hi all,
i am carrying out a small test in where i need to introduce 1
db of noise. Since i don't have any noise generator can i decrease the
input signal by 1 db.Is it equivalent to injecting noise of 1 db?
I am doing the test in awgn channel. Does it also hold good for
multipath channel models?
Thanks
particlereddy (straydog)
decreasing signal power..does it mean.increasing noise power
Started by ●July 9, 2008
Reply by ●July 9, 20082008-07-09
PARTICLEREDDY (STRAYDOG) wrote:> hi all, > i am carrying out a small test in where i need to introduce 1 > db of noise. Since i don't have any noise generator can i decrease the > input signal by 1 db.Is it equivalent to injecting noise of 1 db? > I am doing the test in awgn channel. Does it also hold good for > multipath channel models? > > Thanks > particlereddy (straydog)That depends. If you're going into real hardware that might overload with higher signal and/or noise power then no, increasing one power level is different than decreasing the other. If you're going into hardware that does ratiometric detection, and that you trust is operating in its linear region then yes, you're fine. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●July 9, 20082008-07-09
Tim Wescott wrote:> PARTICLEREDDY (STRAYDOG) wrote: >> hi all, >> i am carrying out a small test in where i need to introduce 1 >> db of noise. Since i don't have any noise generator can i decrease the >> input signal by 1 db.Is it equivalent to injecting noise of 1 db? >> I am doing the test in awgn channel. Does it also hold good for >> multipath channel models? >> >> Thanks >> particlereddy (straydog) > > That depends. > > If you're going into real hardware that might overload with higher > signal and/or noise power then no, increasing one power level is > different than decreasing the other. > > If you're going into hardware that does ratiometric detection, and that > you trust is operating in its linear region then yes, you're fine.Tim, You threw me for a loop here. I thought the question was meaningless. You apparently found a context that gives it meaning. Here;'s my take: Let's start out with a 1-volt, noise-free signal (0 dBv). I want to introduce 1 dB of noise. The first question is relative to what? Relative to the signal seems to be the only useful measure. OK. We introduce enough noise to raise the level to 1 dBv, or 1.122 volts. Suppose that instead of injecting noise, we reduce the signal level to .9 dBv. That gives us .902 volts and no noise at all. Where's the equivalence? Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●July 9, 20082008-07-09
Jerry Avins wrote:> Tim Wescott wrote: >> PARTICLEREDDY (STRAYDOG) wrote: >>> hi all, >>> i am carrying out a small test in where i need to introduce 1 >>> db of noise. Since i don't have any noise generator can i decrease the >>> input signal by 1 db.Is it equivalent to injecting noise of 1 db? >>> I am doing the test in awgn channel. Does it also hold good for >>> multipath channel models? >>> >>> Thanks >>> particlereddy (straydog) >> >> That depends. >> >> If you're going into real hardware that might overload with higher >> signal and/or noise power then no, increasing one power level is >> different than decreasing the other. >> >> If you're going into hardware that does ratiometric detection, and >> that you trust is operating in its linear region then yes, you're fine. > > Tim, > > You threw me for a loop here. I thought the question was meaningless. > You apparently found a context that gives it meaning. Here;'s my take: > > Let's start out with a 1-volt, noise-free signal (0 dBv). I want to > introduce 1 dB of noise. The first question is relative to what? > Relative to the signal seems to be the only useful measure. OK. We > introduce enough noise to raise the level to 1 dBv, or 1.122 volts. > > Suppose that instead of injecting noise, we reduce the signal level to > .9 dBv. That gives us .902 volts and no noise at all. Where's the > equivalence? > > JerryI misread his dang post! You're right, it is meaningless if you read it correctly. It's just so close to being sensible... What I _thought_ he was asking was if increasing SNR by decreasing signal (into a real, and therefore noisy receiver) would have the same effect as injecting noise. The answer, if you mind your 'p's and 'q's, is yes. But you can't equate injecting 1dB of noise to dropping the signal by 1dB, if for no other reason than no one answered the question "1dB relative to what?". But with his question as written, my response is nothing but fertilizer, to make bigger and tastier tomatoes. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
Reply by ●July 9, 20082008-07-09
Tim Wescott wrote:> PARTICLEREDDY (STRAYDOG) wrote: >> hi all, >> i am carrying out a small test in where i need to introduce 1 >> db of noise.I misread this the first time around (thanks Jerry). You can't introduce 1dB of noise because such a thing doesn't exist. A dB is a measure of a ratio, so you could increase your noise by 1dB, or introduce 1dBm (1mW * 10^(1/10)) of noise power, or introduce 1dB of noise relative to your signal (but this latter won't necessarily decrease your SNR by 1dB if there is already significant noise -- think about it).>> Since i don't have any noise generator can i decrease the >> input signal by 1 db.Is it equivalent to injecting noise of 1 db?You can decrease your signal by 1dB and decrease your SNR by 1dB. But the phrase "injecting noise of 1dB" is meaningless, so nothing is equivalent to that.>> I am doing the test in awgn channel. Does it also hold good for >> multipath channel models?I believe so.>> >> Thanks >> particlereddy (straydog) > > That depends. > > If you're going into real hardware that might overload with higher > signal and/or noise power then no, increasing one power level is > different than decreasing the other. > > If you're going into hardware that does ratiometric detection, and that > you trust is operating in its linear region then yes, you're fine. >Correction: If you're going into hardware that does ratiometric detection, and you can reliably decrease your signal so that it is overwhelmed by your front-end noise (this is not trivial with a good radio receiver), and you can measure your SNR by other means, and if my above caveats still apply, then yes, you're fine. -- Tim Wescott Wescott Design Services http://www.wescottdesign.com Do you need to implement control loops in software? "Applied Control Theory for Embedded Systems" gives you just what it says. See details at http://www.wescottdesign.com/actfes/actfes.html
