I have an application and am trying to find practical info on how well an ADC will work here. This may require ADC functions implemented in a DSP. An audio CD contains a continuous analog tone (a pure sinusoid) that I need to measure in amplitude and phase as accurately as practicable. I need to be able to measure the original's amplitude to an accuracy of 1% and phase to an accuracy of better than 1 degree. Does it sound like sampling the original tone with a 44.1 ksample/sec 16-bit ADC will give the amplitude and phase measurement accuracies I need? Or should I plan to start out with higher sample rate and/or ADC resolution? Is it likely that a custom DSP implementation may be needed? Can anyone recommend a Website with application info on ADCs for this type of application? Henry
ADC Performance Question
Started by ●April 5, 2008
Reply by ●April 5, 20082008-04-05
On Apr 4, 9:34�pm, "Henry VIII" <donteventh...@emailingmehere.com> wrote:> I have an application and am trying to find practical info on how well an > ADC will work here. �This may require ADC functions implemented in a DSP. > > An audio CD contains a continuous analog tone (a pure sinusoid) that I need > to measure in amplitude and phase as accurately as practicable. �I need to > be able to measure the original's amplitude to an accuracy of 1% and phase > to an accuracy of better than 1 degree.Phase relative to what?
Reply by ●April 5, 20082008-04-05
"Henry VIII" <donteventhink@emailingmehere.com> wrote in message news:0jDJj.930$GO4.269@newssvr19.news.prodigy.net...>I have an application and am trying to find practical info on >how well an > ADC will work here. This may require ADC functions > implemented in a DSP. > > An audio CD contains a continuous analog tone (a pure > sinusoid) that I need > to measure in amplitude and phase as accurately as > practicable. I need to > be able to measure the original's amplitude to an accuracy of > 1% and phase > to an accuracy of better than 1 degree. > > Does it sound like sampling the original tone with a 44.1 > ksample/sec 16-bit ADC will give the amplitude and phase > measurement > accuracies I need?Wait a minute! Isn't the data on the CD already digitized?
Reply by ●April 5, 20082008-04-05
On Sat, 5 Apr 2008 07:01:31 -0400, "John E. Hadstate" <jh113355@hotmail.com> wrote:>Wait a minute! Isn't the data on the CD already digitized?You beat me to it! Extract the digital data directly from the CD and work your mathematical magic on that. Most CD burning software will allow you to extract a .wav file from an audio CD. Ignore the .wav header info (which is the first 44 bytes, in the simplest of cases, but may be MUCH more) and all you have is pure 16-bit PCM audio stored L|R|L|R. Greg
Reply by ●April 5, 20082008-04-05
On Sat, 5 Apr 2008 07:01:31 -0400, "John E. Hadstate" <jh113355@hotmail.com> wrote:> >"Henry VIII" <donteventhink@emailingmehere.com> wrote in >message news:0jDJj.930$GO4.269@newssvr19.news.prodigy.net... >>I have an application and am trying to find practical info on >>how well an >> ADC will work here. This may require ADC functions >> implemented in a DSP. >> >> An audio CD contains a continuous analog tone (a pure >> sinusoid) that I need >> to measure in amplitude and phase as accurately as >> practicable. I need to >> be able to measure the original's amplitude to an accuracy of >> 1% and phase >> to an accuracy of better than 1 degree. >> >> Does it sound like sampling the original tone with a 44.1 >> ksample/sec 16-bit ADC will give the amplitude and phase >> measurement >> accuracies I need? > >Wait a minute! Isn't the data on the CD already digitized?Hi, Yep, that would be nice if Henry VIII can extract digital samples without using an A/D converter. And it seems to me that 16-bit samples would be sufficient for him to achieve the amplitude- and phase-measurement accuracies that he needs. [-Rick-]
Reply by ●April 5, 20082008-04-05
Unfortunately I only have access to the analog audio because the CD audio source is in a separate location. Regarding the question "phase relative to what?," to be specific I need to measure the precise time of zero-crossing of the original signal. So I have to measure amplitude within 1% and zero-crossing time to within one degree. Or if zero-crossing time measurement is not precise enough, I could measure positive and negative peaks, then interpolate the mid-point. Or, any other suggestion?? It seems to me that 44.1 ksample/sec sampling and 16-bit resolution should achieve these accuracies, but I'm hoping someone can confirm this so I don't start out with the wrong implementation. Thanks. Henry "John E. Hadstate" <jh113355@hotmail.com> wrote in message news:2pWdnagyG9KtwWranZ2dnUVZ_hudnZ2d@supernews.com...> > "Henry VIII" <donteventhink@emailingmehere.com> wrote in message > news:0jDJj.930$GO4.269@newssvr19.news.prodigy.net... >>I have an application and am trying to find practical info on how well an >> ADC will work here. This may require ADC functions implemented in a DSP. >> >> An audio CD contains a continuous analog tone (a pure sinusoid) that I >> need >> to measure in amplitude and phase as accurately as practicable. I need >> to >> be able to measure the original's amplitude to an accuracy of 1% and >> phase >> to an accuracy of better than 1 degree. >> >> Does it sound like sampling the original tone with a 44.1 >> ksample/sec 16-bit ADC will give the amplitude and phase measurement >> accuracies I need? > > Wait a minute! Isn't the data on the CD already digitized? > > > >
