Some interesting experiences with modern digital gear

What is "ASR"? link?
https://www.audiosciencereview.com

The most well known community on the Internet for audio enthusiasts who believe in the importance of objective measurements of audio gear. People on that site regularly take careful measurements of audio gear and publish detailed results, and also publish details of apparatii and processes they use for these measurements. The site strongly rejects advertisements to avoid conflict of interest, accepts donations, and permits audio gear manufacturers to send the members their products for measurements to be published.
 
DAC filters have easily measurable differences.
Yes, I've seen those measurements quite often on ASR. However, the few times that I've tried to hear audible differences, with just two appliances so far, I've failed. Which doesn't surprise me, because I'm 60, and can barely hear up to 15 KHz, and all these filter response graphs seem to affect 18KHz and above.

All this makes it even more remarkable that I am hearing the differences between these two CD players.
 
Just thought I'll write about my very interesting and somewhat disturbing experiences in the last one month listening to music through a few CD players. This post is not about building anything, and there's no DIY here, though I'm a DIYer. Hope this is the right forum for this. The "Lounge" forum seems to be for non-audio stuff.

Background

I've been learning to design speakers since 2002 or so, and I've built a few. I'm an engr by qualification, and I tend to believe in systematic design and measurement. I do not believe that power strips or power conditioners impact the sound of (well designed) modern audio gear, for instance, and I don't believe that CD transports do either. (I could be wrong, but I'm explaining my starting point.) I like ASR. And in terms of sound signature preference, I like clean, accurate uncoloured sound, with full detail retrieval. I don't like "rounded" sound, or "inoffensive, pleasant" sound -- it puts me to sleep. I would probably like the sound of any good studio monitors. I love the sound of my friend's PMC 3-way passive monitors with 12" Volt woofers, 3" dome mid, driven by his Luxman amp stack.

Now starts the story of the last month.

What happened

I used to believe that a DAC is a "solved problem" -- any two DACs which have flat frequency response and inaudible levels of SINAD across their frequency spectrum will be indistinguishable from each other. By SINAD, I don't mean a single number at 1KHz -- I mean the entire harmonic distortion profile across the spectrum, and I include 32-tone IMD in the measurements. But I have now had some listening experiences which have made me question this.

For the first time in my life, I had the following things set up in my bedroom:
  • very good, transparent headphone amps (the JDS Labs Atom Amp 2, and the RME ADI2 FS Pro in preamp mode)
  • a bunch of really transparent and uncoloured IEMs (Moondrop Chu II, Truthear Zero x Crinacle Red, Tanchjim One, Etymotic ER4PT), and a pair of headphones (Sennheiser HD650)
  • two expensive CD players, in the $2,000-5,000 range
I was expecting that these two CD players will sound the same. They didn't.

I first bought one CD player. (Let's call it Player A.) I found it to be super-clean in sound, but I noticed that my brain was reacting to it by losing concentration -- I'd start one track, listen for a minute, and my mind would stray. And in objective terms, there seemed to be an etched quality, an edge, to the sound. I did not think much about it, and attributed it to the CD's mixing/mastering, and/or the earphones.

Then, due to a strange mix of circumstances (I was buying for my friend) I bought a second CD player (Player B), thinking I'll sell one anyway. I've never had the funds or the reason to do this before, in my life. And when I heard both in quick succession, the sound difference was quite remarkable. With the second player, gone was the edge, though I could hear full details. In fact, I could hear more detail now. And gone was my tendency to lose concentration. I was using exactly the same cables, headphone amp, and earphones.

I know that comparisons need level matching. So I went one step beyond level matching -- I tried to actively change the gain up and down while listening to one player, while keeping the gain unchanged when listening to the other player. The subjective impressions remained unchanged -- Player B sounded better, even when I made Player A sound louder, or softer.

