Lobby > Amp and DAC Measurements

Standard Distortion Spectrum Measurements For DACS (Changstar)

(1/1)

firev1:
Some points for DAC tests or low distortion measurements. I blew up my measurement channel on my EMU so nothing from me for now :# The aim for this suite of tests is to be somewhat comparable to those on Stereophile and also those using RMAA and such. Suggestions are welcome for those knowledgeable of the subject.

Scaling: Something like -60 to -160 db for low distortion test. Actually anything that makes easy comparison is fine.

Windows: Anything that lowers spectrum skirting. For ARTA/Wavespectra, that means 7 term windows. Highest FFT lengths with averaging to be used.

Filter reconstruction test with 19khz overlayed: I just emulate the ones posted on Stereophile.

Tone generation: There is a tool called Wavegene which is totally free and handy for tone generation. Just don't use in inbuilt player. It does up to 8 tones and there is an option to use filters. http://efu.jp.net/soft/wg/wg.html

Suite:
1. Impulse Response

2. Filter Reconstruction Test:
single tone 19khz at full scale @ Fs:44100 overlayed on spectrum analysis of white noise @ Fs:44100.
 ADC capture rate: 192Ksps, linear scale to look at the wideband performance of the filter.

3. THD@ 1khz other suggestions welcome, maybe 40hz supplementary to look at spectral skirting.

4. -90dbFS 1khz sine wave

5. SMPTE, DIN, CCIF IMD tests. perhaps CCIF test should be looked at a wideband perspective?

6. Jitter. whenever possible use a true 24 bit JTest file.

Supplementary:
Frequency response: I'm not sure there is a point since we are looking at the HF performance with the filter reconstruction test. I would not bother unless there are obvious deficiencies.

firev1:
Updated with test suites. A couple of thoughts,

Loaded testing for suitability with preamps, what impedances and how?
Square wave testing but looking at the frequency domain for non-harmonic distortion?

ultrabike:
Suggest to test with maybe 10k?

Square wave for harmonic distortion proly not a good idea because square waves have harmonic components that will mask harmonic distortion.

Navigation

[0] Message Index