VHS Transfer (352x480 Vs 720x480)

[hovhaness]

Which do you prefer?

352/360x480 interlaced
352/360x480 deinterlaced

Or

704/720x480 deinterlaced

Or

Other

And why?

Aficionados like lord smurf would say half cd 352/360x480 mpeg2 interlaced is the very best way to go.
They say this is where VHS res tops out & that all the detail that is needed is right there & going too large with resolution is overkill.

What do you guys think?

I kind of agree that it does top out there but I have always felt that recording classic videos in this resolution gives the movies an over-saturated & less authentic feel.
There is too much color bleed & the noise is also more prevalent.

I personally like to capture at 720x480
then re-encode to 640x480 & use AVC/mp4.
In an odd sort of way it seems to give more breathing room to the video
It seems to have more space. which is easier & more pleasing to my eyes.
There is also less noise & in many cases I find I need 0 noise reduction.
Plus I can make my videos nearly 1/3 smaller while maintaining the quality.

Here are samples of both ways
For both I used Yadif for a fast & effective non destructive deinterlace.

Here are the samples

352x480.yadif

Format : MPEG Video
Format version : Version 2
Bit rate mode : Variable
Bit rate : 2 124 Kbps
Nominal bit rate : 2 500 Kbps
Width : 352 pixels
Height : 480 pixels
Display aspect ratio : 4:3
Frame rate : 29.970 fps

http://www.mediafire.com/?7owmdkr3pvu794o

640x480.yadif

Codec ID : avc1
Codec ID/Info : Advanced Video Coding
Bit rate mode : Variable
Bit rate : 1 600 Kbps
Width : 640 pixels
Height : 480 pixels
Display aspect ratio : 4:3
Frame rate mode : Constant
Frame rate : 29.970 fps

http://www.mediafire.com/?32yyyy4bb2pum30

[Blue126]

I would go with the 352x480, I like the noise, because it is keeping the detail, as you see you lose it in the 640x480. Do you have the option of going for say 512x384 ? maybe get the best of both worlds.

That's a option I have been using a lot recently, mainly because I have been using a tablet for viewing which looks great and also a decent size for viewing on a computer.

Like you say, sometimes when you go to a larger picture ratio you do lose detail and it can look kind of plastic looking with color bleed, smaller sometimes gives you a better viewing experience, but at the end of the day its down to your source material.
You can sometimes increase picture ratio and get more detail, but not always, but it is possible.

[hovhaness]

Thanks for the feedback
Yes we do have to keep in mind both source and also the destination.
Certainly for tablets & smartphones & ipod touch etc.
The smaller size is wonderful.
Also it is good for DVP type DVD players because it upconverts to 480 then stays there.
The issue arises when watching like I do most times
On a 22 inch progressive scan monitor opened up full size with a natively high monitor res.
I am gonna do more testing.
It is just difficult to gauge which way the people are flowing with adult movies.
Many have moved away from the DVP method & to the HTPC way which requires a larger aspect ratio & more kbps..

I don't have the option of 512x384.
It is either smaller 1/2 DVD 352 with a mpeg 4000kbps transfer
or full dvd 720 with either a full res AVI or 8000kbps mpeg transfer
352x240 requires no deinterlace but removes 1/2 of the info & detail with it so I won't go there.
I am not too sure of how the math works with a 512x384 encode.
I have heard that you have to crop 9 pixels from the left & right border, before scaling to get a 4:3 res.
For this I would have to encode first in 720 & then drop down. This would work better than upconverting?
I will do more research on this

K I found this helpful thread on it
http://forum.videohelp.com/threads/3...-D1-or-Half-D1

Please more suggestions
All are welcome.
Feel free to bitch slap me on my ass with great stuff, or tell me I am wrong.
I don't take anything personal & I love to experiment & learn.
The more I learn how to improve the better the encode result.

[bigbobnobody]

Honestly......I've never understood why anyone would transfer something
off of a VHS tape, laserdisc, or what have you at anything less than what
it was to begin with on that tape. There's a finite amount of pixels to begin
with from the original, and to me it seems nonsensical to further reduce the
amount of those pixels by reducing the resolution. Once encoded......that
pixel amount can't be changed again. Sure......a video can be upsized, but
the pixel amount isn't increased......only the size of the pixels themselves.
With the bandwidth that most people have at their disposal these days
on their internet connections, reducing the size of a video for those with
slower connections simply makes no sense. So I guess you can count me
in that group, that says that to create anything less than at it's best potential
is a waste of time. As far as interlacing and de-interlacing goes, I always
de-interlace during a transfer.

[hovhaness]

Thanks Bob
I am in the same party with you that I would rather be over than under.
Then shrink it with all the larger sized pixels i need.
:-)

[i4004]

Originally Posted by bigbobnobody

As far as interlacing and de-interlacing goes, I always
de-interlace during a transfer.

and by doing that you probably destroy any gains you got by using higher resolution on capturing.

