Questions over capturing on a basic home setup as opposed to a pro service

[Wills]

Hi all,

My I-O Data GV-USB2 Cable Adapter USB 2.0 finally arrived, and I'm using it with AmaRecTV 3.10 and I chose Huffyuv for a compressor.
It works - the quality is what I expected, given that I'm capturing footage I shot on a 1986 Sony shoulder mounted camera, and as we only had 10 tapes, I transferred footage to 4 hour VHS tapes as I went, and then reused the tapes.

I do have a couple of questions, though.
My 30 year old Aiwa HV-FX5500k VHS player is pretty basic. It has SCART out, and I'm using a composite RCA at the moment into the I/O Data capture dongle.

Is it likely that a basic (not awful Aiwa from 1995 £250 player) would have s-video (or composite) wired up its their SCART out?
My guess would be no....and if it's a no - is getting an s-video going to have that much better quality output?

I'm wondering about paying to get the transfer done - and that's my other question - given that these VHS tapes are second generation copìes of something that was pretty basic in terms of image quality to begin with - I wonder if I'd see a dramatic quality of capture shift if I paid for it to be done rather than doing it myself.....

Would a professional grade player see a measurable increase in quality....?

Any thoughts?
Thanks to any kind souls who might be able to proffer an opinion!

Cheers.

[dellsam34]

As far as I know regular VHS players do not have an active C pin on the SCART connector, Only S-VHS machines did. Whether you personally notice an improvement when playing a VHS tape in a S-VHS machine or not the advantage is beyond just an enhanced sharpness and richer colors, It's the whole package, better tape handling, better image processing (TBC, DNR, DOC...) which make the output signal very tolerable to most capture cards.

Paying to get the transfer done also means nothing, you could be paying someone who does an inferior job than what you are doing now, unless you know exactly what gear they use, how they process the files and also seen their work.

[Alwyn]

Is it likely that a basic (not awful Aiwa from 1995 £250 player) would have s-video (or composite) wired up its their SCART out?

No. In my (limited) experience, only SVHS machines have S-Video Out, as well as some DVD/VHS Combos; the Panasonics being examples (ES-35, EZ-48). There has been some conjecture about whether they actually produce true S-Video and not just Composite over the S-Video Out ports. Those Pannys though are good because they have the Panasonic Diga stabilisation system that works well in stabilising wonky/wobbly/jerky tapes.

You can test the S-Video on the Scart plug: just connect an S-Video cable to the Scart and to the GV and choose S-Video on the AmarecTV Graph 1 tab.

if it's a no - is getting an s-video going to have that much better quality output?

Once again, in my limited experience, not "much" better. Of more importance is the condition of the machine itself. A good-condition Composite VCR will be as good as a poor-condition SVHS VCR.

All in all, a conundrum!

You could send one of your tapes to Colin at Video99UK and compare his transfer with yours.

If you'd like an appraisal of your current efforts, you can attach a 15 second AVI capture here (max 500mb).

[Sharc]

.... only SVHS machines have S-Video Out, as well as some DVD/VHS Combos; .....

Some DVD/VHS combos do have an S-Video out but it is only operational for DVD playback and not for VHS playback which is composite only. Worth to check the manuals (including the footnotes).

[Alwyn]

Worth to check the manuals (including the footnotes).

Example?

[Sharc]

Worth to check the manuals (including the footnotes).

Example?

Sorry but no as I don't like the idea of bothering your Grandma
https://forum.videohelp.com/threads/418295-Bluray-Players-Does-Yours-Play-Bluray-Data-...ks#post2775104

[Alwyn]

Sharc, if you can't back up your claims with your suggested manual references, I suggest that you don't make such claims. There is ZERO mention of this issue in the EZ-48 manual.

[Sharc]

Alwyn, it was the result of earlier discussions about a particular device of which I don't remember the brand and type. I really don't find the time to scan through posts and manuals. The OP may take it as a hint.

[Brad]

.... only SVHS machines have S-Video Out, as well as some DVD/VHS Combos; .....

