Hi Matt
one question about dac tests.
In Audioreview magazine there was a test on impulse with 44,48,96 and 192 kHz.
One example in attach
I have some wav samples.
One 48 kHz that maybe is fine. The other at 192 more complicated.
Do you think is possible to have the same results with 403 set to Time testing a dac?
Thanks
Walter
Walter
Impulse testing done for DACs or CD players was historically done to determine the presence of linear phase (oversampling), non-oversampling and the overall polarity of the DACs (inverting or non-inverting).
A single 0dBfs sample was played by the device and the resulting waveform captured by a 'scope.
In reality, such a signal is illegal,as it cannot be captured in “nature”. It cannot exist in the real analog world. It can only exist in the digital or electrical domain as a manufactured signal- a bit like a perfect square wave.
There are plenty of test discs and programs that can create such signal and I use ask me for the file (Denon standard test disc) to quickly check polarity etc. In the present day, there are plenty of oversampling DACs that can apply whatever “flavour” they want to the FIR filtering to resemble NOS DACs or whatever they want.
As the QA403 is analog in and out, the ADC will apply its filtering and the DAC does the same, so you cannot drive a single sample impulse into a device to see its response in the analog domain.
Sure, you could create a file, “play” it via the mirroring function into a DAC and capture the DAC’s output, but the waveform you would collect would be limited/affected by the QA403’s ADC filtering AND and the sample rate restrictions as well at the DUT DAC’s own impulse response. With the QA403, the sample rate of the “mirrored” source signal must match the analyzer sample rate too.
The best device for capturing an impulse is a 'scope with a decent bandwidth:
Here’s a 1st generation Sony CDP-101 (non oversampling infinite impulse response (IIR))
Hi
thx
It is clear your explanation and match with the info I have
But if the 403 is a instrument even it has the ADC and DA from ESS.
When I plot a FFT at 1 kHz (just an example) coming from a dac I see the real until it is generated with a samplng frequency that is 1/2 of the max (192kHz) in theory
This is an impulse read with AP Sys2 , 48 kHz of sampling test impulse
Hi @walge, you could set the QA403 to a large buffer that would take 5 seconds or so to run. And then start it, and then immediately play your impulse response. Then wait for the capture to end. That would place the impulse somewhere in the buffer. Then zoom in in the time domain to look at it. Or export as a wave and use whatever tool you want to examine and analyze the wav.
Restorer-John is right: the ADC has its own brickwall filtering that will make it hard to discern what you are looking at.
Hi
Thx.
I will test it.
But if the Adc has a filter ( of course) the results of FFT ( for example) are real ?
Or the impulse is treated different from frequency domain?
Walter
But if the Adc has a filter ( of course) the results of FFT ( for example) are real ?
Hi @walge, yes, the ADC has a filter and the DAC will probably have a filter too. And so, a wav file of the impulse response will result in a band-limited impulse response, and the FFT of the impulse response will reveal the composite (both ADC and DAC) filter function. And if that is what you are aiming to see, then the results are interesting.
Now, the QA403 ADC uses the ES9822, and the filter is (I think) the linear-phase, fast-rolloff filter. You should be able to capture the impulse response, save as WAV, give that to AI, tell it to window the IR, produce the spectrum, then give the AI the ES9822 filter response picture, let it extract the response from the picture, subtract that from the overall response, and then ask the AI to render the DAC filter response.
