Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
42
No, I meant the recordings of the two different playbacks, so that I can listen to them with my ears, instead of your ears, and hear the differences myself. Or better yet, run them into a computer and find out exactly what “more air, bigger soundstage, added clarity” look like.
I do have a hypothesis of my own, which I’ll share. I believe you when you say you hear a difference. (Which is why recordings would be so helpful; we could see if there actually is a difference, or if you’re just imagining it.) Let’s assume you’re not imagining it. Since the limited amount of empirical evidence we have shows that the Raspberry Pi is perfectly transparent, I might hypothesize that the difference you’re hearing from the “better transport” is because that “better” transport is actually introducing distortions somehow, perhaps because it’s badly designed, and you’re hearing that lack of transparency as “better”, because that’s just how your mind works. Of course, that would just be a hypothesis, and we’d have to conduct experiments to validate it (more properly, to invalidate it).
This is why basing judgements on what some random person hears is so tricky.
… whereas “more transparent with more air, bigger soundstage, added clarity etc.” are what? Indefineable? Unmeasurable? This is all rather convenient to your argument.
The thing is this: you can’t have your cake and eat it!
That is, you can’t argue, for example that the RPi is inferior to transports that cost more without defining the reasons a more expensive item is superior. But if, as you say, you can’t measure these things, how can you design and build that superior transport?
“defining the reason” is simple, they sound better. I don’t need to be able to measure it.
And you can measure the noise from an RPi4, both directly from the USB and internally. Its the effect that noise have on the DAC that seems to be hard to measure but quite easy to hear.
It’s possible that the measurements done by ASR and Archi are not complete enough to tell the whole picture. But still, there’s a two-fold problem with your POV:
“Quite big” differences would have to show up somehow in very basic measurements of audio - distortion, jitter, SNR, etc. Even if a mysterious “something else” we haven’t measured is causing a difference in what we hear, it would have to be reflected in those measurements even a little bit if it was causing “quite big” changes in the audio output. What we hear is based on frequencies, distortion, etc. If you hearing a big change it’s not logical to say none of those fundamental aspects change, but the sound is changing in a big way.
Very few audiophile comparisons are done unsighted. This invalidates most of them. Even if done by Amir and Archi. That’s just how the human brain works-conscious or not, sighted listening changes our perception.
Recently an experiment was done where listeners heard several playback streams for comparison. The was zero difference between them.The only difference was the color of the volume knob associated with each stream in the graphic that accompanied it.
In spite of each steam being identical, the streams were perceived and ranked as being different in volume, with the red colored knob associated with the highest volume by most listeners…So that volume difference they reported was also “hard to measure” (actuallly impossible to measure a non-existent difference) and yet “quite easy to hear”. Of course the listeners DID actually hear it - because their brains took the identical signals that reached their ears and created a difference for them.
If you hear something better and want to spend money on it, go for it. But at least admit that your sighted comparisons may very well be just your mind fooling you. It happens to all of us. You might save a lot of money if you did some unsighted, tests at identical volume - or even double blinded (although that’s hard to do).
“Big” is of course a subjective term, maybe “clearly audible” is a better term in this case.
But that sighted tests automatically is wrong is a myth. Its not very hard to learn to compensate for bias, and the number one error people do is listening over a short time. At least for me, if I listen a few days I get over the bias period, which I have discovered several times when listening to something new that the first day sounded very good and then sounded clearly worse the second day.
To listen to a new HiFi equipment or tweak I do this:
Listen to the new stuff for at least 2 days
Switch back to how it was before and listen
At that point I am usually able to tell what really sounds best, and several times I have discovered that a tweak or change in my system actually sounded worse even though it sounded very good in the beginning.
No, that’s an equally woolly and meaningless term since what you say you hear is neither obvious nor without doubt fact. If it was we wouldn’tbe having this discussion.
Human behaviour is influenced by what we know or believe, so you’re only fooling yourself if you think you can prevent expectation influencing your findings.
Whilst I regret igniting this, It did drive me to research more and spend time critically listening. For my wife and I, at least, we do not hear any appreciable difference to warrant the extra $ for the HAT. We neither have an expensive, revealing system nor golden ears.
I have expectation bias like everyone else, but its not hard to learn to compensate for it, and its not hard to know when you have learnt it. We are not talking about some absolute universal physics law, but a trick your mind plays on you.
If you don’t know how to handle bias, you should learn.
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
51
These past several months I’ve played around with a variety of streaming architectures. Here’s the list:
Fanless Windows 10 WASAPI and ASIO with Matrix Audio Element H USB card with USB into Chord Qutest. Tested internal, external, and battery power for Matrix USB board. This PC is my Roon Core device in all of these experiments.
Computer above but using the mother board (ASUS) USB outputs instead of the Matrix Audio.
Raspi 4 with Roon Bridge USB out to Chord Qutest
Squeezebox Touch with the various SBT tweaks for USB output, higher bit rates, etc. USB out from SBT to Qutest. Tested LMS server and Roon squeezebox output.
Squeezebox Touch (one from above) using Coax BNC input (Blue Jeans RCA to Coax cable) into Chord Qutest. Tested LMS server and Roon squeezebox output.
All this feeds McIntosh solid state preamp and amp to Dynaudio speakers.
