Forums

audio clock jitter effect

Started by qbq7...@yahoo.com.cn July 23, 2009
I have a question , every one must know that the jitter will effect the performance of audio ,such as SNR . However , if there is no data input to audio dac (you also can think the data are also 0), would it still effect it? such as noise floor .
QBQ-

> I have a question , every one must know that the jitter will effect
> the performance of audio ,such as SNR . However ,
> if there is no data input to audio dac (you also can think the data
> are also 0), would it still effect it? such as noise floor .

If by "jitter" you are referring strictly to slight variations in sampling time, then there would be no effect on:

-any constant dc input, including zero

-100% random (white) noise

But for signals with audible variation in frequency, then SNR would be affected, depending on how much jitter.

-Jeff
QBQ:

> Thanks for your reply!
> You mean that if the input pcm signal is zero , there will be
> no effect from clock jitter , am I right ?

If the sample data seen by the A/D is truly zero, then yes. But in practice it's hard to make that actually happen.
There will be noise from the input circuitry, and A/D converter itself.

> If I'm right , there is one phenomenon that I can not explain. I have
> made an experiment  by changing crystal to
> active crystal ,  I found that the noise floor has been improved.
> What would you think about it ?

It sounds like the active oscillator is more accurate, with less jitter. If you have an appropriate digital scope,
you could make some jitter, or "eye", measurements, and compare.

-Jeff

PS. Please post to the group, not to me.

> --- 09年7月23日,周四, Jeff Brower 写道:
> 发件人: Jeff Brower
> 主题: Re: [audiodsp] audio clock jitter effect
> 收件人: q...@yahoo.com.cn
> 抄送: a...
> 日期: 2009年7月23日,周四,下午11:41
>  
>
> QBQ-
>
>> I have a question , every one must know that the jitter will effect
>> the performance of audio ,such as SNR . However ,
>> if there is no data input to audio dac (you also can think the data
>> are also 0), would it still effect it? such as noise floor .
>
> If by "jitter" you are referring strictly to slight variations in sampling time, then there would be no effect on:
>
> -any constant dc input, including zero
>
> -100% random (white) noise
>
> But for signals with audible variation in frequency, then SNR would be affected, depending on how much jitter.
>
> -Jeff
>