About fft and distortions analysis using a sound card

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello

Even wen we use a very low distortion sine wave oscillator the biggest problem wen using a fft or distortion analysis software using the sound card are the sound card distortion itself. Is there a way to do a first calibration with the sine wave oscillator alone and keep in memory the sound card + oscillator distortion spectrum and substract that distortion spectrum for the amplifier analysis, wen you do a spectrum and distortion analysis you will mostly have only the amplifier distortion showed in the analysis, so even if it's not a perfect method it would do a better job.

BTW, I have an Ensoniq Audio Pci sound card and use the auxilliary input. I know there is much better sound card but they cost to much.


Thank

Gaetan (Canada
 
Hi,

Distortion of the oscillator and the soundcard for FFT is multiplicative rather than additive. That means that these distortions cannot be subtracted from the total result. These distortions set your measurement limits. So use equipment with has low distortion by itself.

The Rightmark tool is pretty good, simple to use and free. I personally prefer ARTA which is more advanced and can do long FFT’s and averaging that reduces the noise floor. However it is relatively easy to use for this kind of software. The demo version is full functional and not time limited, only you cannot save to files. But you can copy the images from the screen.

Cheers ;)
 
The thing about the Audigy cards is the awful drivers. I used a 2Zs for a couple years and got very good results from it, but each time out was an adventure and a wrestling match trying to get things to work properly. I changed out cards and drivers to an M-Audio Audiophile192 and have not had any problems whatsoever. Raw performance is slightly better, usability is hugely better.
 
Hi SY,

Have looked at the M-Audio Audiophile192. But is appears that it doesn’t do sample rates below 44.1 KHz. Is that correct?

Cheers ;)

Btw. Here is the distortion of the M-Audio Audiophle 2496 at 24 bits:
 

Attachments

  • distortion_24bit.png
    distortion_24bit.png
    15.6 KB · Views: 559
SY said:
The thing about the Audigy cards is the awful drivers. I used a 2Zs for a couple years and got very good results from it, but each time out was an adventure and a wrestling match trying to get things to work properly. I changed out cards and drivers to an M-Audio Audiophile192 and have not had any problems whatsoever. Raw performance is slightly better, usability is hugely better.


That's good to know - I was trying to decide which card I wanted to use. Have you heard anything about their Revolution 7.1 PCI card? It would seem to be a dandy output for an active PC crossover.
 
SY said:
Nope, I'm pretty dumb about these things; the M-Audio was chosen as a result of some testing by John Atwood which showed its performance to be close to the best at a fraction of the price.

By the way, you might want to know Rip Off Press has a two column complete collection of the Freak Bros, including biographical tidbits about Shelton and RO Press.
 
SY said:
the M-Audio was chosen as a result of some testing by John Atwood

Do you mean this? http://www.clarisonus.com/Research Reports/RR001-SoundCardEval/RR001-PCsoundCards.html

I would take it with some salt. Anyway the 500 Hz artefacts he found with the Audiophile 2496 I cannot second. Have a look at the attachment some posts earlier. Ok it was taken at -10 dB. Also the hum he noticed is greatly influenced by noise coming from mains cables entering into the external signal cables. With proper layout this is not an issue (at least not at my side).

End of this week an Audiophile192 will arrive. Let's see what that brings.

;)
 
Pjotr said:
Hi SY,

Have looked at the M-Audio Audiophile192. But is appears that it doesn’t do sample rates below 44.1 KHz. Is that correct?

Cheers ;)

Btw. Here is the distortion of the M-Audio Audiophle 2496 at 24 bits:


My Delta410 has 0.0019% THD and 0.0042 THD+N in the very same test with ARTA.

However, in the measurement setup, I cannot define more than 16 Bits. I can toggle to 24 bits, but when I try to do a measurement, Windows XP gives me a wave out error message. Toggling between No dither and 18 / 20 bit differ makes no difference at all.

Wonder if the Audiophile has better converters or whether the difference is due to the word length.
 
Hi capslock,

Did you install the latest drivers for it?
Did you select the M-Audio Delta driver in the ARTA hw set-up panel?
Did you use “extended” in the ARTA hw set-up panel?

I have no Delta 410 but with all other M-Audio cards I have used, the M-Audio drivers were rock solid (with W2K).

;)
 
You can increase resolution by changing the approach - Intermodulation products give information about the amplifier's distortion charateristics and the testing can be arranged to avoid super low harmonic distortion source requirement

separate DAC channels for each sine component, buffered, then cleanly added together gives a test signal with low intermod product content

as long as harmonics of the well chosen test tone frequency pairs/triples... are reasonably low they won't interfere with measurement of the IM products created by the tested amp's distortions

it is still good to filter/null the major test tone components at the ADC input

with enough DAC channels you could create digitally phase/magnitude scaled null signals for the IM test
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.