Filmscanners mailing list archive (firstname.lastname@example.org)
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[filmscanners] Re: Density vs Dynamic range
I think Laurie is right, and to take a break. However as I am away for a
week and hence an enforced break, here are some frustrated responses of
mine to a couple of points in your last post, I am ignoring your less
temperate and more insulting(!) bits, just concentrating on a couple of
things that I think might help to crystallize the problem.
> > Austin's number = (max - min)/noise
> > = (max-noise)/noise [here]
> > = 8,599/1
> > = 8,599
>I am very sorry to say, but you are playing games. I have CLEARLY stated
>that you are using the equation incorrectly, and you chose to ignore that.
>That is dishonest.
>The CORRECT use of the equation is:
> = (max-MIN)/noise
> = (8,600-0)/1
> = 8,600
Now this wasn't being dishonest, it was my *honest* attempt to interpret
things the way you do. I wrote this based on your previous post which
seems to be different - here it is again:
Austin said some time ago:
>The mistake is you are using noise as your divisor instead of min signal
>level. Your equations should be:
>DR = 10log10(max-min)/noise)
>where in case 1, range = 1-1000mV, noise 1mV so 1000-1 = 999/1 = 999
Do you see why I am confused? Can you tell me which of these two uses of
your equations is correct? And can you tell me again, how you define this
Is MIN = noise, or is it zero? Why? Why is it different in these two
examples? If MIN = zero, why? (since this signal level is impossible).
You can see why I have had a problem interpreting what you are saying I
think. I am not saying you are WRONG, but I am saying that from where I
sit, these two posts of yours are contradictory. Obviously they are not
contradictory to you, but to me they are. So without getting steamed up,
can you just explain to me what is the difference betw these two, given
that in both cases the minimum of the range is determined by noise.
Or have you changed your view of what the top of this equation means?
>Sorry, Julian, I can not continue this conversation with you. What I have
>said is entirely correct, and you seem to have a problem handling that, and
>need to IGNORE the correct use of terms and equation, and that in dishonest
>and makes your analysis not credible.
>I am not even going to waste my time reading the rest of what you wrote. It
>is PLAINLY clear that everything I have said is ENTIRELY correct, and YOU
>NOW AGREE with it, as what I have said gives the EXACT same results as what
>you are saying, it's just that you won't admit that, and need to modify my
>statements to somehow show I am wrong, when I am not.
It is a pity you won't continue, especially since, as you have done before,
you seem to be using this as a reason NOT to reply to the hard bits of my
post. I really wish you would reply to ALL my post, even when the answers
may not please you.
>Please don't bother answering any of my questions, or responding to this
>post. You have unfortunately shown that you have no intention of being
>"honest" in this discussion. I can only guess so that you do not come out
>as having been wrong about what I've been saying...since it is not obvious
>that your answers agree with mine, given the same numbers. There is NO use
>in continuing this.
Sorry Austin - that you think I have been dishonest. I may have been
confused as to what you are saying, and I think anyone, engineeringly
literate or not, would have been equally confused by your apparently
self-contradictory statements. I admit to being completely bemused and
frustrated by your posts, but not by my own understanding, and I admit to
believing that I am trying to accommodate myself to a moving target in this
argument, as you *seem* to constantly change your mind. I am tying myself
in knots here trying to re-phrase everything in your terms, but when I do
you tell me I am misinterpreting or dishonest. I should have just
continued to talk in my own terms, although I didn't see how that would
resolve. But I assure you, I have been scrupulously " honest".
I really would like you to address what I wrote yesterday on the examples
of the log amp and the detector system. Otherwise it is too easy to draw
the wrong conclusion as to why you won't address them. But in the
meantime, in one last desperate attempt to get to one of the hearts of this
matter - - -
As you can tell, I find your discussion is still hopelessly confused in
terms of its definitions. Can you please on one piece of cyber paper, that
is in one place, define what your terms are:
And can you relate these to minimum discernable signal and noise? Please
draw on your celebrated diagram what this minimum signal is, because it
apparently is not what you call "smallest", and since it is not on your
diagram I don't see that you can say it is defined. If it is the line at
the bottom, then you are in fact talking about a DC offset, which is what I
believed when I hopefully wrote my previous post, but you told me this was
wrong. So ... what?
I need a diagram showing these values:
Smallest discernable signal
and which of these signals is equal to which, and what determines
them. Particularly what determines Minimum signal.
I do NOT need your current diagram which shows two terms that are not
included in any definition anywhere and which is therefore useless as a
definition diagram - "largest" and "smallest" signals. Please use ONLY the
terminology in the definition of DR.
I think one of our basic difficulties is that I don't believe that we need
this thing called minimum signal, in addition to noise or minimum
discernable signal, and in the absence of DC offsets. And I think your
introduction of this extra value is what causes many of the
difficulties. If minimum signal is something that you introduce as an
arbitrary range limit, then as I say in other places, why? and if so, you
are no longer measuring the dynamic range or resolution of the system or
signal, you are measuring the dynamic range or resolution of some arbitrary
part of the signal, maybe the part that you choose to digitise. If we are
in fact talking about the part of the range that you digitise, then, once
again, you have now defined the dynamic range of your ADC, not the original
signal. And if you choose to do that, then it just means that min
discernable signal of the ADC is not equal to noise, as I have been
suggesting is true of any other non-linear system. It doesn't mean you
need a new term, you can still calculate DR of the ADC in terms of only Max
signal and Min discernable signal. You can still calculate resolution by
dividing that dynamic range into noise -sized chunks. In neither case do
you need to define this new term MINIMUM signal. So, to get to the bottom
of this, WHAT IS MINIMUM SIGNAL and why is it necessary to add it into the
If all else fails, I look forward to reading your paper.
Unsubscribe by mail to email@example.com, with 'unsubscribe
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or