate MPEG2 captures (8+Mb/s) in modern times for average consumers that don't want to deal with extremely large file sizes or expense, or possibly those that don't want to involve computers (via DVD recorders). You can definitely do better with lossless capture, but you can also do muuuuuch worse with quite commonly used methods like the elgato or cheap RCA2HDMI adapters with HDMI recorders.

This video is worth a watch if you haven't considered MPEG2 and does compare against lossless capture: https://www.youtube.com/watch?v=j4sIgQWNHWA

I again think some of the similarity could be due to some detail getting blurred away by YouTube compression.

Seems like the input levels are a bit off on the MPEG2 capture which look to be adjustable in menus, or you could just use an analog proc amp in the chain also. The other takeaway is that going above 9Mb/s may not actually result in a significant visual difference when it comes to MPEG2.

[Alwyn]

I do still think there's a place for higher bitrate MPEG2 captures (8+Mb/s) in modern times for average consumers that don't want to deal with extremely large file sizes or expense

What are the exact options for MPEG 2 hardware though, and how much, what OS.

I think the paranoia about YT destruction is a bit of a furphy. Pump it up with plenty of bitrate and YT videos are fine, including when doing comparisons than are 1152 or above.

[aramkolt]

I think the problem with Youtube is more that it starts to round similar colors to each other creating essentially macro blocking. I've heard this isn't as much of an issue with VP9 formatted uploads, but VP9 is something I'm not too familiar with.

For the average user, the easiest to use MPEG2 solution would just be a regular DVD recorder and burning onto either DVD-R or RW and then ripping discs. Definitely has some extra steps, but the results can be nice without much complexity and if you kind of get a bonus physical backup copy. If you had 50+ tapes, I probably wouldn't go that route, but if you only needed it for a few, should work well.

I read a lot of Jwillis84's older posts who has tested countless devices and methods and he really seems to like the Toshiba RD-XS series of DVD/HDD recorder machines as an option. They can be modified to accept SATA drives and this was sort of my implementation so that you could just remove the drive easily and transfer the MPEG2 files off whenever you want using ISObuster. The RD-XS55 actually had a feature called "net dubbing" which would allow you to pull files off of the internal hard drive over ethernet, but I'm sure that is rather slow compared to directly hooking up the drive, but something that no other DVD/HDD recorder had apparently.





As far as other standalone devices that might be better supported by modern systems, I've heard good things about the Dell Angel and apparently it works with any modern version of PowerDirector (though bitrate is fixed at 8Mb/s - which should still be fine) and Nero Vision (no idea how new of a windows OS that can be). The Dell Angel uses a similar or same NEC chip as at least some of the the RD-XS series.

The Dell Angel also supposedly works super well with Snappysoft Monsoon which is XP only, but there's no way to buy a license these days as the company is long gone. Supposedly it was super easy to get around activating it according to those that used it back when it was more popular, but I don't see anything online for how you'd use what is essentially abandonware at this point.

[Alwyn]

@aramkolt, interesting, thanks.

[spapakons]

I would capture in analog using S-Video connection for best image quality and either a lossless codec or very little compression.

Capture a Y/C time base corrected signal 4:2:2, lossless, interlaced, 720x576.

Of course I would then edit the video and save as DVD or video file with more compression to keep file size reasonable. I would capture at the maximum resolution supported by the card, in my case 768x576 25fps for PAL

Analog PAL SD signal is framed to 704(702)x576, so capturing at 768x576 is a useless degradation of quality (and probably a distortion) by the internal resizing performed by the card.

so I have less waste when cropping the black borders.

Black borders have nothing to do with the resolution, and if you have less black borders is because distortion of the frame.

I would then deinterlace it to improve quality and resize to 720x576 for DVD.

A resizing, even in software, is another potential loss of quality. In addition to the one present because capturing at 768x576. Capture at 720x576 since the beginning.

Today's graphics cards offer hardware acceleration, so I can capture to HEVC directly using a very high bitrate to appear almost lossless without wasting disk space. For standard definition video I use NVEnc HEVC at 2000K bitrate or more. This is better than 10000K MPEG 2, to get an idea.

There are zero reasons today to do not capture lossless, specially if you plan a restoration.

