I reduce the level as on my monitor setup, I find the output from my DAC/pre-amp to my power amp a bit too hot, so I don’t get enough play on the volume knob. With a -12 db, I get better control on the volume pot. Also, the pot becomes closer to the middle where it performs its best. I will be upgrading to something stepped eventually.
I have 2 Parametric EQ lined up. One is just flat for monitor use at -12 db. The other EQ is -6db with actual EQ changes and for headphone use.
It doesn’t really matter in the grand scheme time-wise. I’d still like to confirm whether reducing gain flat by PEQ is the same as Headroom Adjustment if both attenuated by the same amount?
Completely agree. My experience was that a recent update of Roon changed the default to ‘DSP ON’. I did the update and sat down to listen to music a day or so later. What happened? Why is the sound quality degraded? Then I remembered that Roon had been updated. I went to the forum to see if others were complaining. Saw nothing online. Then I remembered that I played around with the DSP some time ago and did not like it. I checked the DSP setting and found it ‘ON’. I turned it off and my system sounded great again.
I switched to Roon from Audirvana because i could not distinguish between them sound wise. I will go back to Audirvana if a future build degrades the sound quality. I like Roon, but sound comes first for me.
Of course, any perceptions I can have only apply to my hardware and hearing ability. Perhaps my MacMini does not have the chops to do the DSP.
I just wanted to say that I hear similar effects as Frank in my system.
MacMini->Netgear switch->ultraRendu/LPS-1->Berkeley Alpha->Ayre KX5/VX5->Revel Studio 2
JE_Stout,
Just to reiterate, It was later found not to be the DSP Engine enabled that caused lower SQ but the “Automatically Update Changes” on the same page that caused a lower SQ for me. I found the DSP Engine to be transparent. Roon is saying the Auto Save should not make a difference but it does in my system, my ears, my room, my etc…
The DSP on/off does nothing unless you also have something to turn on/off, like equalizer settings, volume adjustments, etc. If you turn it off and hear an improvement, its probably because you had some settings that changed the sound. Try turning it on and check the signal path.
“Automatically update changes” just tells Roon to save as you make changes, or wait with saving until you press the “Apply” button. If this in itself has any influence of sound than Roon would have very embarrassing bug on their hand (but I somehow doubt it).
Just to follow up this: unfortunately I did not get the time I needed to do this test. Annoying as I had it all planned re. how to do the test. Now it will take quite some time before I again have a chance to do this.
I spent 30 minutes or so with the DSP engine this afternoon. NOW I can say that I cannot detect a difference with the main DSP switch on or off. (note that I did not play around too much with any actual DSP ‘sound affecting’ settings.) I’ll spend more time with it and update if my perception changes.
I also do not find the ‘Automatically Update’ switch to matter.
I do, however, stand by my previous experience. Something was going on…I also had the perception if you entered a delay in the device settings, it had a deleterious effect.
OK guys, tonight I listened over an hour trying different DSP Engine configurations - on/off, auto save on/off, etc. and I could NOT detect any change is SQ. A couple of nights ago I reported a lowered SQ with the Auto Save on but now it does not seem to matter. Maybe the way my legs were crossed or maybe the wine…who knows? Good news is - it ALL sounds terrific! So enjoy.
You ought to reduce the gain in the analog side of your system if possible. By cutting the input signal to your dac, you are going to hear more of the noise that is inherent in the dac and preamp relative to the music signal.
Is it possible to lower the output signal from the DAC? Some have jumpers or dip switches inside to adjust the output level (gain). You might also be able to use a single ended output from one device instead of balanced.
My DAC/preamp (Dangerous Source) does have internal jumpers, but none for reducing the volume of outputs. I looked into it already. I and another person had this same problem with too much gain with Dangerous DACs, Source and Convert-2, with too much gain on Amphion systems. Dangerous recommended the use of “Whirlwind IMP Pad” that come at different levels of attenuation (dB).
Dangerous told us that it would have no impact on SQ, but when the other try guy used a few of the IMP Pads with different attenuation, he found they all messed with the sound. I would have given them a shot just to see for myself as they are $20 in US. I tried to get some here in Canada and they were over $150+ at the time. The reseller here apparently sells them in hundreds, and I only wanted a pair.
Short of upgrading my DAC/preamp or getting a separate high quality preamp, this is what I’m stuck with for the moment. I’m open to suggestions.
Use the Headroom adjustment feature. Start with about -6 dB and keep reducing the “gain” until you reach a value that you are satisfied with. Leave your hardware as is.
I use these between an old AVI pre-pwr amp combination because the volume control pot has become noisy at the bottom end. No effect whatsoever on SQ but no idea why there would be under the circumstances.
I’ve never laughed so hard in my life reading a thread. The confirmation bias is so strong in this thread. I can’t believe people hear these imaginary ghosts and goblins and still find any enjoyment from listening to music at all.
If you switch on DSP, the signal processing goes to frequency domain from time domain:
time domain music -> frequency domain conversion (FFT) -> frequency domain DSP processing -> conversion back to time domain.
The problem came from FFT; all engineer knows only infinite FT (Fourier Transformation) chain can model the time domain without distorsion. It means infinite processing power necessary which is not given. The problem audible.
An example what you saw already; convert a good quality loseless photo (BMP/TIFF) to JPEG. You will see the same picture but fine details become unclear. The JPEG transformation is an FFT conversion on photos like DSP FFT conversion on music.
If you don’t hear the difference that means;
You have infinite processing power for DSP
Or your DAC/amplifier, your listening room do higher distorsion on music than the DSP FFT conversion itself. I bet it is the case. (The most important to have a good DAC around or under 0.0001% THD, check technical details of your DAC)
Why people use DSP if it is not so good ? At point 2) there is the room distorsion. Room walls can cause +/- 15dB attenuation, reflections at some frequency ranges or more at same specific frequency. People looking for HiFi speakers with +/- 3dB or 6dB maximum deviation in frequencies between 20Hz and 20000Hz while their room can cause +/-15dB distorsion in music.
For this reason that could be an important question; how to create linear room frequency response (e.g.:by DSP) while keeping loss of fine details on a minimum level.
For example I have absorber/diffuser panels on walls instead of DSP (I have DSP hardware but had same problem like discussed in this forum) difference is big in audio quality.
We all enjoy listening music just want the best what is possible from our equipment.
Uh, I am not sure which, if any Roon DSP operations are performed via FFT. Parametic EQ, for example, uses IIR filters. So, those are serial operations in the time domain. Convolution uses FIR filters. Those, too, are serial operations in the time domain.
When I do this -3db headroom adjustment, the final bit depth conversion is 64bit float to 24bit with my DAC. Since this is decimation, roon should be adding dither, increasing the noise floor in exchange for reducing quantization error. This is not reversible, that is, if the noisier dithered output could be reprocessed to add back the 3db gain, the result would not be the same as the original. And it may even be audible too!
Is roon applying dither? If so, what dither function is it using?
If roon is adding dither, and DSP volume is applied in the same manner as reducing headroom (bracketed by the two sample size conversions), then DSP volume should be shown in the “Signal Path” along with the conversions when it is attenuating. Otherwise, the display is misleading by showing an unaltered signal path.
I am with you @edman, I have raised the questions a few times but never got a clear answer. It is worthwhile to understand these details of the DSP implementation, so that the full chain can be optimized, especially. for 24bits Outputs (32 bits are less an issue).