Some DVD/VHS combos do have an S-Video out but it is only operational for DVD playback and not for VHS playback which is composite only. Worth to check the manuals (including the footnotes).

This is common on DVD player / VHS combos. Sony example: https://www.amazon.com/Sony-SLVD370P-Progressive-Combo-Player/dp/B000ELC0GM/

Are there DVD recorder combos like this too? I feel like they would all implement the "composite separated into S-Video and YPbPr" route. I don't think I've seen a DVDR/VHS model that has the "DVD exclusive output" section at the back.

[Alwyn]

@Wills, here's a test showing a comparison between two different VCRs, Composite and S-Video, using the GV-USB2. The JVC looks to have been lightly used, based on the condition of the insides; the EZ-48 is essentially brand new.

No processing on either capture was done (apart from the "Diga" effect of the EZ-48).

[Sharc]

@Alwyn: Should the subtitle of the picture on the right read JVC5700 S-Video? Which is which?

(Btw both units produced a glitch at frames 525 ...527).

[Alwyn]

@Wills, here's another comparison: Run-of-the-mill LG Composite VCR compared to a different JVC SVHS VCR.

[Sharc]

What we compare here are more the properties of player 1 vs player 2, rather than strictly comparing Composite vs S-Video IMO. The clip has relatively low details, little high frequency content and low color saturation. The drawback of Composite materializes most with sharp transitions (high contrast edges), high details and high color saturation causing the typical Composite artifacts like dotcrawl and rainbows due to spectral overlap of luma and chroma. But yes, a good Y/C filter (usually comb filter) mitigates those effects significantly, and in this comparison little of the differences (like halos due to oversharpening) can really be attributed to Composite vs S-Video as I see it. So the Composite players may have a decent Y/C filter which is fine. But this clip does not really challenge the Y/C separating filter IMO. The LG composite player may also have (over-)compensated the sharpness/details loss caused by its luma/chroma filtering with post processing causing annoying halos.
Maybe someone else has a different view.

[lollo]

I agree with the commas too

[Alwyn]

Just doing my bit to provide the OP with some evidence so that they can make a reasoned decision.

Perhaps, Sharc, you could put up an comparison that shows my comparisons are not representative of the scenario you mentioned.

Everybody says S-Video is so much better than Composite but is it really?

[aramkolt]

S-Video is definitely better in certain situations, generally with fine patterns and can also get rid of some edge rainbowing. The issue is when repeating lines which are varied DC voltages get confused for AC waveforms that are then interpreted as chroma when they are not supposed to be. Another easier example is if you watch movies that are supposed to be purely black and white films. Often you'll find rainbows all over the place without S-Video. The mixing of chroma and luma also seems to cause detail loss in luma as the AC waveform seems to sometimes smooth out luma to some degree as well.

Now, if you're eventually uploading to YouTube, the YouTube compression can smooth out a lot of the S-Video gains, so it can be a bit harder to see it there. It'd probably be better to have side by side comparisons of a zoomed in area as opposed to the full image so the detail is less likely to be smoothed over in compression.

The other thing is that all composite sources then go into a comb filter at the display or capture device and some of those comb filters are better than others. So it is possible with a really good comb filter that there'd be minimal difference visibly in S-Video on certain monitors, but you're gambling a lot on a comb filter that may not play well with the content it is being used for.

The test I'll eventually be doing is the Snell and Wilcox SW2 pattern captured via composite through a variety of capture cards and TBCs which has a moving zone plate that likes to confuse comb filters. You'll almost always get a strong rainbowing effect as the sone plate moves via composite, but never with S-Video. A good comb filter via composite will have minimal rainbowing on diagonal line segments and with the moving zone plate.

[Alwyn]

I didn't include one in my above comparison workflows, but I recommend a stabiliser such as a DVD recorder-passthrough with S-Video Out.

[Sharc]

Everybody says S-Video is so much better than Composite but is it really?

