THD% - THD+N% vs Frequency and THD% - THD+N% vs Power
Posted: Thu Oct 03, 2013 9:12 amTHD% - THD+N% vs Frequency and THD% - THD+N% vs Power
I've tried a comparison between the MI Pro3.1 and the top FFT analyzer of Audio Precision. For this reason I had bought the sound card ASUS Xonar Essence STX, because its packaging includes a booklet with measurements of the card with an Audio Precision analyzer. Trying to achieve consistent results with MI Pro3.1 i found that this is done with small FFT sizes, from 2048 to 4096. Particularly, if the record length is set exactly twice the size of the FFT, then the spectrum analysis is much cleaner. E.g. If you set the FFT size to 2048 and the record length in 4096, then the resulting spectrum after the analysis shows very clearly only the main harmonics (I mean those who have a considerable width) while the very short interim harmonics disappear almost completely. I expect a comment to this by Virtins.
Making a comparison between the posibilities offered from the MI Pro3.1 and from the Audio Precision, I have found two significant lacks in your software:
1) FFT analysis of THD% and THD + N% versus Frequency (usually as a function of power in power amplifiers)
2) FFT analysis of THD% and THD + N% versus Power with input a sinusoidal signal of 1KHz, increased gradually to drive the output up to its maximum power (usually as a function of load in power amplifiers)
I must tell you that the above measurements in audio devices are very important for the good promotion of a product as these are related with the final consumer. If you read the test e.g. of an amplifier in any audio magazine, you will see these measurements anyway. Also, all major manufacturers of audio devices do the same as above.
I would like to ask you to do something for this. Please do not take it negligently, it is mandatory. I have read comments from many people interested for the MI Pro3.1 about this lack.
Kind Regards
Fotios Anagnostou
Greece
Post at 1-10-2009 05:26
0) The record length being twice of the FFT size would only mean that the result will be the average result of the two FFT segments. It is difficult to comment further without further information such as other parameters and screenshots.
1) THD vs Frequency, THD+N vs Frequency can be done via Device Test Plan using frequency stepped sine wave as the stimulus. We will provide some sample test plans on this in the next version.
2) THD vs Power, THD+N vs Power can be done via Device Test Plan using amplitude stepped sine wave. Note that the power variable can be obtained from the power-related DDP (derived data point) in MI, and you can specify the load factor for them via [setting]>[calibration]. We will provide some sample test plans on this in the next version.
virtins
Post at 2-10-2009 18:39
Thank you for your instructions.
Indeed, from yesterday by looking some deeper in DTP i perceived that these measurements can be obtained from MI Pro3.1, simply by programming the requested parameters. Unfortunatelly those parameters are so many, so as i believe that only a small percentage of MI Pro3.1 users will be available to programming these. Me, i am patient but i am only one. The 90% of users, they prefer to get ready offered panels for these measurements. It is not so difficult for you to make 2 or 3 ready panels.
If you have the kindness to take a look in my first finished power amplifier presented in my web page www.eal.gr you will have the opportunity for a better comprehension of the lot of work needed for finishing a project, and that we don't have excess of time to spend in calibration, programming etc of our measuring instruments or repairing our tools etc.
Kind Regards
Fotios Anagnostou
Post at 2-10-2009 19:38
I've tried a comparison between the MI Pro3.1 and the top FFT analyzer of Audio Precision. For this reason I had bought the sound card ASUS Xonar Essence STX, because its packaging includes a booklet with measurements of the card with an Audio Precision analyzer. Trying to achieve consistent results with MI Pro3.1 i found that this is done with small FFT sizes, from 2048 to 4096. Particularly, if the record length is set exactly twice the size of the FFT, then the spectrum analysis is much cleaner. E.g. If you set the FFT size to 2048 and the record length in 4096, then the resulting spectrum after the analysis shows very clearly only the main harmonics (I mean those who have a considerable width) while the very short interim harmonics disappear almost completely. I expect a comment to this by Virtins.
Making a comparison between the posibilities offered from the MI Pro3.1 and from the Audio Precision, I have found two significant lacks in your software:
1) FFT analysis of THD% and THD + N% versus Frequency (usually as a function of power in power amplifiers)
2) FFT analysis of THD% and THD + N% versus Power with input a sinusoidal signal of 1KHz, increased gradually to drive the output up to its maximum power (usually as a function of load in power amplifiers)
I must tell you that the above measurements in audio devices are very important for the good promotion of a product as these are related with the final consumer. If you read the test e.g. of an amplifier in any audio magazine, you will see these measurements anyway. Also, all major manufacturers of audio devices do the same as above.
I would like to ask you to do something for this. Please do not take it negligently, it is mandatory. I have read comments from many people interested for the MI Pro3.1 about this lack.
Kind Regards
Fotios Anagnostou
Greece
Post at 1-10-2009 05:26
0) The record length being twice of the FFT size would only mean that the result will be the average result of the two FFT segments. It is difficult to comment further without further information such as other parameters and screenshots.
1) THD vs Frequency, THD+N vs Frequency can be done via Device Test Plan using frequency stepped sine wave as the stimulus. We will provide some sample test plans on this in the next version.
2) THD vs Power, THD+N vs Power can be done via Device Test Plan using amplitude stepped sine wave. Note that the power variable can be obtained from the power-related DDP (derived data point) in MI, and you can specify the load factor for them via [setting]>[calibration]. We will provide some sample test plans on this in the next version.
virtins
Post at 2-10-2009 18:39
Thank you for your instructions.
Indeed, from yesterday by looking some deeper in DTP i perceived that these measurements can be obtained from MI Pro3.1, simply by programming the requested parameters. Unfortunatelly those parameters are so many, so as i believe that only a small percentage of MI Pro3.1 users will be available to programming these. Me, i am patient but i am only one. The 90% of users, they prefer to get ready offered panels for these measurements. It is not so difficult for you to make 2 or 3 ready panels.
If you have the kindness to take a look in my first finished power amplifier presented in my web page www.eal.gr you will have the opportunity for a better comprehension of the lot of work needed for finishing a project, and that we don't have excess of time to spend in calibration, programming etc of our measuring instruments or repairing our tools etc.
Kind Regards
Fotios Anagnostou
Post at 2-10-2009 19:38