Future resolution?

[ronnylov]

EDIT: Take a look on this explanation on what lines of resolution is:
http://www.dvddemystified.com/dvdfaq.html#3.4.1

Well, now I am very confused. Maybe I have confused myself. I thought the value 336 on NTSC was "lines of resolution" and not pixels so it should be equal to around 448 pixels if calculating the total picture width, but perhaps I got it wrong (as many of the resolution links I posted). Remember that it's not only the Kell factor but also the aspect ratio in the calculation because of "the circle rule" of TV resolution.

The Kell factor is it the same at low resolution as at a high resolution?
Lets image a picture that is 12 pixels wide and 9 pixels high. I'm pretty sure I will see all the pixels on the TV when scaled to a TV picture. In this case the Kell factor would be 1.0 in both directions.

Lets take a picture that is 12000 pixels wide. The resolution of the picture when displayed as a TV picture will not be higher than the maximum resolution on the TV system itself, the Kell factors must be very low because the TV is not cabable to be better than a cetain limit.

So what happens at VCD resolution? Is the TV capable of showing all the pixels? Yes, I think it will display all pixels because the TV resolution is probably higher than that. So I'm not sure that the extended Kell factor would be the same at VCD resolution as at full D1 resolution. Meybe we can forget it completely because it's just confusing...

I guess that we must made some tests of our own to check what our equipment is capable to do. Make different captures on the same source using different horisontal resolution and compare the detail level of the picture. If we can't see any difference at all with a higher resolution then there is no need to use it. Even if the TV system may be capable of a certain limit your receiver, capture card, cables and playback device will make the results worse. Now we should also consider the end format. When making DVD we can consider using 720, 704 or 352 pixels width, that's what we can choose. I suggest using the full height 480 or 576 pixels. Use the available options and compare the results.

It seems that we can agree about that half D1 is enough on VHS sources. If it is enough on better sources depends on the quality of the source and your equipment to capture and playback the signal.

[SatStorm]

Exactly what I and other older users state for this subject for years now!
A simply search on the forum can show you endless postings for this....

Once again in short terms:
For analogue captures using mainstream hardware/ cables:
With NTSC, you capture the higher you can, you filter what is neeeded and you encode (best choice: 352 x 480).
With PAL, you capture at the framesize of your target framesize, you filter the less possible and you encode (best choice: 352 x 576)

Follow those simply practical rules and enjoy excellent results.

The alternative root is harder, much more expencive and you gain something like 5% at best in terms of quality.
For PAL users, and only from VHS sources, there is a small alternative: Interlace 352 x 288 as a target framesize, from a proccessed 352 x 576 (or higher) capture. This alternative has the minus I already posted in this same post.

[lordsmurf]

EDIT: Take a look on this explanation on what lines of resolution is: http://www.dvddemystified.com/dvdfaq.html#3.4.1

Again, that info is based upon the original AOL page of Allan Jayne, and I find errors in some of that information. The DVD Demystified site is an encyclopedia of knowledge about DVD, but again, not flawless and not necessarily perfection of the theory in use with video. Much of it sounds great on paper (the AOL site and information based off of it), but doesn't work out as explained. I'm going to give them the benefit of the doubt, and attribute it to miscommunication and stilted use of language rather than intentional errors or incorrect information. Putting technical and theoretical information into normal language is not easy. Given that I'm a journalist with a decent grasp of technology, I find it easier to do than most.

You can make video as confounding as you want. As somebody wrote me today, "working with digital video is akin to experimental witchcraft." It gets tougher when jargon (with no common language equivalent) is introduced and misinterpreted on a large scale (as is the VCD=VHS myth). I'm finding that this has also happened when attributing digital values to analog source.

And again, remember that discussing Kell with videographers is like discussing the Israeli-Palestinian issue. Heated debates from each side, and nobody can agree on anything.

SatStorm:

I've read most of your posts. I especially researched most of your information on resolution and PAL. I would say with AVI (for advanced editing), the theory of "NTSC cap highest" may ring true. But not for MPEG. Especially not if conversion is the only goal (using 352x480). It does still ring true for VCD quality. It is best to capture high and convert down to MPEG1 VCD (at least on non-ATI cards).

On PAL, the 352x288 isn't a goal for everybody, and some players reject it (cannot find definitive information on interlaced PAL MPEG1 support being in the DVD spec). It is actually undercutting original resolution of the VHS PAL source (about 260x560 from what I've found at a glance) but given the interlace and relative quality on well-used MPEG1, it's a "cheap trick" as you've said before to allow for many hour of video on a PAL DVD with very decent quality. I'm not aware of any capture device that allows a capture resolution of MPEG1 x288 interlaced, but encoders can surely do it. Capping must still be 352x576 for most (unless you know of good software for x288 interlaced capture).