To put it simple and pragmatic: IF one has the choice between Composite and S-Video, use S-Video because it AVOIDS the deficiencies of Composite a priori.
And if one has S-Video available, using a high quality and properly screened S-Video cable is imperative because a poor cable will readily spoil the benefits of S-Video and one might come to wrong conclusions. Examples have been discussed before as well.
(Sidenote: One could expand the discussion and claim that one should prefer Component (Y,Pb,Pr, sometimes mixed up with RGB) over S-Video .... ).

The rest depends on the quality (details, true resolution, sharpness, color saturation, structure ...) of the source and - in the case of Composite - on the performance of the luma/chroma separating filter. Examples of sub-par Composite with its typical artifacts have been presented and discussed in this forum many times in the past.
For "tame" VHS sources - when using a decent Y/C separating filter (not the poor one of the Hauppauge USB live2 for example) - one won't expect eyecatching benefits from S-Video - as your example has shown, and the final quality depends more on other player properties than on Composite vs S-Video.

So reverting to the OP's question: Provided his current VHS captures have no basic flaws I agree that there is most probably no good reason to redo everything, just for the sake of replacing Composite by S-Video.

[Wills]

Thanks all, for the detailed and nuanced points.
I'm quite happy with the caps I'm getting at the moment; they're from sometimes 2nd gen sources (Video8 > VHS to reuse the Video8 tapes) so some of them are pretty crappy from the source, but it seems to be coming through pretty well and is totally watchable.

I do have a few questions though, if anyone fancies helping out a relative newbie to the cap game.

1. Now that I've got my Huffyuv files, I may simply want to convert them into sections to be able to put them on hard drives for the family.
I use Handbrake for that sort of thing - but is there an option that would give me a lossless version?
I've cap'd the VHS tapes in one long file, and some of them are 3 or 4 hours - would love a way to be able to chop them into event pieces but I don't want to do it in a lossy way.
I may decide to do a highlights reel, for which I'd probably get a sub for Premiere Pro for a few months.

2. I do have a question about formats - given that cap'd TV episodes can be downloaded at around 1gb for 45 mins in 1080, is there a way I can do that once I'm finished with the content? MKV or MP4 ? I'm getting an hour at around 35gb with Huffyuv, which is fine for what I assume is a sort of RAW content - but I'd imagine there's a way to package that afterwards to something smaller?

Thanks all!

[The 14th Doctor]

I'll try and answer these best I can.

1. There are codecs out there that are less lossy than others, but since you're probably just giving these out to family members, you wouldn't know what media players they are using to view your files. So your best bet would be x264 mp4 with a bit depth of 4:2:0 to maximize compatibility

You can use a Non Linear Editing program like VirtualDub to cut the recorded files into clips, just make sure to set VirtualDub's compression type to direct stream copy, so that no re-encoding is required.

2. Those files probably have compressed bit rate, so the 35gb per hour avi files are your raw 'master copies' you can encode them to a lossy delivery (ie x264 .mp4) format latter down the line.

[Alwyn]

I've cap'd the VHS tapes in one long file, and some of them are 3 or 4 hours - would love a way to be able to chop them into event pieces but I don't want to do it in a lossy way.

As the 14th Doctor has said, you can use VDub in Direct Stream Copy mode. This will leave your chopped-up file pieces in HUFFYUV, unrecoded. VDub has a job batching function where you can do your cutting into numerous files and then do the actual processing all at once. If you're unfamiliar with editing in VDub, check out my site here.

TV episodes can be downloaded at around 1gb for 45 mins in 1080, is there a way I can do that once I'm finished with the content? MKV or MP4 ?

Yes. Commercial video is encoded with (I assume) high-end gear/software which allows high compression/low file sizes/bitrate but with very good quality, which you're probably not quite going to get with consumer software. That said, with appropriate settings/bitrates, you can still achieve very good quality, almost as good as the original capture. I work on around 3GB per hour for 768x576 (SD) VHS captures encoded to H264/AAC. Others use CRF, 18 being a good all-round value. The exact settings available to be used will be dependent on your encoding program.

