OK, so we have 2 different filters that produce a different sound. Nothing really shocking here.
An oversampling digital filter shifts the HF noise and aliasing products to higher frequencies, making the job of the analog filter easier (you get buy with a much less steep analog filter), but you still need an analog filter.
You may, but as I don't have any DACs available that don't have the required filter, it is rather pointless.
Functionally it does exactly the same as any normal low pass filter - that is what it is.
Well, now I understand (from Ken`s post 36) why I appreciated more the sound out of a non filtered DAC system. The integrated filtering in the DAC chip it may be enough for some peoples...
It might be enough that the built-in low-pass function of subsequent analog stages might be sufficient.
We don't need to. If the DAC is a current source, the pole will be dependent on the resistance of the equivalent resistance and the capacitance.
If a single pole filter is placed after the I/V, the results will be the same.
The only difference would be due to interactions with the DAC not behaving with a capacitive load. (If the I/V is perfect, the input impedance zero, the cap doesn't even have an effect!) That's going to be negligible unless we do something to upset the DAC by placing a load that it's not designed to drive on it.
Here's what I will suggest. If you won't first try this rather simple experiment, that's your prerogative, but do not waste your time (or mine) commenting in this thread? There is a difference between healthy skepticism and closed-mindedness.
Last edited:
No, you don't a specific analog filter circuit. The oversampled alias images are low-pass filtered by interconnect cable capacitance, by amplifier bandwidth limits, by loudspeaker bandwidth limits, and by human hearing bandwidth limits. The only potential technical problem due to high oversampled alias products is the potential for inter-modulation distortion, which would be quite low in level at any rate due to the zeroth-order hold operation of most DAC chips, which inherently overlays a SINC function shaped frequency suppression mask on top of the image products.[/QUOTE]
Julf, see my post #46 comment to FoMoCo, as it applies to you as well.
Last edited:
Pardon me great master of all that is rubbish. For I did not mean to waste your time with simple textbook theory that infringes on your ability to conjure something from nothing and call it anything you please. You do have the right to call anything by any name you please. Science shall not trod upon that right.
You are correct. Your closed mind refuses to accept that there is nothing special (beyond a simple low-pass filter) here, even though there are stacks of books and papers that will prove you wrong. Sit down and analyze the circuit presented and (Shazam!) you'll see nothing but a simple single-pole RC filter composed of the added capacitor and the parasitic elements of the circuit.
Of course, Zen Mod makes the point that we don't know the output stage, and that can indeed make a difference, but as long as we don't violate the constraints that the manufacturer of the DAC gives, we still have a simple low-pass filter. If we do violate those constraints, we most likely have a stability problem. Possibly some oscillations. But, most current source outputs of DACs behave just fine with a capacitive load.
Perhaps you should spend some time with basic circuit analysis before claiming that I shall not post simply because I won't perform an experiment that you feel is requisite. I've made RC filters before, no need to do it again to appease you.
I think it is you who are showing the closed-mindedness. This thread is supposedly about investigating a "new" effect, but you don't seem to want to hear a simple, rational explanation based on well-known principles. It is almost as if you are enthusiastic about having finally found an unicorn, albeit one without a horn, and get really offended when we tell you it is a horse.
So clearly you agree we need some form of analog low-pass filter. How is the capacitor inserted by Joe Rasmussen any different than the cable capacitance? All he has done is ensured *sufficient* capacitance.
trying my own I/V stage with input impedance of 3R , balanced , is still on list to do ;
prior to that , I'm speaking from experience with xformer based I/V stages , and I wrote with which DAC chips I tried it .
sound wise - they're more then good enough , and mentioned 6db filter was necessary
however - dispute - is it plain filter or necessary patch (solution) - I don't care
I see you changed your post after I responded to it. Who gives you the right to tell me I cannot comment?
It looks like that low pass filter creates a slight roll-off at the end (and not-so-near-the-end) of the audio spectrum, and it has been reported to sound "psychoacoustically better". I remember that Mr. Linkwitz mentioned it in the last section of his article about "A Three-Enclosure Loudspeaker System with Active Delay and Crossover".
Yes, while at the same time removing the aliasing products and high frequency noise from the digital-to-analog conversion process (as specified by the principles of sigma-delta conversion).
Yep...that's right, but it seems that many people like "the noise" at the sigma-delta DAC's output
Yep...that's right, but it seems that many people like "the noise" at the sigma-delta DAC's output
Sure, just like a lot of people like "soft" valve distortion and other enhancements to the sound. See the Aphex Aural Exciter.
some comments on the scope traces posted earlier :
1) http://www.diyaudio.com/forums/atta...ods-using-working-configuration-cap-value.jpg
Is the right trace measured at the DAC output or at the opamp IV output ?
2) http://www.diyaudio.com/forums/atta...dac-processing-circuits-optimized-largest.jpg
Same question.
Could you zoom in ? I'd be very interested in the edge rate (V/µs) of these transitions. Something like 10ns/div should produce interesting results.
Also, lots of overshoot, noise, and clock feedthrough here, is it the circuit, the layout, or the probing ?
3) http://www.diyaudio.com/forums/atta...al-level-worse-case-scenario-degraded-aud.jpg
In this trace the opamp is probably oscillating (quite badly...) from capacitance overdose at the input.
1) http://www.diyaudio.com/forums/atta...ods-using-working-configuration-cap-value.jpg
Is the right trace measured at the DAC output or at the opamp IV output ?
2) http://www.diyaudio.com/forums/atta...dac-processing-circuits-optimized-largest.jpg
Same question.
Could you zoom in ? I'd be very interested in the edge rate (V/µs) of these transitions. Something like 10ns/div should produce interesting results.
Also, lots of overshoot, noise, and clock feedthrough here, is it the circuit, the layout, or the probing ?
3) http://www.diyaudio.com/forums/atta...al-level-worse-case-scenario-degraded-aud.jpg
In this trace the opamp is probably oscillating (quite badly...) from capacitance overdose at the input.
First of all, when I edit posts, it's either for greater clarity or for more effective tone. In short, for more effective communication. I do this regularly. Your response had nothing to do with my edit, as I hadn't read it prior. Frankly, I resent your implication. I would simply post a new comment if I had something further or difference to address to you.
Secondly, I did not tell you that you cannot comment. It's interesting how you mis-characterize my comment, and then overreact to your own mis-characterization.,You protest too much. The bottom line is this, if you want to maintain a closed-mind regarding something which is very easy to test for yourself, that's your free choice. Just as it's my free choice to stop wasting my time attempting to open that mind. Those who are open to Joe's discovery will benefit, while those who aren't open won't. Simple as that, and audiophile life goes on.
Last edited:
- Status
- Not open for further replies.