MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider $a(t)\in\mathbf{L}^{2}(\mathbb{R})$ and $a(t)>0$, is a low pass smooth function with $\hat{a}(f)=0, |f|>f_{max}$. Can we have a upper bound on the following, $\Big|\frac{a'(t)}{a(t)}\Big|$? Using Bernstein's theorem we can upper bound $|a'(t)|$ alone based on $f_{max}$ but how can we upper bound the ratio mentioned here. Any suggestions for it.

share|cite|improve this question
Please edit your question for English and math. What is $a(t)$? is this the same as $x(t)$? What sort of estimate do you want? $x(t)$ can be zero at some points where $x'(t)/x(t)=\infty$. – Alexandre Eremenko Nov 28 '12 at 22:42
Thanks Alexandre, I have corrected the notation and the question. – Neeks Nov 28 '12 at 22:51
You did not tell me what sort of estimate you want. There is no uniform estimate, of course: $a$ can have complex zeros as close as you wish to the real line, condition $a>0$ does not help, and $a'/a$ can be arbitrarily large at some points. – Alexandre Eremenko Nov 29 '12 at 2:07
In the case I am dealing with, only real zeros of $a(t)$ are of interest. But you brought out interesting possibility. Now taking a simple example, $a(t)=1+\mu\sin\omega t$, with $0<\mu<1$ we have,\newline $\Big|\frac{a'(t)}{a(t)}\Big|=\frac{\mu\omega\cos\omega t}{1+\mu\sin\omega t}\leq\frac{\mu\omega}{\sqrt{1-\mu^2}}, \forall t$. So above it can be bounded, though a simple example. Can we have a bound for sum of harmonic sinusoids and generalize it. But polynomials are also entire function and hence band-limited but the bandwidth is too large for it. Please correct me. – Neeks Nov 29 '12 at 6:24

$a(t)=\cos(t+iϵ)\cos(t−iϵ)$ is a low pass signal and $a^\prime/a$ can be as large as you wish at the point $t=\pi/2,$ if $\epsilon$ is sufficiently small.

share|cite|improve this answer
Please see my above comment. – Neeks Nov 29 '12 at 6:26
1. What "real zeros" are you talking about if one of your conditions is that $a(t)>0$ ? 2. Example which I gave shows that in general there is NO upper bound for $a'/a$ under your conditions. What else are you asking? – Alexandre Eremenko Nov 29 '12 at 23:38
sorry for the delayed reply. I had missed the comment. Under what further conditions on a(t) can we have an upper bound on the quantity mentioned? For example, in the example i gave we have a bound dependent on ω – Neeks Dec 5 '12 at 17:23

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.