Maybe you could try wrapping the whole cable connection up in a sleeve made of electrical shielding tape to fend off electromagnetic interference (EMI). Even gold plated, 3.5mm TRS mini jacks still aren't the best choice for analog high end audio connections so probably, I would leave out mini jacks and go with the full fat "version" of the DIY route instead.
Also, onboard audio interfaces aren't galvanically isolated from the other components of a motherboard. Even with the best voltage regulators and well-engineered electrical noise filtering techniques, electrical noise patterns still creep in through the power signal that feeds the onboard audio interface. The digital path entering the DAC chip is still electrical so, it too carries unwanted electrical noise patterns into the onboard audio interface's more sensitive, analog intestines. Not all outboard DACs are necessarily always better than onboard audio, though, as the common observation that all onboard audio interfaces aren't created equal also applies to outboard DACs, prepros, and AVRs.
That said, looking at SNR and THD+N alone is like judging marriage by looking at a wedding picture or studying a river by taking out a bucket of water and then staring at it on the shore. Audio is part science, part subjectiveness. Whatever you do, don't waste your time with blind, or double-blind listening tests (DBTs) like ABX tests, for example. When we perform the task of listening specifically for differences during a DBT, we don't just eliminate hearing bias. Rather, we replace it with different types of hearing bias. I will try to explain a few things about what this means.
When we partake in a test, we aren't consciously aware of all the human stress factors caused by performing the test, but these factors still nevertheless introduce listening bias to some non-negligible extent. The fact that humans under stress can't tell the difference between quite surprising things has been proven many times over.
To make matters only worse, our brain tends to fill in gaps. Simply rememebering that a certain consonant belongs somewhere in a certain spoken sentence causes us to hear that consonant even after it was removed from the sample with audio editing software, for example, and, these traits are inherent of how human hearing factually works. It's very easy to prove. As a result, we can switch to a better playback system that reveals subtle details previously masked by distortions from the inferior system, then when we go back to the inferior system, we potentially can always hear these details (or partially) despite that they are masked. Because we have memory, switching back and forth skews the test results towards "hearing no difference".
Further, different characteristics of sound arrive on different time scales. Some sounds are heard immediately whereas when we enter a concert hall it can take up to ~15 minutes before we start listening fully into the space. Put simply, our brain externalizes sounds as sound objects against a background, and builds a map of this soundscape over a relatively longer period of time. When we do rapid switching, that natural process breaks, which limits the usefulness of ABX comparison tests if the test samples are being kept fairly short. Like maybe only a few seconds, which tends to be a common recommendation of sample duration for doing ABX comparison tests by self-professed objectivists who choose to stay ignorant of every important nuance or almost, fighting bias by adding more bias or... the blind misleading the blind.
Finally, setting up and conducting a proper, valid, DBT is very challenging, and takes a lot of resources. Test subjects who can be admitted to such listening experiments have to be professionally trained and meticulously selected. Otherwise the test results usually turn out inconclusive, and, a lot of people seem unaware that DBTs can only be used to prove that differences can be heard (not to prove that differences are inaudible to humans nor even to support the claim, that it is unlikely that differences are audible to humans).
Also, onboard audio interfaces aren't galvanically isolated from the other components of a motherboard. Even with the best voltage regulators and well-engineered electrical noise filtering techniques, electrical noise patterns still creep in through the power signal that feeds the onboard audio interface. The digital path entering the DAC chip is still electrical so, it too carries unwanted electrical noise patterns into the onboard audio interface's more sensitive, analog intestines. Not all outboard DACs are necessarily always better than onboard audio, though, as the common observation that all onboard audio interfaces aren't created equal also applies to outboard DACs, prepros, and AVRs.