I am going to deinterlace the video and crop the black borders anyway, so what looks better, 768x576 cropped to, say, 740x560 and resized to 720x576 or 720x576 cropped to 704x560 (or less) and resized to 720x576? Even if the difference is not perceivable, I prefer the 768x576 option, just in case. Also 2000K or 4000K in HEVC for standard definition is indistinguishable from lossless and takes much less space, common sense.

[lollo]

I am going to deinterlace the video and crop the black borders anyway, so what looks better, 768x576 cropped to, say, 740x560 and resized to 720x576 or 720x576 cropped to 704x560 (or less) and resized to 720x576?

What you do not understand is that Analog PAL SD video according to the standards should be digitized to 720x576. 768x576 is the display resoultion.

What you do not understand is that no resizing is necessary at all. Capture at 720x576 and crop to 704x576. End of the story.

Even if the difference is not perceivable, I prefer the 768x576 option, just in case.

Non sense, your preference causes degradation of the quality and non standard capture resolution.

Also 2000K or 4000K in HEVC for standard definition is indistinguishable from lossless and takes much less space, common sense.

Common sense is that is not indistinguishable, especially if you go through decoding/coding iterations because restoration.

Hello! I know it's an old thread, but I want to help whoever reads it.

You are not helping anyone here, just giving bad advices, so stop this non sense!

[spapakons]

My understanding is that PAL is an analog format with 576 viewable horizontal lines, but each line has infinite points. The 720 pixel line was decided because standard definition cannot give more detail beyond that limit and was easier to encode with then technology than 768 which is the 4:3 PAL resolution for square pixels. The respective 16:9 resolution is 1024x576 which is actually used in standard definition digital channels. Just record a video with a DVB-T receiver and confirm with Media Info utility. So stating that the image is distorted in 768x576 is wrong. It is just digitized with more samples. Compare 48000Hz vs 44100Hz in audio. That's exactly why NTSC DVD is 720x480 and not 640x480, to make it look better. Also have you made an A-B comparison of lossless video and HEVC video at very high bitrate, say 4000K? Can you tell the difference? So why not use HEVC and save valuable disk space? Besides it's for personal use, not professional. Even if it were professional you could not tell the difference. So use a very high bitrate for the capture and then save to lower bitrate for the final file. That's the common sense I was talking about.

[dellsam34]

What's your evidence that your card samples natively at 768 samples per scan line versus sampling @ 720 and resizing to 768 on the fly in software, which what capture cards do because they use chips that are limited to 720 samples by standard.

[lollo]

My understanding... So stating that the image is distorted in 768x576 is wrong.

You still do not understand; I was talking about a captured frame at 768x576. dellsam34 already gave you some hints.

Just record a video with a DVB-T receiver and confirm with Media Info utility.

Are you serious? There are none (zero) channels broadcasting at 768x576 resolution in DVB-S and DVB-T, and I am dumping dvb streams since 1999!

Also have you made an A-B comparison of lossless video and HEVC video at very high bitrate, say 4000K? Can you tell the difference?

I made a comparisons in a recent thread between lossless and x264 at CRF 17 (off-line compression of a lossless stream, so higher quality than your HEVC 4000K), and guess what? (search the topic by yourself)

So why not use HEVC and save valuable disk space? Besides it's for personal use, not professional. Even if it were professional you could not tell the difference. So use a very high bitrate for the capture and then save to lower bitrate for the final file. That's the common sense I was talking about.

Because lossless is better and because you do not stress the hardware with the risk of dropping frames. There is no common sense, just bad advices if quality is the aim.

Once more, stop your non sense!

[spapakons]

No use trying to convince you. Anyone can do some tests and judge what settings work best for themselves. And I was talking about 1024x576 resolution in digital channels, not 768x576. 1024x576 is what Handbrake reports when opening the video.

[lollo]

No use trying to convince you.

It is not a question of convincing someone, but of science and facts versus idiocies.

Anyone can do some tests and judge what settings work best for themselves.

Sure, and then nobody will capture analog video 768x576 h265 compressed.

And I was talking about 1024x576 resolution in digital channels, not 768x576. 1024x576 is what Handbrake reports when opening the video.

That's just another useless GUI, reading flags. Learn AviSynth, use it, and you will understand the real nature of the video.