Re MKV or MP4, both are only "containers" for the underlying video and audio codecs. MKV is a magical panacea/holy grail for all things video. In your case, with simple VHS captures, I'd just use MP4 as the container and H264/AVC for the video and AAC for the audio.

You can use MediaInfo to analyse a file. In "Tree" or "text" modes, it will display all sorts of good info about the file, including the codecs used and the bitrates.

[Wills]

Thanks The 14th Doctor and Alwyn, appreciate it.

One last question - I have 6 or 7 hours of Super8 cine footage (from the '70s) that I'm going to pay to have digitalised, plus around 20/25 hours of Video8 footage ('80s).

Thus far, I've paid for one cine reel to see what the results were like (pretty good), and I've done a couple of the VHS tapes to my PC.

My question is about upscaling.
I tried TopazLabs for an upscale preview of a few minutes of both sources.
The Super8 upscale came out quite well - detail enhancement, less blurriness, not bad.
The VHS upscale was really rather ghoulish, to be honest. Not good.

So the question is - should I? Is it worth it?
There's an option within Topaz which is a one-off payment, and then it just takes time to process, and I can choose to use or ignore the results.

Any thoughts in this space? If it is worth it, is there a vendor that anyone'd recommend?

As always, ta very much for any answers! Really appreciate it.

[Alwyn]

Haven't tried Topaz; I'm tempted (curiosity killed the cat), but the apparent processing time would kill it for me so I haven't tried it yet.

[Wills]

Thanks Alwyn.

Before I start the main work of digitising all the VHS stuff, I'd like to sense check my settings.

I'm happy with the results, but it occurs to me that i haven't tried all the options - and whilst of course I should, I'm also aware that it could take hours to test and check every variation of the many many options within AmaRecTV, and it also occurs to me that this may not be necessary if some kind soul here already knows the answers.....

I've left pretty much all settings at default - but had to change a couple to get the source to show up properly in AmaRecTV.

Here are two screenshots of the main two settings pages - any thoughts?

1. 4:3 PAL is 720x576, right?

2. In the 3rd pic, the Recording tab - Video File Setting is set to a frame rate of 999 fps (which was the default) and it does work as is but clearly it's not capturing at that frame rate - perhaps it should be set to 25 ?

3. I've left deinterlacing as auto, which shows up as this txt file when you click Edit

; AmaRecTV Live Auto Deinterlacing List --------------
;
; *** Format ***
; - ip
; p : Progressiv (not use deinterlacing).
; i : Interlace (use deinterlaceing).
;
; - field
; t : Top field first.
; b : Bottom field first.
; a : Auto select.
;------------------------------------------------------

; ip field min_w min_h min_fps max_w max_h max_fps
#DATA p t 1280 720 59.94 1280 720 60.00 ; D4
#DATA p t 720 480 59.94 720 480 60.00 ; D2
#DATA p t 1920 1080 23.97 1920 1080 24.00 ; 1920@24p
#DATA p t 1920 1080 59.94 1920 1080 60.00 ; D5
#DATA i t 0 625 25.00 1920 625 25.00 ; PAL
#DATA i b 0 525 29.97 1920 525 30.00 ; NTSC
#DATA i b 720 486 29.97 720 486 30.00 ; D1 for Intensity Pro
#DATA i a 720 480 29.97 1920 1080 50.00 ; Other interlace format

#END

Many thanks indeed if you can help out here!
Cheers.

[Alwyn]

@Wills, lossless capture is pretty simple: it's either Yes, or it doesn't happen at all. That said, there are a few settings in Amarec.

I've made a detailed guide here. Those HUFF settings look OK.

[Wills]

Thanks Alwyn!

I compared all of my settings with the ones in your super helpful guide, and the only difference was the Deinterlacing which I had on AutoSelect but is set to Not Use in your guide, so I'll compare and contrast between off and on.

Cheers!

EDIT
I redid some one of the Video8 caps for a few minutes without deinterlacing - and the results are a slightly sharper picture; there's more noise, but also more defintion.
You lose on smoothness but gain quality - not massively so, but worth it, I think, IMHO, given that the quality is not great to begin with.