In my system and with my ears I don’t hear any difference between these architectures playing the same tracks on each. All of them sound equally excellent. Based on the various viewpoints from information sources online I really didn’t expect that outcome. Your experiences may be different.
I’m happy that for me I can pick whatever architecture is most convenient and not worry about it. For me I just need an async USB output to my DAC and I’m happy.
14 Likes
Bill_Janssen
(Wigwam wool socks now on asymmetrical isolation feet!)
53
Welcome to the world of digital! Bits really are bits.
Agree with you guys, had several set up and finally choose the more convenient but I ve never really heared a difference between the 3 dacs, 3 streamers and two cores I tried.
Final set up is
Core on a Nuc I5 8th gen Rock
Streamer: RPI 4 touchscreen with Ropeee XL
Dac: Cayin MK2
Pré ampli/amp: Mcintosh MA7200
I tried more and less expansive for every part except my beloved amp that I’ll bury with me, with also different types of RCA and XLR from my buying compulsory collection…to end up with this conclusion: no matter what I do it is always great to listen through any of those.
I’ve just bought an RPi 4, put it in a fanless case, and use it with Ropieee. It feeds a Quad Artera Link + Artera Stereo + ESL57s. Previously, I had an Allo USBridge as the endpoint here. Frankly, the RPi4 sounds just as good to me, I’ll be keeping it here.
Edit: I should say that the first track I tried was the opening track from Max Richter’s Voices (a new acquisition). I was somewhat shocked to hear pops and clicks, and then relieved when I realised that they were from the original recording of Eleanor Roosevelt’s voice… Whew.
Such a nice thread. I have been thrown into a very similar uncertainity related to RPI 4 via USB or a HAT Digi Board/COAX. Well initially I heard some cracking noise coming out from the USB port of my RPi 4 running RoPieee XL connected to my Topping e30, so I have added an iFi iSilencer and the cracking disappeared. Then I have added a HIfiberry Digi + Pro HAT and connected to the e30 via COAX to test whether the SQ improves. I am certainly not the bast in hearing, I have to admit that, but I have not observed any differences in noise levels. So I have ended up using the USB connection, but I have changed now from iSilencer to the iPurifier 3 (as one of you guys above in the thread) and admitting again I am really a deaf audio guy, what I have heard in difference could be best described as the difference I hear when comparing FLAC 192 with a DSD 64, e.g. the iPurifier 3 providing a reclocker produces a complete silent background. I could imagine a costly HAT Digital board with a dedicated clock generator would produce similar results, but I have opted to stay with the USB/iPurifier as that provides for more than 192k, which is the limit I could get via COAX, and that is important for me. I am also using a Allo Nirvana SMPS for both, e30 and RPi. I fully admit I am cognitively biased, but who is not… I am happy with the RPi 4 in this configuration and the SQ is very similar to what I am getting on my second setup with Auralic Aries, where I do not have to fight interconnection issues. May everyone stay with a setup he/she likes the most with his/hers personal cognitive bias -)
I have two RPi4’s, both running RoPieeeXL and using USB out to a DAC. One is using Meridian Prime headphone amplifier and power supply feeding AEON headphones. The other is using an Audioquest Cobalt feeding Sony WH-1000Xm3 headphones. It is battery powered. Both sound glorious.
I have used a RPi3 with wired ethernet running Roon to drive an OPPO BDP-105 via USB for several years and thought the combination sounded excellent. (For reference: the BDP-105 differential analogue outputs are connected directly to homebrew MOSFET amps driving Martin Logan SL3).
However, I did notice occasional, very short drop-outs which were sufficiently rare to be a nuisance rather than a show-stopper.
After changing to a RPi4 driving the OPPO via USB I have not heard any drop outs. Otherwise the system sounds the same as before: excellent!
As an engineer, and a hobby designer of audio gear, that’s all I have is measurements. There is no other way to engineer audio circuits. You use math combined with measurements of components and systems. I don’t know what to do with “this sounds wooly”, does that mean I need more feedback, etc? Over years I have gained some insights as to what circuits sound good and what can make a circuit perform the best.
As someone who has made quite a few cone and dome speakers, I have tried to voice crossovers by ear. It’s a very enlightening exercise, what sounded good to me at that moment wasn’t even closely flat. Just a simple two way crossover is nearly impossible to optimize by ear, a person’s ears are just not reliable for analytical use.
So as much as it’s entertaining to think that there are things under the surface that we can’t measure or perceive with delicate measurement gear, but somehow come out when the Beatles play, that does me no good because I can’t use that information or let it guide my design decisions. Math and measurements is all I have. The metaphysics does me no good…
I am not sure if you agreed or disagreed with me, but as I said measurements has its uses. Having said that, some high-end brands do it mostly “by ear” even during manufacturing.
But trying to determine the final sound quality (or transparency) of a HiFi equipment like a DAC with measurements alone is simply not possible, at least not with the basic measurements often used by for example Amir.
Many pi’s at my home amongst other more expensive components they sound great coupled with the right gear and in now way sound any lesser of a transport than more expensive components. I have tried my RME with a number of different transports including a pi3 and pi4 via USB and Spdif, Bluesound Node 2, NUC via USB. Tried the NUC with LPS , Pis with batteries, stock supply. It makes zero difference which I use. They all sound fantastic and to me better than my main Naim system at twice the price.
-)