MPEG2 capture using ATI AIW 9600pro - settings

[lordsmurf]

That list of 8 items is still not all facts. Some of those are a mix of opinions, theorem, hypothesis, and experience, especially towards the end. There is a difference.

You're welcome to draw your own conclusions, but their is plenty of evidence and practical application to back up the idea that 352x480 will easily capture all the information needed to create a high quality image that is adequate to retain all of the detail found in the source. Especially perceived detail, if not perfect mathemical accuracy.

For like the 3rd time in this thread, do what you think is best.

On some medium-res sources, 720x480 will give an extra edge to an image, especially if you add detailers to the mix. But it's not an amazing leap in image quality, it's very minimal and easily overlooked. I've said this before many times. The drawback here is you have to use more than double the bitrate and can put less video on a SL disc (no more than 60-90 minutes, unless you want to start really compressing).

[Nelson37]

Specify and detail ANY inaccurate information contained in 1 thru 8.

Specify and detail HOW a 352 x 480 capture is created by the ATI card.

Using 3:2 Pulldown removal, a 2 hour source will easily fit on a single disk with a specified average bitrate of 8.5.

7.5 to 8 without the IVTC. This is at 704x480.

[vhelp]

Evening everyone

The forgotten facts from past (and documented) research..
In all honesty, I believe that the issues brought out in this topic are based on some
forgotten findings and facts (found through research) though hidden inside these old
topics. In brief, one of the forgotten facts has to do with the way *each* capture card processes the video inside, aka, resizing and/or are incorporating certain low-pass filtering based on certain parameter conditions.
And, as such, some users here may be still be using these devices in their video
endeavors, and need a readup on the following threads for clarity sake.

A little re-reading is good for the mind..
Mind you, I resorted to re-reading/re-researching this to refresh my memory of some
of the key elements and what-nots. The topics were refreshing to read.

So, I recommend that before you all go on about in continuing with your "who is
is wrong or right" scenarios, etc. that you all at least first re-read the topics
below, before going any further in your responses. And, since neither one of
you bothered, I started first, in re-reading much for a refreshing course.. after
all, I participated in the research, too

Anyways..

The key(s) of infomation lye somewhere..
I note with you all, that inside the topic of the, "Amazing discovery.." holds
much of the key information of the various tests that have been done already to
give valid reasons why each *best* (I mean, subjective) resolution is dependant
upon the capture card; and its drivers; that govern the image quality, hence the
optimum capture resolution.

You just have to re-read everything (the link below) in their entirty, to remeber.
Then, we can get back to the topic, the original reason and question that the
member brought up, now on page one

Anyways..

Reading the sources to recall what some have forgotten..
The following are resources for reading, as such, please bare in mind, some
of the comments/puns are of that ERA in the timeline, and some things should
not be taken out of context or other -- in other words, be nice

Also, some of these threads have exhaustive demos of images depicting the findings.
Please review them for clarity. And remember also, that during *that* ERA, capture
equipment are reflective of that timeline.

Now, because of the above and mixed in with todays ERA of discussion in this *very*
similar topic, there may be some things: forgotteen; confused; changed; or revised;
or what-have-you. So, bare all these points in mind when reading and researching
these issues further via the link posted below.

(All those who participated, I included in the credits as listed)

The facts leading toward the "optimum capture resolutions":

--> Amazing discovery in capture resolutions (IMHO) -- by fmctm1sw; -- 01.14.04
** fmctm1sw; indolikaa; DereX888; lordsmurf; SatStorm; trevlac;
** W_Eagle; Alan69; vhelp; You_Are_Alive; BrainStorm69; racerxnet;
** FulciLives; andie41; spiderman2k1; Zetti; vmesquita287; Northstar;
** lacywest; holistic; SLK001; ronnylov; tito13kfm; deadpac;


--> TV Resolution -- by c_hernandez32; -- 11.19.03
** non-linear; trevlac; xtreemkareem; lordsmurf; fvd72; rtkeen;
** wulf109; vitualis; proxyx99; Cornucopia;
** Nelson37; SatStorm; energy80s; Heavensent;
** FulciLives;

There are other previous topics that covered this area already, but I'm too tired
from all this re-researching, etc. I did today for you guys

-vhelp 4090

[Nelson37]

Trevlac's findings are the ones that confirmed my original observed results. I could quote a fair amount of that from memory.

Initial investigation of these results came from yet another post noting specific degradation in ATI capture going from SVCD to CVD res. Have seen no definite evidence ATI introduces an additional filter at some res between these similar to BT chips or if the resize is just particularly inadequate at 352 res. I tested this myself and noted the quality dropoff from 480x480 to 352x480 seemed more significant than the minor res change would indicate.

Capture res cannot be set to match source res. That is the basic point. It is fixed.

The repeated reason for capping VHS at 352 is either to match capture res to source res or because Nyquist suggests 352 as best. Whether or not these theories are accurate does not matter as the capture res cannot be changed.

Now if you just think 352 looks better than fine and dandy. If you think or can even prove that 352 is theoretically better as a capture res also fine and dandy. ATI simply will not capture at that res, only resize to it after the capture.

To put this another way, if you were handed a file with a resolution of approximately 672x468, what would you resize this to in order to make it DVD compatible?

[davideck]

Davideck, I was unaware that Nyquist addressed resizing. My reading of it was primarily where it addressed the "correct" amount of oversampling for capture of an image. My rejection of it is based on the fixed nature of ATI capture card resolution. What exact numbers the theorem recommends for capture are therefore irrelevant, since the fixed capture res cannot be changed.

Nyquist does not specify an optimal capture resolution for a given source resolution, but rather a minimum requirement.

Stated differently, Nyquist specifies the maximum source resolution (bandwidth) that can be preserved for a given capture resolution (samples_per_line).

For your example of 672 samples_per_line, Nyquist identifies a maximum preservable source resolution of approximately 504 lines. If this image is resized down to 352, then the maximum source resolution is proportionally reduced down to approximately 264 lines. These "ideal" results assume infinite filtering and therefore need to be further reduced due to practical circuit implementations.

On some medium-res sources, 720x480 will give an extra edge to an image, especially if you add detailers to the mix. But it's not an amazing leap in image quality, it's very minimal and easily overlooked.

Detailers introduce higher frequency components into the video by sharpening edges to create artificial detail. These higher frequency components can also exist as actual detail in original source content.

I agree that some may not notice this detail in their sources or they may simply prefer to trade off bandwidth for a longer record time. Others may notice it and wish to preserve it.

Picture quality is subjective. I am not making claims with regard to the importance of the additional detail. I am merely claiming its existence and what is required to accurately preserve it.