Reply by ●April 5, 20082008-04-05
On Fri, 04 Apr 2008 21:34:06 -0700, Henry VIII wrote:> I have an application and am trying to find practical info on how well > an ADC will work here. This may require ADC functions implemented in a > DSP. > > An audio CD contains a continuous analog tone (a pure sinusoid) that I > need to measure in amplitude and phase as accurately as practicable. I > need to be able to measure the original's amplitude to an accuracy of 1% > and phase to an accuracy of better than 1 degree. > > Does it sound like sampling the original tone with a 44.1 ksample/sec > 16-bit ADC will give the amplitude and phase measurement accuracies I > need? Or should I plan to start out with higher sample rate and/or ADC > resolution? Is it likely that a custom DSP implementation may be > needed? > > Can anyone recommend a Website with application info on ADCs for this > type of application? > > HenryWhat Ron said, what John said. I think you need to tell us more about what you're really trying to do. If you can, just tell us what your end goal is, and we'll be in a much better position to help out. AFAIK a 16-bit ADC and 44.2kHz sampling rate will match the capabilities of the CD; assuming that your samples aren't perfectly synchronized to the CD you would gain a little bit of accuracy with a higher sampling rate. Ditto, if your bits weren't perfectly aligned with the DC output you may gain a bit from a more precise ADC. But you've left out some necessary information to tell if you can make your amplitude and phase estimates to your required precision. You need to know the expected amplitude of the signal with respect to the full-scale of the ADC (or the digital full-scale if you take it straight off the CD) so you have a figure for the quantization noise. You need to know the frequency of the sinusoid, and the duration of the sample. You need to know the level and frequency content of any other interfering signals. If you're really going to do this with an ADC, then the analog electronics leading up to the ADC are critical -- not only must you concern yourself with noise, but you need to worry about phase and gain variations with frequency; this requirement puts astonishingly strict requirements on component variation if you're going to use 'normal' anti- aliasing techniques, or raises your required circuit bandwidth astonishingly high if you're not. -- Tim Wescott Control systems and communications consulting http://www.wescottdesign.com Need to learn how to apply control theory in your embedded system? "Applied Control Theory for Embedded Systems" by Tim Wescott Elsevier/Newnes, http://www.wescottdesign.com/actfes/actfes.html
Reply by ●April 5, 20082008-04-05
Henry VIII wrote:> An audio CD contains a continuous analog tone (a pure sinusoid) that I need > to measure in amplitude and phase as accurately as practicable. I need to > be able to measure the original's amplitude to an accuracy of 1% and phase > to an accuracy of better than 1 degree.The task is trivial. Is it what your majesty got as the homework assignment? VLV
Reply by ●April 5, 20082008-04-05
Rick Lyons wrote:> On Sat, 5 Apr 2008 07:01:31 -0400, "John E. Hadstate" > <jh113355@hotmail.com> wrote: > >> "Henry VIII" <donteventhink@emailingmehere.com> wrote in >> message news:0jDJj.930$GO4.269@newssvr19.news.prodigy.net... >>> I have an application and am trying to find practical info on >>> how well an >>> ADC will work here. This may require ADC functions >>> implemented in a DSP. >>> >>> An audio CD contains a continuous analog tone (a pure >>> sinusoid) that I need >>> to measure in amplitude and phase as accurately as >>> practicable. I need to >>> be able to measure the original's amplitude to an accuracy of >>> 1% and phase >>> to an accuracy of better than 1 degree. >>> >>> Does it sound like sampling the original tone with a 44.1 >>> ksample/sec 16-bit ADC will give the amplitude and phase >>> measurement >>> accuracies I need? >> Wait a minute! Isn't the data on the CD already digitized? > > Hi, > Yep, that would be nice if Henry VIII can > extract digital samples without using an > A/D converter. And it seems to me that 16-bit > samples would be sufficient for him to achieve > the amplitude- and phase-measurement accuracies > that he needs.To repeat Ron's question, phase relative to what? The phase relative to an external reference will depend on exactly when the CD is started. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●April 5, 20082008-04-05
Henry VIII wrote:> Unfortunately I only have access to the analog audio because the CD audio > source is in a separate location. > > Regarding the question "phase relative to what?," to be specific I need to > measure the precise time of zero-crossing of the original signal. So I have > to measure amplitude within 1% and zero-crossing time to within one degree. > Or if zero-crossing time measurement is not precise enough, I could measure > positive and negative peaks, then interpolate the mid-point. Or, any other > suggestion?? > > It seems to me that 44.1 ksample/sec sampling and 16-bit resolution should > achieve these accuracies, but I'm hoping someone can confirm this so I don't > start out with the wrong implementation."The precise time of zero crossing." What then is your time reference? There will be no zero crossing at all until the CD starts. After that, the crossing timess will be determined by when it starts. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