(only the very slow deinterlacers are very good..some of them are found as avisynth functions, ala mcbob etc.)

during the tests we did for doom9 capture guide it turned out that there is some difference bretween 480 vs. 352 horizontsal pixels.
not much, but it's there...

after deinterlacing, denoising is another can of worms as doing it usually involves some desctruction of the details (and so does the avs/h264 codec, as it employs inloop smoothing filter that less people turn down or turn off completely). again just like for deinterlacing motion compensated avisynth filters are usually the best way...because they will be best at preserving detail during motion, this being important because human skin in motion (like porn is) is actually toughest thing to denoise.
offcourse those filters are slow too.
few fps on 2 core 2ghz machines etc.

hovhaness, 640x480.yadif is an excellent example of too smoothed video which probably lost a lot of detail mostly during avc encoding (even though it is a question what detail vhs has to begin with..some vhs copies just don't have that much detail).
i would say avs is acceptable (not "great", not "good") codec for hd content with less noise, as new codecs usually dislike noise.

sometimes one should weigh: do i want smooth looking video without noise, but also without details, OR a bit of noise but much more details preserved.

i usually just cap to mpeg4 codec at constant quantizer (not doing any processing while capping, at 480x576, as i'm in pal land), say 3 or 4, and then i have the thing capped, and i can afterwards try to improve it, perhaps when cpu power costs less (ie faster processing) etc.
but even if i don't improve it, it's ok, as i didn't really lost any quality vhs had there.
(and i'll probably soon upload just such a thing, vhs straight to mpeg4 and constant quant..)

another thing is the display: best results will be with crt tv displaying sd interlaced content. watching interlaced vhs directly (or wathing dvd copy of that vhs tape) on flat panel tvs results in garbage (because then the tv will do deinterlacing, and new tvs are usually pretty bad at scaling and deinterlacing).

should be better quality (on flat panel tvs) if you use pc to process it on the fly (dscaler) or go the whole way of deinterlacing and denoising with those slow filters...

it is kinda weird situation when with new technology you get worse quality on old content, but it's exactly what's happening, if you pay attention to details.

incase i want something really decent i'll cap to mjpeg and encode to mpeg2 (it is least destructive to sd video even though it requires higher bitrate...there's no free lunch even in video compression..heh) with some avisynth denoising and sharpening. won't even bother with deinterlacing, as i have crt to watch it on...
for web i would probably pick some mpeg4 (not h264, though) if filesizes of mpeg2 got too big.


to make 512x384 you don't need to crop 720x480, just resize it and eventally crop later to multiple of 8 if there's too much black border around the image.
if you crop first, you'll mangle the original aspect ratio...

[bigbobnobody]

Believe I'll take deinterlacing, and eliminating the shadow lines during movement
with an interlaced video, 7 days a week, and 3 times on Sunday. I guess it's
one's own preference as to which you can stand the least, but to me it's
annoying as hell to watch an interlaced transfer. I always assumed that it
was the result of someone just not knowing what they were doing......but I
guess there's those that prefer their videos that way. Not me......ever.

[chip]

I read i4004's post as meaning that deinterlacing on-the-fly as the source is being captured produces undesirable results. And that doing so once the raw video is done recording produces better results. I don't think he was suggesting not deinterlacing?

Now if the intended playback system is a CRT then you don't have to deinterlace. But for most of us the reality is that we don't have one anymore and likely never will again.

[bigbobnobody]

Originally Posted by chip

I read i4004's post as meaning that deinterlacing on-the-fly as the source is being captured produces undesirable results. ........

Guess I read it another way Chip, when he said that de-interlacing
defeats the purpose of encoding to full resolution 640x480 in the first
place. I prefer to do my editing on my transfers, after I've got them on
my computer as opposed to on the fly while transferring, so the thought
of de-interlacing as you go never entered my mind. I also prefer to bring
the files to my hard drives as VOB files, rather than un-compressed, mostly
because I can't afford the un-godly amount of space uncompressed video
takes up! lol The whole point of my original post, was to get the point
across, that shrinking a video down to a smaller resolution, (at least thru
my eyes) serves no purpose other than taking up less hard drive space.
No way no how could anyone ever convince me, that creating a smaller
resolution video, would give it better clarity. Heck .......everything looks
pretty clear on an ipod right? lol

[hovhaness]

Thanks so much i4004 & chip for adding input..
Your post was a brainful i4004 & I will come back to do some experiments as time permits.
I do really think that nowadays the important thing is to think about tomorrow.
How many are watching on tablets & portable devices.
I have noticed that many of these devices do a decent job at upscaling to 480
For those web streams & such I think I would rather keep them smaller & interlaced.
Still for viewing them on larger monitors with huge native resolutions of 720+ I like beginning at 480 & using a faster & less destructive deinterlace like Yadif.
Are there any others that I should be giving a shot to that do a better job than Yadif?
I can sacrifice speed but I don't want to give away too much.