[SatStorm]

You can't capture interlace @ 352 x 288, but you can encode interlace at this framesize, using TMPGencfor example. This is a good alternative and it is "VHS quality" for sure (visually at least). The problem is that most DVD players won't play it correct. Most players gonna treat it like a progressive mpeg 2, a "null" mpeg 2 if you prefer.
This trick is too old, I still remember Sefy searching for it, back in late 2000 if I remember correct....
Personally, I still use it a lot, expecially for stuff I want to store in digital, from VHS sources...

From the other hand, when I say capture, I always mean capture to avi, not direct to mpeg 1 or 2.
You see, most people when they start with this hobby, believe that they can save time by capping direct to mpeg 2. Even if this is possible and has really good results, expecially if you do it with 1/2 D1 framesize and DVD bitrates, when you have to encode analogue material like VHS/SVHS home movies, you realise that ain't enough!.
The truth is that from VHS sources, you can fix things with filters you coan't imagine that is possible. For this step, avi (uncompressed or with mjpeg or Huffyuvis yet neccessary.
Filters can do miracles: They can emulate TBC, they can smooth/restore/stablise a picture, or restore the original progressive frames from the interlace source!
Soon, the ones who want to convert to digital their VHS stuff, end up capturing avi, filtering, encoding. It is still the only 100% quality way!
Also, DVB transmissions are bad. They also need filtering. So, when you cap direct the stream transmission (like I do) or capturing analogue like any analgue transmissions, you need filters to remove the transmission macroblocks and - very often - some noise of the source. Just imagine a DVB channel transmitting music videos from VHS tapes. You need filters to clean the DVB macroblocks (if the transmission sucks, as it does always....) plus filters to clean what is possible from the original source...
You can't do those things correct IF you capture to mpeg 2...

Capturing direct to mpeg 2 is only good for LD, really good satellite transmissions (which I repeat, there are plenty other ways to capture direct the stream copy with no quality loss, than capturing the picture the analogue way...) and good aerial transmissions (terrestrial analogue). For VHS sources, it is not a good idea, even if you use CBR 15.000kb/s!
It is ofcourse a faster way, but the factor "Time" is relevent when we have our precious Family Home movies to encode, or our old (and so difficult to find) VHS collections.


When you cap direct to mpeg 2, then it is natural to do it direct to the framesize you intent to burn on DVD/CD. With todays mainstream technology everything from analogue to digital beyond ~440 x ~ 500 is an overkill, for both PAL and NTSC. So, expecially for storing to DVD, capturing at 352 x 576/480 is OK and for most desent cards around, the best alternative. With cheaper cards, capturing Higher might help on NTSC, but again, I suggest to the owners of those cards, new ones!

My final word: Theory and Praxis ain't always compatible. Reading the theory helps, but only practising those theories can show you the truth. Practising and testing many times can proove that theory ain't working as it should. The best example for this is Kwag: If theory and praxis was alike, he would never existed!

[lordsmurf]

They can emulate TBC

I wouldn't say it exactly like that (don't want the newbies to think software can replace a TBC). There are things a TBC can do better because it is hardware (pre-processing of the captured video, too). It's always best to fix the problem as early up the line as you can (which is why a good VCR is a must on VHS), before it hits the digital world.

But filters are DEFINITELY something everybody should learn.

Good talking to you.

[SatStorm]

I like reading from you too! You can express the things you know for this hobby in a much better way than I can do, expecially because I don't have great english skills...

I didn't say that filter can do TBC, I said filters can emulate TBC.
Time Base Corrector, can help a lot for framedrops, stablise old VHS tapes and to make sure to fill missing fields from tapes with dummies. So visually, most of the time the result is so better.
The minus of TBC is that on some cases, something like "pixelaration" can occur from VHS sources, expecially on full Coulors areas (Blue, Yellow and Red). So, on rare cases, there are reports that TBC make VHS tapes look worst! I saw this happening couple of times on pro VHS PAL tapes!
I don't know for NTSC, since with this system, TBC also stablises the colours, a great nightmare of this TV broadcast system...
Anyway, the PC alternative of the hardware TBC, is the temporal filters. I suggest to anyone interest to convert VHS to digital, to read about them. You can succeed really great resutls, some time unbelievable, with those filters!
I was able to convert 15 year old VHS tapes from not so clean aerial transmissions, in an ubelievable way because of those filters! I couldn't believe my eyes! Noise 15 years old plus the typical VHS noise disapear and I have a picture of some really important for me stuff, in an amazing quality, I never dreamed of!

[lordsmurf]

I don't have great english skills...So, on rare cases, there are reports that TBC make VHS tapes look worst!......You can succeed really great resutls, some time unbelievable, with those filters!

Your English is just as good, if not better, than most people that I know and/or write on these boards. Is it not your first language or something?

Final Cut Pro has image stabilizing filters too, but take days just for a few hours of footage. Those filters amaze me. That's something a TBC cannot do (at least not one I've seen).

Yeah, my TBC in my JVC SVHS can make tapes look worse on rare ocassions. Luckily I've figured out most work-arounds and methods for improving video. Most of my DVDs look better than the source tapes.