I called my 18-year-old to listen. He's not interested in audio gear, but loves music, plays the piano, and has ears 42 years younger than me. I told him I'm hearing differences, and asked him to try the two players out. He used the HD650 (he diesn't like in-ears) and he played a CD he loves, which I had not used for my listening. It was "Retrospectacle" by Supertramp. He listened to just 1-2 tracks and said he can hear a difference. His exact words were "The second one is softer, but I like it more." I had not told him which was worse or better to my ears, but his identification of the "second one" matched mine.

Then I switched from the Atom Amp 2 to the RME ADI2 FS Pro, which, as you know is a complex piece of gear, and does A2D and then D2A before driving the headphones. I was not sure it would be accurate enough to let the differences come through, but it did. I could hear the same difference through it too.

To make things clear: I am not able to hear differences due to (a) changes in the RCA cables, (b) changes in the digital filter in one of the CD players -- it has two. (The other CD player does not give any choice of filters.) So, there are many things which are not impacting the sound in any way I can identify.

The subjective problem with Player A is not there if I listen to Hindustani Classical vocals. With that class of music, Players A and B sound different but both are enjoyable. But with Frank Sinatra, the subjective enjoyment is affected, partly because a lot of Sinatra (and Shirley Bassey, etc) are recorded with an edge to their voices, and partly because their orchestra arrangements have violins, etc which trigger that edge and make me shift to Player B.

Then a third little chapter was added to this -- I was lent a professional CD player (which is used for installations, 1U rack mount). Player C. The sound of this third player did not have any of the edge, any sharpness, of Player A. It was pleasant to listen to. But when compared side by side with Player B, it had less detail retrieval, and its sound had music notes which were kind-of one-note. For instance, the hard drumming in some of the Supertramp tracks sounded like "one note" from Player C and had layering and textural detail in Player B.

Why now? What changed?

I've been listening to CDs (and to FLAC ripped from CDs) for about 30 years. But I've never had two CD players at the same time in my bedroom, both of which are expensive enough for me to expect no audible differences. Plus, I've never listened to CD players (or any other digital source) other than USB DACs through revealing, high-resolution IEMs before, because these current amps were a very recent addition. It's possible I won't notice any differences if I use just speakers and the room modes start playing up.

My distress

This challenges my earlier beliefs. I used to believe that the DAC was a "solved" problem, and any two DACs which are made today without being forced to cut corners due to pricing pressure will have inaudible levels of noise and distortion, and will deliver a flat frequency response. I am ok being told that the professional-install CD player has less detail and realistic sound than Player B, but I am upset to admit that Player A sounds worse than Player C. And of course, I'm upset that two expensive CD players sound audibly different, and more important, one very clearly reduces the subjective enjoyment I feel, while the other makes me feel like losing myself to the music in CD after CD.

It's upsetting when old beliefs, based on which I've guided my audio journey for 20+ years, are challenged and I can't hide behind denial. I didn't do double-blind tests, but short of that, I did many cross-checks, and I can't deny that (a) there are audible differences, and (b) they affect the subjective enjoyment of music quite a lot.

Two reactions I expect

Some of you will be like my earlier self, and will say "You're deluded. Just do a double-blind test and your claims will fall flat. There is no audible difference between two such players, or the output of, say, the the Fosi Audio ZD3, which costs one-twentieth of some of your CD players." This is what I would have said, in less charitable language, two months ago.

Others will say "We have been saying that every bit of gear differs in sound from every other bit. We've been saying this for fifty years. The cables, the power socket, the audio rack on which you keep your stuff, the thickness of the metal used to make the enclosures, everything matters. You fools have been laughing at us -- now the shoe is on the other foot." Except that you'll say it in less charitable language.

I expect that these two templates will account for 99% reactions.

I am hoping that the remaining 1% will find value in my experience.

