Not sure I'm following... my above comment on SQ was over 1121 (mono - SE). My 1021 is 50km away form here and covered with snow
//
I hope your 1021 is not actually covered in snow... especially given its recent increase in market value
Your conclusion sounds suspicious though... How can dual mono sound audibly worse? Have you tried measuring to make sure that dual mono is not somehow getting more distortions by some statistical rarity?
sorry didn't see your edit on the previous post... You shouldn't have asked me for test tracks then since you know I don't collect music for their audiophile reference quality. But I'm guessing an easy way to get good recording quality is to go for the newer ones...
Edit: it's late and I shouldn't be on this forum. But which recording were you referring to in terms of better musical qualities?
And in this case it was older - I checked the recording date - who nows about transfers...
Just don't blame a DAC for inferior recordings. Think about all the bad stuff the music has passed before ending up in a fil.e Especially in the beginning of the digital era.
Or is it 0,1 or 0,01 ohm that is the whole culprit?
//
the recording date - who nows about transfers...Just don't blame a DAC for inferior recordings. Think about all the bad stuff the music has passed before ending up in a fil.e Especially in the beginning of the digital era.
Or is it 0,1 or 0,01 ohm that is the whole culprit?
//
And in this case it was older - I checked
Just don't blame a DAC for inferior recordings. Think about all the bad stuff the music has passed before ending up in a fil.e Especially in the beginning of the digital era.
Or is it 0,1 or 0,01 ohm that is the whole culprit?
//
the 0.01R could be the problem. Paul certainly thought so. But I don't have it on so I have no idea. I completely agree that the quality of the digital recordings is really critical.
So I assume your version is the one you posted earlier?... Tidal might be the way to go if you want to compare versions.
Quick question, does anyone know why putting on the top cover (aluminum with point contacts to other panels) adds 3rd and 5th harmonic mains noise to single channel balanced raw out in dual mono? SE RCA output has no mains noise (though there're some other serious problems...). Cheap balanced cable from Amazon uses copper shielding braids as ground wire which probably doesn't help the case... There is a 10R ground loop breaker between mains ground/chassis ground and signal ground. See screenshot for FFT.
Btw, don't get Behringer UMC202HD. If only because input impedance is 3KR and doesn't match 1.3KR balanced output impedance.
Fun fact, touching the RCA cable terminal produces -80db mains noise in umc202hd... connecting the terminal (center pole) to the signal ground produces -100db mains noise...
Also, the rest of the 60hz is all from my SK Lite laminate UI transformer (and with an oversaturated core flux thanks to 10V AC over spec...). I have no idea where the rest of the 300hz comes from though...
Btw, don't get Behringer UMC202HD. If only because input impedance is 3KR and doesn't match 1.3KR balanced output impedance.
Fun fact, touching the RCA cable terminal produces -80db mains noise in umc202hd... connecting the terminal (center pole) to the signal ground produces -100db mains noise...
Also, the rest of the 60hz is all from my SK Lite laminate UI transformer (and with an oversaturated core flux thanks to 10V AC over spec...). I have no idea where the rest of the 300hz comes from though...
Attachments
Last edited:
Of course, the real question is what we can do with the harmonics, if we should worry about anything at all... Would it be efficient and effective enough to search through the resistor error space (maybe discretize 5th bit error to +/-0.005%, 0.01% to limit search space, and increasing granularity from there to MSB) using FFT plot as a guide to whether our guess is close? Soren does this idea work electrically? If each user can reasonably calibrate a channel running a software in the background for a few hours, might be very much worthwhile.
If there's a mathematical relationship between bit errors and harmonics ratio, maybe we can even come up with a better algorithm than exhaustive search.
/
Or maybe we input constant level signals (if it doesn't cause physical damage..) and compensate for the non-linearity directly? We get the levels of each of the 28 bits first, that's 28 measurements with 2 decimal place precision, average to find a rough center for all the resistors, and then correct each one directly in software. If we want we can only correct for the 10 MSBs which will be much faster. We can reach thermal stability first to make the most out of this process. This assumes the ADC has good linearity.
If there's a mathematical relationship between bit errors and harmonics ratio, maybe we can even come up with a better algorithm than exhaustive search.
/
Or maybe we input constant level signals (if it doesn't cause physical damage..) and compensate for the non-linearity directly? We get the levels of each of the 28 bits first, that's 28 measurements with 2 decimal place precision, average to find a rough center for all the resistors, and then correct each one directly in software. If we want we can only correct for the 10 MSBs which will be much faster. We can reach thermal stability first to make the most out of this process. This assumes the ADC has good linearity.
Last edited:
I don't build my chassis out of metal anymore due to the same finding as you did. Once you close a metallic box, the sound gets restrained. I know how one could be led to think that the physical roof leads to an audible "roof" - but my own and combined impression from fellow builders leds to this conclusion. I now build with plastic enclosures. They are dead silent.
Picture of my prototype build. I'm so satisfied with it that I will make something better looking. DAC smps power is located mora then 1 meter away - regulators are close to DAC. I feed each with a 5 meter opto link - speaker cable is < 0,5 meter. Goal is shortest possible analogue path.
//
Picture of my prototype build. I'm so satisfied with it that I will make something better looking. DAC smps power is located mora then 1 meter away - regulators are close to DAC. I feed each with a 5 meter opto link - speaker cable is < 0,5 meter. Goal is shortest possible analogue path.
//
Attachments
Last edited:
The “roof” finding seems suspicious... it’s easy to dream up something like that without proper ABX...
I already have 10R between chassis and signal ground, maybe it needs more?... doesn’t seem like good practice though...
I don’t see ABX tests... heck I don’t think the 0.001% guy even knows what 0.005% sounds like. He just went for it...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.