VHS to SVCD, crazy or ... ?

[TMcD]

I am converting a TV series from VHS to SVCD.

My question is, since the VHS is 352(320?) X 240 (regular VHS not using S-VHS and captured from normal not digital TV signal)... Is it doing me any good to capture and convert to SVCD format 480 X 480, or am I just wasting my time?

I want the best image I can produce through my TV.

Thanks,
Troy

[Martí]

Hi!

Yes, VHS has the resolution you said, but if you encode it at 352*240 for VCD, you will miss half of the resolution because it's not like a VHS to VHS copy...it's a compression to another format with very low resolution.

What I'm asking for is if it's better SVCD or XVCD for this process... for me SVCD gives excessive blockiness.

Any opinion?

Martí

[TheInformer]

TMcD:

Capture your video at 352x480. 352x240 is only half of the video signal. There's really no reason to capture at any resolution higher than 352x480 because you don't gain anything.

Once you've captured the video at that resolution, you can use TMPGEnc to encode to your target MPEG-1/2 file.

[TMcD]

Marti,

352 X 240 Interlaced won't lose half the resolution, since two frames are shown per second. (Field a/b).

XVCD, seems like an option, but is non-standard (curious to see if it works on my Apex 1500.)

The thing that confuses me, is that I've done a VCD sample of the show and it looks really good on TV, but the SVCD has blockiness ocassionally. (Odd since it has more data.)

Not to lose sight from my original question.. any make sense of this, and has the experience, heh, no "Here's what I think people."

I've got 30 shows to encode. I want to get this figured out.

thanks,
Troy

[TMcD]

A lot of lookers on this post. It seems that people are curious to the answer, but it also seems that no one knows it.. or is not willing to part with their wisdom.

Troy

[thxkid]

well 1st problem, VHS is allready a crappy picture, so you can't improve on crap(why I only use SVHS). But if you must use a VHS tape, I would capture it as 352x240 but not just 1150 bitrate, go for the 2520 bitrate for VCD, I use this bitrate all the time(never use 1150) and I get no blockines at all, picture is a little soft but since you allready have a soft picture with VHS, should not be problem, I personally tape iin the higher SVHS (400 lines) and then capture to MPEG2 SVCD(350 lines) with fantastic results.
I hate to death Crap VHS is why I use only SVHS and store all my programs on VCD or SVCD so much better than 24 year old technology of VHS(RCA 1977) beta 1st 1976. I had owned VHS since my 1st one in 1980. Owned nothing but SHVS since 1992.

[KoalaBear]

Riddle me this, Batman: what is it about digital video that makes new guys think it's sooooo easy that skills that have taken some of us months or years to accumulate can be imparted to them at the drop of a question?

An experienced hobbyist can do exactly what you'd like, that is, transcode a broadcast television program from VHS to [S]VCD with excellent picture quality given the limitations of either media. But the key word here is experienced, and by that I mean 6 to 18 months of capturing, processing, editing, encoding, authoring and burning your way to success.

352x480 is more efficient than 480x480 at the same bitrate, about 25% better. Advanced deinterlacing techniques like IVTC can boost your effective bitrate another 20%. Noise reduction methods exist that can dampen extraneous motion and prevent some artifacts from appearing in the first place. The encoder that creates your files and the decoder that plays them back both play a role in the number, type and severity of artifacts you experience.

If you have the patience, the interest and the fortitude you'll eventually learn all of these things. But you won't learn them today, this week, or for many, many moons to come.

I'm not trying to discourage you, Troy, but I am genuinely curious to know what you expect.

[Stansell]

My best advice is to record a short section of one of the shows using SVCD and VCD resolutions. Find a portion of the show that is typical of the amount of motion in most of the shows.

Encode using various methods, try out some filters and such. Sit down in front of the DVD player and see which method look the best. When you find what looks best to you, go with that.

[TMcD]

Stansell,

I appreciate the advice, and even though CDRs are cheap... I've gone through lots testing (and CDRs) so far... just looking for some tid bit I may be missing.

I have found so far that CBR looks lots better than VBR (even trying various parameters.)

Troy

[flaninacupboard]

that's a bit harsh, but i see your point.....

my advice, its coming from vhs, use vcd. capture at 352X480, deinterlace combine to 352X240 and adjust bitrate for disc size. (assuming 45 minute show, should be around 2.15 mbps) you may want to switch on noise reduction in tmpgenc, but only use it lightly. its really not worth trying svcd for vhs rips, because the resolution just isnt there, and also noise from vhs means your bitrate needs to bemuch higher to preserve the image, which just isnt going to happen with svcd.

[TMcD]

flaninacupboard,

Deinterlace?

I thought that since the VCDs/SVCDs or whatever they turn out to be... need to be kept Interlaced, since the final viewing source is a television. That is what I have done with early tests and they look fine on my TV.

???

Troy

[Anonymous65]

VCD - 352x240 always non-interlaced
SVCD - 352x480 can be either

It's the player, the DVD player that encodes the mpeg file into a standard interlaced NTSC video signal, the source can be either interlaced or not.

[TMcD]

skittelsen,

Thanks.

Please make sure you re-read your posts before you hit submit, since me and other greenies can get confused and take things literally.

I think you mean to say that the DVD player decodes the signal into interlaced, so the media (ie CDR) can be interlaced or deinterlaced.

