Sure.
But there is a difference between the measurement equipment, where, of course, harmonics must be as lower as possible, and the sound systems.
According to statistics from users of some of the my EES'based DACs - ~40% prefers to switch OFF THD Compensation. ~20% prefers it ON, ~10% prefers "anti compensation" (means increased harmonics), and the rest - they do not hear the difference at all!
I would bet that's because they are using the distortion to compensate for other problems in their systems. For example, a muddy sounding speaker driven with a bit of 3rd harmonic may sound a little more 'clear.' Its not right of course, but often people don't know why the sound is muddy in the first place, or maybe they aren't in the market to acquire better speakers so they do what they can with the toys they already have.
Oh come on guys the DACs in question already have their THD well below the threshold of audibility with the THD compensation turned off. The feature is only of academic use when chasing numbers, for the likes of those on audiosciencereview, or for those of us trying to create a system for measuring things.
THD at -110dB, or, compensated down to -130dB, no one is going to hear any difference and if they are trying to tell you that they can they are deluding themselves. Of course you can use the THD compensation to do the exact opposite and inject copious amounts of 2nd and 3rd. Now this you could end up hearing, providing the levels were high enough, and preferring.
THD at -110dB, or, compensated down to -130dB, no one is going to hear any difference and if they are trying to tell you that they can they are deluding themselves. Of course you can use the THD compensation to do the exact opposite and inject copious amounts of 2nd and 3rd. Now this you could end up hearing, providing the levels were high enough, and preferring.
Not necessary, look at IVX and his DAC/ADC products. Since he's in the business, his work (splitting the chips in classes, individually tuning boards, etc...) suggests there's a market for the number hunters, ready to pay a premium for the ultimate performance even if it's, of course, 40dB or more under the audibility thresholds.
Oh absolutely there's a market for it. Those at audiosciencereview will pay for that extra dB in SINAD. Not that they'll hear it, which was supposed to be my point. I'm a bit of a number chaser myself, so I understand the want/need/desire to have the DAC that measures the best. I know I can't hear it but I still prefer it in my system.
I didn't mean the forum in and of itself. I meant the people that the forum attracts and then influences the purchases of.
Nothing, except speakers and acoustics in general. That's the root cause why the High End Audio industry went pretty much the Dodo way, quite some time ago.
Personally, I take it as a strictly technical challenge, If I could do a 0.01ppm power amplifier, I would for sure build it and measure it, only "because I can". I was considering going into speakers years ago (I even got all the instrumentation, microphone, software, etc...), but then I gave up, since I concluded woodworking is not something that could be done in a basement, and otherwise Toronto is not the best place for working outside, for about 6 months/year.
But audio is not my only DIY fun challenge, I am having fun with microwave black magic circuitry (you can do a lot of circuitry, before adding a single piece of semiconductor to a PCB), instrumentation, etc... Opportunities to learn something new and keep the grey cells active.
Now, if only we could get those chips back. All my projects are on hold for now
.
As was said above me, at the state of the art, and by this I mean the actual state of the art, not some audiophool idiocy, absolutely nothing. Although there are always differences of opinion in what trade offs to make when it comes to loudspeakers. KEF vs Revel for example. Very few manufacturers bother to design excellent coaxial drivers. It's not impossible to do, as KEF has shown, and clearly there are technical benefits and very few, if any, draw backs to be concerned about. So pushing that further would be something I would look at.
What's left to deal with is bringing that state of the art into the mainstream and making it cheap and accessible.
Doing anything right is never cheap. I guess that is relative, but the build quality is often pretty poor. Circuit design could be excellent, but almost is never even competent. The average audio component sold isn't very good, and designs from the very early 80's were often better. A lot more reliable too. Back then quality and performance mattered in the main stream.
Speakers have come a great long way, and further improvement is possible. So this is really the only area for technical improvement. Electronic product improvement is a question of eliminating contract manufacturing and providing products designed well. So while we could have a better performing amplifier (or whatever) today, try and buy something like that.
Speakers have come a great long way, and further improvement is possible. So this is really the only area for technical improvement. Electronic product improvement is a question of eliminating contract manufacturing and providing products designed well. So while we could have a better performing amplifier (or whatever) today, try and buy something like that.
To be honest the biggest issues with budget speaker design can be solved relatively cheaply using active speakers and the abundance of modern single chip amplifiers. Those with built in DACs and DSPs. The worst offender to decent sound quality is passive crossover parts costing an arm and a leg plus cheap drivers generally needing complicated filters to sound half decent. Not to mention the knowledge on how to design a decent crossover. Use the onboard DSPs for the crossover work and you've just solved most of the problems.
I've seen a handful of small, cheapish, DSP based speakers using TIs single chip amps with excellent measured performance. I think Sonos go this route with some of their speakers too.
The technology is there most don't take advantage of it though.
I've seen a handful of small, cheapish, DSP based speakers using TIs single chip amps with excellent measured performance. I think Sonos go this route with some of their speakers too.
The technology is there most don't take advantage of it though.
Sonos Connect - S1 were the only digital appliances that sounded like crap to me, from he very beginning, and that after costing an arm and a leg. To add insult to injury, Sonos obsoleted them, no longer supporting the SW, and trying to sell me the new version. No thanks, I just threw everything branded Sonos out of my network and installed Rpi4's with Nanosound DAC hats and cases.
It was not trivial configuring Volumio to have it working right (for example, before enabling nfs on the NAS and configuring Volumio accordingly, it took almost 24 hours for a full library scan, then Rpi4's were losing wifi connection every few hours and never recovered before power cycling, I added a cron job running every minute, checking if wifi is healthy and restarting the service otherwise, etc...). The end result is sound wise as good as I ever heard from a digital source, at 1/3 the price of the effing Sonos devices.
@5th element:You'd get pitchforks and flaming torches at your door saying that in some forums! I should note that I agree entirely with you, I'm just not brave enough to argue it and I do have a lot of admiration for those who can design a really good passive crossover as it does my head in.
Why so many people still aim for a 1960s architecture in their systems is a confusion.
@Syn08: I had to rebuild my server recently as the latest version of software for the kitchen system needed a Java version my old debian install refused to support. Gigantic pain to rebuild as lots of little things had changed so even setting a fixed IP address took me ages to work out. I still can't remotely hibernate it. Luckily it will scan a 3TB archive in under 15 mins. USB out in the living room to an SPDIF converter and into miniDSP. Gapless playback eludes me which is a big annoyance on some albums.
Why so many people still aim for a 1960s architecture in their systems is a confusion.
@Syn08: I had to rebuild my server recently as the latest version of software for the kitchen system needed a Java version my old debian install refused to support. Gigantic pain to rebuild as lots of little things had changed so even setting a fixed IP address took me ages to work out. I still can't remotely hibernate it. Luckily it will scan a 3TB archive in under 15 mins. USB out in the living room to an SPDIF converter and into miniDSP. Gapless playback eludes me which is a big annoyance on some albums.
Last edited: