#DPS/charge for output.
10 messages in this thread
>>Anybody want to place any wagers. <G>
I won't stoop to this level to decide whether the equations that engineers use
to determine resolution, video bandwidth, and other specifications are
relevant.
>> through instead of 972 (486×2) the number should be 1440 (720×2) which is
totally erroneous.
If the recording device is capable of doing 720 lines of resolution then the
pixel level of 1440 is correct. However if the recording device is speced at
720 pixels then the input would be 720 pixels. 720 pixels would have a maximum
resolution of 360 lines. That is not to say that graphics would look worse
than on a deck of 450 lines because as I said before a more precise way of
evaluating graphics resolution is rise time rather than lines of resolution. .
>> D1 is 720 pixels so that by this reasoning it has only 360 lines of
resolution.
That may be so as I do not have the D1 specs infront of me but I do for D2 and
it is 6Mhz BW which equates to 636 lines.
>>The whole scenario about rise time is related to electrically generating
output from a character generator, and is not relevant.
Oh, but it is most relevant. In fact, rise time is probably one of the most
reproducible and accurate ways to measure resolution in a graphic environment.
The generation of graphics with a CG or with a computer is essentially related.
In my opinion, a closer relation than the generation of graphics with a camera
which is where lines of resolution are used.
>>Even more critical to picture quality is Signal to Noise ratio and Crosstalk,
but this also has nothing to do with the discussion.
Right!
>>If one follows the full logic of your position, it means that if I grab a D1
image with a Matrox CCIR 601, and output it again I've lost half the resolution
of the original. Not!
Wrong! This is not what I said. If you grab an image at 720×486 pixels and
output it again it should in theory be 720X486 pixels. The point both Greg and
I are making is not that you are loosing anything. it has to do with a
relationship between lines of resolution and pixels. The mathematical relation
that lines of res. = pixels is silly because it takes two pixels to produce one
line at the very least except when there is only one line then it takes 3.
Let me explain it another way.
If we have a horizontal line made up of 10 equal sized units. Each
unit is white and they are adjacent. Each unit represents the horizontal part
of a single pixel. Now, how many pixels do we have? answer: 10. How many
stripes do we have? answer 1 Now for the hard part, if we set every other one
of the units to black how many stripes do we have now? answer 5 With every
other stripe set to black we produce the maximum number of stripes with 10
units (pixels) which is 5 (lines of resolution.) Can you show me how to create
10 stripes with only 10 units when they are adjacent?
If you didn't get it this time, then I give up.<G>
BTW the formula Pixels=picture trace time/rise time is an approximation since
the rise time is never a square wave, therefore the actual formula is quite a
bit more complex but this simplified version is considered by engineers to be
within 10% accurate.
Also, you should note that resolution is generally never mentioned in VTR
specs. Instead, bandwidth is listed and usually any figures you see for
resolution are extrapolated or approximated from the bandwidth using the
equations I stated. Consumer deck specifications use this enterpolation to
assist non engineers in evaluating the equipment and relating it to monitors
and cameras where lines of resolution are the standard.
Don – Sorry to say but you really dug yourself in on this one!
-JE
>>Don – Sorry to say but you really dug yourself in on this one!
Which one: Agreeing with GregP. that his math was correct? or acknowledging
the engineering texts that a pixel is not the same as a line of resolution? Or
do you feel that when "god" says he has ten pages of notes from PHD's that the
the issue is settled. Would it matter if I said I had 11 pages of notes that
say otherwise. Probably not. You believe what you want. It really doesn't
make a bit of difference to me. I could not find your premise in any text or
hand book or catalog of specs. You, or Paul, have not offered any evidence to
support your position. You have not listed your sources for all to examine.
You have not been able to prove your relationship mathematically. On the other
hand I have offered sources, page numbers, quotes, mathematical proofs, and
logic to show the relationship that it requires 2 pixels to produce a line is
true. I have also shown that the use of pixels alone is not a valid
measurement of resolution. I know you have your mind made up so I won't
confuse you anymore with the facts and I really don't care to spend anymore of
my time presenting logic as an argument against Paul's money and power and your
stubbornness. Just show me the logic, the math behind " it only takes one pixel
to produce a line" and I'll be happy to help you change all the math texts and
engineering equations based on this new postulate.
Now let's do something different. How about you taking the position that the
earth is flat and I will prove that it is round. <G> No! I've got a 6 second
animation and ten slides to get out by Monday AM so it's back to work for me.
It has been fun!
PS I found the problem with the Cirrus Logic VLB card. It turns out that the
board had two different speed ram chips mixed that limited its ram to only 512K
with 2 meg chips installed. I put the correct speed drams in and now it's
working fine. full 800×600 at 24 bit. A diagnostic software pointed me to
look at the numbers on the chips.
Don I'm glad you solved your Cirrus Logic card problem. But on the rest of this
stuff, I think you're way off base. What you don't seem to understand or are
unwilling to accept is that I did talk to Quantel's engineers and I did talk
with Textronix's engineers. I trust what Paul told me because we both came to
the same understanding in talking with very knowledgeable video engineers. But
not only that, Paul is a VERY credible person. I don't view him as a "god" but
as a person who makes an especially great effort to be sure of his facts, knows
his own limitations, and when in doubt, seeks the advice and counsel of those
who do know when a question arises. I respect that very much. Sorry but I
really think that your are quite mistaken in your understanding of horizontal
resolution as it applies to this discussion.
-JE
Who is Textronics? Do you mean Tektronics?
Here's another documentation I just looked up that supports one of the
formulas I presented earlier.
On the spec sheet for the Laird Legend CG it states:
Specifications:
Resolution: 35ns (1506 pixels x 483 lines)
Does this mean that it is 1506 lines? No, of course it doesn't because that
again would be a mathematical impossibility at 35ns rise time.
Are you going to say that not only Greg Pyros and I are mistaken but the
engineers at Laird are too?
Why won't you answer my request? Just give me one spec sheet or documentation
from any credible source such as the ones you mentioned so I can look it up. If
you don't have written documentation then how about giving me the name(s) of
the engineers you claim told you this stuff. At least I will have a direct
conflict of data that I can take to Laird, Sony, Chyron, Dubner, Tektronics,
and JVC and let them defend it.
Tomorrow I will call Sony and request a spec sheet on the resolution
capabilities of the DVR-2100. But I seriously doubt that they will quote it's
maximum resolution in pixels. All my other spec sheets list resolution in
bandwidth which you claim is one of the areas that I'm way off base.
Its "Tektronics" Don. But you've been using values from the output of video
keying Character Generators to prove your point, (and I did point this out
earlier) which explains your total misunderstanding of horizontal resolution as
it applies to the discussion at hand. You state that I claimed your values on
bandwidth were off. Check again! You're not very accurate in your accusations,
in spite of the way you call me on the spelling of one letter in a name. Or
your claim that the whole discussion was based on my confusing horizontal and
vertical resolution. Challenge still stands! BTW do me a favor and leave Greg
out of this. If he wants to speak up he will. Obviously he wants to check his
position before he takes a stand.
-JE
Don,
Sorry to jump in like this, but what resolution(s) does the PAR use?
Bailey
Bailey, DPS does not publish this on their spec sheet, but I believe Gus Grubba
said it was 752×480. You would use that pixel ratio when rendering to the PAR.
Don – Since you seem to have chilled a little, I'll give you a hint as to why
the CG resolution comparison is erroneous. Hint 1: Why does the CG have a pixel
resolution of 1506×483? Hint 2: How does a CG anti-alias?
-JE
Bailey:
The PAR resolution is 752 x 480.
Greg Pyros