Since my rip from the VCR is interlaced, I'll keep it that way, plus it cuts down in time.. Not having to run it through a filter.

Thanks,
Troy

[House de Kris]

TMcD,

You know, skittelsen is a quite helpful member of these forums. What he wrote actually makes more sense than your interpretation. And, speaking of re-reading posts before submitting, perhaps you should take this advice yourself.

Was it not TMcD who wrote further up:

"352 X 240 Interlaced won't lose half the resolution, since two frames are shown per second. (Field a/b)."

If you think that only two frames are shown per second, you know even less than you think. OK, perhaps you really did intend to say, "...since two fields are shown per frame." But you didn't.

More importantly, 352x240 does INDEED throw out half the resolution. People ARE helping you, but you refuse the information by 'correcting' them with erroneous data.

I, for one, am not inclined to offer any assistance in this case.

[TMcD]

House de Kris, you said "I, for one, am not inclined to offer any assistance in this case."

And you didn't. Calm down. No need to flame on. I was basically saying that newbies (such as myself) are in the take things literal stage.

No further comment from you is needed, unless you are trying to help the situation.

Troy

[ngnr]

Hi Troy. I have been trying (note I said trying) to encode good video off and on for about a year and I am still not satisfied with my results, despite numerous long encodes with varying settings. I have some 8mm video that to me is priceless (daughter's day of birth) and I am trying to get this into as clean a digital video as possible, while sticking to standards that I hope will remain around long enough for my daughter to enjoy it so I am restraining myself to either standard VCD or SVCD.

I have tried VCD and while the quality is not bad, I find it is too "soft" so I am now experimenting with SVCD. I hear a lot of opinions about how this is a waste of resolution but to me, the result is better. SVCD gives me more sharpness and clarity but as you say is more blocky in motion scenes. I am using interlaced encoding which I believe gives me better results in motion scenes as long as the field order is right, which took me a while to get under control. I will continue to tweak so until I hit upon a combination that works for me.

What's my point? I think you will have to find out for yourself what works for you. Frankly, what some people call high quality, I would call crap and what others call crappy video, I sometimes find is aceptable. It just depends on what works for you and it's affected by so many variables that I don't think you are going to get any quick answers. Sorry but that's been my experience so far. But I still keep reading this forum...

Good luck.

<font size=-1>[ This Message was edited by: ngnr on 2001-12-12 19:39:18 ]</font>

[Anonymous65]

I actually wrote what I ment to say

In the DVD player, there is a NTSC and/or PAL encoder. The DVD player reads the disk, and loads each frame into memory, no matter how big the frames are, 352x240, 480x480, 720x480 and so on, and no matter if the source file is interlaced or not. Obviously, if the source file is interlaced, it will already be in the correct output format. Anyway, the NTSC encoder reads the information it needs, and makes one interlaced NTSC signal.

Also, the notion that VCD at 352x240 is only half the resolution of VHS is not correct. I don't have the formulas and math in front og me, but the nature of interlacing, and that two fields are not exposed at the same time, plus the effect of line paring during display, you do not get 480 lines out of a 480 lixel/lines of resolution in an interlaced signal. By dumping one field, you loose something like 30% and not 50% (half) the vertical resolution. And by using vertical 2:1 reduction during capture, you will barely see any difference between 352x240 and 352x480 video. I suggest you give it a visual try, I have...

[ngnr]

<TABLE BORDER=0 ALIGN=CENTER WIDTH=85%><TR><TD><font size=-1>Quote:</font><HR size=1 color=black></TD></TR><TR><TD><FONT SIZE=-1><BLOCKQUOTE>

On 2001-12-12 19:51:43, skittelsen wrote:
And by using vertical 2:1 reduction during capture, you will barely see any difference between 352x240 and 352x480 video. I suggest you give it a visual try, I have...

</BLOCKQUOTE></FONT></TD></TR><TR><TD><HR size=1 color=black></TD></TR></TABLE>

So have I and I would agree - I can see no difference in captures at 352x240 and 352x480 once encoded to VCD. Maybe it's my eyes...

[Anonymous65]

Not sure if I read the above correctly, but when making a VCD, *capture* at 352x480 will give you better detail than capturing at 352x240.
But, if you make a VCD (352x240) that was captured at 352x480 and reduced to 352x240, and compare that to a non-standard video-CD at 352x480, you will only see minor differences in detail between the two.
If you don't like the blurr in action video of 2:1 vertical reduction (basically a deinterlacing process), then capture at 704x240 and reduce it to 352x240 during encoding. This will give you basically the same added resolution as capturing both fields and doing a 2:1 reduction.
So, in other words, *during captures* it helps to oversample and then downsample to the VCD standard. This might be getting too confusing...

[ngnr]

You're right - it is confusing. I thought we were saying the same thing but maybe we're not.

What I am saying is, in my analysis, which has not been exhaustive I admit, capturing in 352x480 and then encoding to 352x240 yields no discernible difference in the final result than capturing at 352x240 and encoding to 352x240. Yes, there is more detail before the encode, but at the standard VCD resolution after encoding, they look the same to me, even under magnification. I did this both with TMPenc and LSX and the results were the same but it's possible I didn't follow the same procedure you did. Can you provide samples or other information that would dispute this analysis? I'd be interested in seeing if this is worth revisiting.