From: wazerface on
On Nov 7, 6:02 pm, Eric Jacobsen <eric.jacob...(a)ieee.org> wrote:
> On 11/7/2009 12:56 PM, Greg Heath wrote:
>
> > On Nov 5, 1:04 am, "BCLIM"<boonchun_...(a)yahoo.com>  wrote:
> >> Hi,
> >>     I have a doubt on jitter generation. Wonder is the method to generate
> >> jitter signal is that same as dithering generation method? Understand that
> >> both are usign random signal generation with different distribution. For
> >> example, triangular, rectangular etc.
>
> > I always thought that for a signal defined over a finite length of
> > time,
> > jittering added a randomness to the amplitude whereas dithering
> > added a randomness to the starting time.
>
> > Hope this helps.
>
> > Greg
>
> I think it'd help for the OP to clarify what he means by his
> terminology.   In my experience jitter usually means random fluctuation
> in period, (usually sampling period), while dithering usually means
> adding small random values to an input to reduce quantization noise.
>
> It's also not clear whether the OP is asking about unintentional or
> intentional jitter.   Some clocking systems add jitter in order to
> reduce spurious emissions related to the clock frequency.   Is that the
> topic of the question?
>
> --
> Eric Jacobsen
> Minister of Algorithms
> Abineau Communicationshttp://www.abineau.com

My understanding is that dithering doesn't reduce quantization noise
(power) in general, but de-correlates it.
From: Jerry Avins on
Greg Heath wrote:
> On Nov 7, 5:16 pm, Jerry Avins <j...(a)ieee.org> wrote:
>> Greg Heath wrote:
>>> On Nov 5, 1:04 am, "BCLIM" <boonchun_...(a)yahoo.com> wrote:
>>>> Hi,
>>>> I have a doubt on jitter generation. Wonder is the method to generate
>>>> jitter signal is that same as dithering generation method? Understand that
>>>> both are usign random signal generation with different distribution. For
>>>> example, triangular, rectangular etc.
>>> I always thought that for a signal defined over a finite length of
>>> time,
>>> jittering added a randomness to the amplitude whereas dithering
>>> added a randomness to the starting time.
>>> Hope this helps.
>> It probably hurts. Jitter randomizes the timing of pulse edges. When you
>> look at such a signal on an oscilloscope, the trace jitters.
>
> Given your additional comment. it helps emphasize that
> terminology is application dependent.
>
> In the fields of classification and regression, the terminology
> is as I have indicated.
>
> Hope this helps.

I think that whoever adopted those terms in their new meanings
misunderstood their meanings in their original fields of application.
Jitter, as I mentioned, was named for its visible effect on oscilloscope
traces. Dither entered the technical lexicon from mechanical
servomechanism design, where vibration applied orthogonally to the
direction of travel eliminated the effect of stick-slip friction (aka
sticktion). (It is the same effect as tapping an aneroid barometer to
get an accurate reading.) It's extension to smoothing the sampling
process, thereby extending the resolvable limit seems natural.

Jerry
--
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������
From: Eric Jacobsen on
On 11/8/2009 9:07 AM, wazerface(a)gmail.com wrote:
> On Nov 7, 6:02 pm, Eric Jacobsen<eric.jacob...(a)ieee.org> wrote:
>> On 11/7/2009 12:56 PM, Greg Heath wrote:
>>
>>> On Nov 5, 1:04 am, "BCLIM"<boonchun_...(a)yahoo.com> wrote:
>>>> Hi,
>>>> I have a doubt on jitter generation. Wonder is the method to generate
>>>> jitter signal is that same as dithering generation method? Understand that
>>>> both are usign random signal generation with different distribution. For
>>>> example, triangular, rectangular etc.
>>> I always thought that for a signal defined over a finite length of
>>> time,
>>> jittering added a randomness to the amplitude whereas dithering
>>> added a randomness to the starting time.
>>> Hope this helps.
>>> Greg
>> I think it'd help for the OP to clarify what he means by his
>> terminology. In my experience jitter usually means random fluctuation
>> in period, (usually sampling period), while dithering usually means
>> adding small random values to an input to reduce quantization noise.
>>
>> It's also not clear whether the OP is asking about unintentional or
>> intentional jitter. Some clocking systems add jitter in order to
>> reduce spurious emissions related to the clock frequency. Is that the
>> topic of the question?
>>
>> --
>> Eric Jacobsen
>> Minister of Algorithms
>> Abineau Communicationshttp://www.abineau.com
>
> My understanding is that dithering doesn't reduce quantization noise
> (power) in general, but de-correlates it.

Yes, and the decorrelation allows an increase of SNR with appropriate
filtering, so it can be used to effectively reduce quantization noise.
But you're right, dithering by itself doesn't reduce quantization noise.

--
Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.abineau.com
From: Randy Yates on
Eric Jacobsen <eric.jacobsen(a)ieee.org> writes:
> [...]
> Yes, and the decorrelation allows an increase of SNR with appropriate
> filtering,

What filtering is that? If oversampling is involved, there can be
filtering. And perhaps there may be a "preference" in frequencies (e.g.,
the emphasis of the human ear in frequencies near 4 kHz), which would
permit some type of filter. But in general, there is no filtering that
can be done and dither actually increases the total quantization-related
noise power. That's not to say dithering isn't worth doing, however,
bacause the decorrelating is a distinct advantage.
--
Randy Yates % "How's life on earth?
Digital Signal Labs % ... What is it worth?"
mailto://yates(a)ieee.org % 'Mission (A World Record)',
http://www.digitalsignallabs.com % *A New World Record*, ELO
From: Eric Jacobsen on
On 11/8/2009 12:52 PM, Randy Yates wrote:
> Eric Jacobsen<eric.jacobsen(a)ieee.org> writes:
>> [...]
>> Yes, and the decorrelation allows an increase of SNR with appropriate
>> filtering,
>
> What filtering is that? If oversampling is involved, there can be
> filtering.

Yup. And filtering can drop the noise floor far further with dithering
than without.

> And perhaps there may be a "preference" in frequencies (e.g.,
> the emphasis of the human ear in frequencies near 4 kHz), which would
> permit some type of filter. But in general, there is no filtering that
> can be done and dither actually increases the total quantization-related
> noise power. That's not to say dithering isn't worth doing, however,
> bacause the decorrelating is a distinct advantage.

Pulling a small, narrow signal (say in the 1 LSB region) out of a
composite signal is easily possible with dithering. Without dithering,
it's very often not possible at all. Dithering can make a huge
difference in reducing effective quantization noise.

This place is going to get really tiresome if every response to a
question has to include every possible nuance in the answer.

--
Eric Jacobsen
Minister of Algorithms
Abineau Communications
http://www.abineau.com
First  |  Prev  |  Next  |  Last
Pages: 1 2 3 4 5 6
Prev: STUPIDENT::Re: Floting Point Saturation
Next: MPEG-Xtreme