That said, looking at SNR and THD+N alone is like judging marriage by looking at a wedding picture or studying a river by taking out a bucket of water and then staring at it on the shore. Audio is part science, part subjectiveness. Whatever you do, don't waste your time with blind, or double-blind listening tests (DBTs) like ABX tests, for example. When we perform the task of listening specifically for differences during a DBT, we don't just eliminate hearing bias. Rather, we replace it with different types of hearing bias. I will try to explain a few things about what this means.
When we partake in a test, we aren't consciously aware of all the human stress factors caused by performing the test, but these factors still nevertheless introduce listening bias to some non-negligible extent. The fact that humans under stress can't tell the difference between quite surprising things has been proven many times over.
To make matters only worse, our brain tends to fill in gaps. Simply rememebering that a certain consonant belongs somewhere in a certain spoken sentence causes us to hear that consonant even after it was removed from the sample with audio editing software, for example, and, these traits are inherent of how human hearing factually works. It's very easy to prove. As a result, we can switch to a better playback system that reveals subtle details previously masked by distortions from the inferior system, then when we go back to the inferior system, we potentially can always hear these details (or partially) despite that they are masked. Because we have memory, switching back and forth skews the test results towards "hearing no difference".
Further, different characteristics of sound arrive on different time scales. Some sounds are heard immediately whereas when we enter a concert hall it can take up to ~15 minutes before we start listening fully into the space. Put simply, our brain externalizes sounds as sound objects against a background, and builds a map of this soundscape over a relatively longer period of time. When we do rapid switching, that natural process breaks, which limits the usefulness of ABX comparison tests if the test samples are being kept fairly short. Like maybe only a few seconds, which tends to be a common recommendation of sample duration for doing ABX comparison tests by self-professed objectivists who choose to stay ignorant of every important nuance or almost, fighting bias by adding more bias or... the blind misleading the blind.
Finally, setting up and conducting a proper, valid, DBT is very challenging, and takes a lot of resources. Test subjects who can be admitted to such listening experiments have to be professionally trained and meticulously selected. Otherwise the test results usually turn out inconclusive, and, a lot of people seem unaware that DBTs can only be used to prove that differences can be heard (not to prove that differences are inaudible to humans nor even to support the claim, that it is unlikely that differences are audible to humans).
Last edited:
My Computers
-
At a glance
11 Homei7 13650HX16GB DDR5GeForce RTX 4060 Mobile- OS
- 11 Home
- Computer type
- Laptop
- Manufacturer/Model
- Asus TUF Gaming F16 (2024)
- CPU
- i7 13650HX
- Memory
- 16GB DDR5
- Graphics Card(s)
- GeForce RTX 4060 Mobile
- Sound Card
- Eastern Electric MiniMax DAC Supreme; Emotiva UMC-200; Astell & Kern AK240
- Monitor(s) Displays
- Sony Bravia XR-55X90J
- Screen Resolution
- 3840×2160
- Hard Drives
- 512GB SSD internal
37TB external
- PSU
- Li-ion
- Cooling
- 2× Arc Flow Fans, 4× exhaust vents, 5× heatpipes
- Keyboard
- Logitech K800
- Mouse
- Logitech G402
- Internet Speed
- 30Mbit/s up, 500Mbit/s down
- Browser
- FF
- Antivirus
- What's an antivirus?
-
At a glance
11 Homei5 1135G716GB DDR4Intel Iris Xe- Operating System
- 11 Home
- Computer type
- Laptop
- Manufacturer/Model
- Medion S15450
- CPU
- i5 1135G7
- Memory
- 16GB DDR4
- Graphics card(s)
- Intel Iris Xe
- Sound Card
- Eastern Electric MiniMax DAC Supreme; Emotiva UMC-200; Astell & Kern AK240
- Monitor(s) Displays
- Sony Bravia XR-55X90J
- Screen Resolution
- 3840×2160
- Hard Drives
- 2TB SSD internal
37TB external
- PSU
- Li-ion
- Keyboard
- Logitech K800
- Mouse
- Logitech G402
- Internet Speed
- 30Mbit/s up, 500Mbit/s down
- Browser
- FF