Further work
  1. I will try measuring the performance parameters of the CD players using this test CD and my RME ADI2 FS Pro, for whatever that is worth. Better to have something rather than wait for someone with an AP5x5 to measure them one day.
  2. I have rips of all my CDs. And both these expensive CD players also accept digital audio in. So I'll feed these CD player SPDIF inputs from my RPi-Hifiberry-Digi-Pro with these FLAC files and listen to the music and see if I get these audible differences.
  3. I have one or two DACs: an Audiolab M-DAC and a Cambridge Audio 851D. I'll take the digital audio out from both these CD players and play them through these DACs, and connect my headphone amp to the DAC output to see if there are still audible differences.
  4. I will play some test tracks on both CD players (from CD-R I'll record from my FLAC files) and feed them to my ADI2 FS and record them to my PC at 24/96 FLAC, to share with you here (when I get the time -- don't hold your breath). I hope the Audacity software, which converts the audio to 32-bit float and then saves as FLAC, will not mangle the sound much.
So, I have a lot of work to do. But till then, I thought I'll just share what I've learned and heard.

BTW, this story is not just about distressing discoveries. One pleasant discovery was the transparency of the ADI2 FS Pro, in spite of its AD-DA chain. Wow. What a device. Second: the impressive transparency and effortless driving power of the Atom Amp 2. Third: Player B has an awesome headphone amp, with three levels of gain and really lovely, transparent sound. I tried it out for just a few minutes, but one thing I'm sure of is that even with the middle gain setting, it has enough power to drive the HD650 to beyond acceptable levels, so power is not in short supply. IEMs sound excellent too, with any gain setting -- no noise detected.

View attachment 1479890


View attachment 1479903
I had a nearly identical experience as you with CD players as you, 15 years ago. Substitute your son for my younger brother, and ditto that story too.

I’m an EE by education and designed speakers professionally for 3 years. Couldn’t deny the differences of DACs.

And…. It still pops up. A MindiDSP Flex8 is significantly better than a 2x4HD which is FAR better than a Behringer DCX2496.

I think instant AB tests mask these kinds of subtle differences rather than reveal them. The difference is heard in a different part of your brain.
 
@Logon - You can easily verify that FLAC compression followed by decompression is bit-perfect.
Thanks Ed, I appreciate your post and I ask you to follow my thoughts for a moment.
I don't listen to music through my PC, but a digital transport because I believe that the PC (its power supply) is a potential source of frightening interference for music reproduction.
Furthermore, the compression/decompression process of a flac file I do not doubt that it is Bit-Perfect, but you will converge that that process will ask the CPU of your PC to perform millions of operations more than a WAV file.
And this is enough for me, simply. 🙂
 
People on that site regularly take careful measurements of audio gear and publish detailed results, and also publish details of apparatii and processes they use for these measurements.

Off topic: except that they measure the noise of moving-magnet phono preamplifiers all wrong, because the boss of ASR doesn't understand what equivalent input noise current is.
 
  • Like
Reactions: A 8 and audio_tony
I'll never use FLAC format to listen to music at home because for me it is a contradiction in terms.
You can rip a CD track to wav, conver it to flac, and then convert it back to wav.

You will find that both wav files are identical (assuming the flac has no tagging information).

This can be done with a utility like md5sum - the resulting signature will be identical.

There's nothing in the decode process that can corrupt the audio, it's a well proven process.
 
There's nothing in the decode process that can corrupt the audio, it's a well proven process.
Sorry, but here I recognize a conceptual mistake.
That's not what we're talking about, but the fact that you're listening to your music being played WHILE your PC's CPU is performing millions or billions of useless operations related to get a good sound.
I don't know if I explained myself well.
The process of decoding the flac file has no use in obtaining a good sound, it's just an extra.
Anyway, it is an easily avoidable extra: just use WAV files! 😉
 
https://www.audiosciencereview.com
The most well known community on the Internet for audio enthusiasts who believe in the importance of objective measurements of audio gear. People on that site regularly take careful measurements of audio gear and publish detailed results, and also publish details of apparatii and processes they use for these measurements. The site strongly rejects advertisements to avoid conflict of interest, accepts donations, and permits audio gear manufacturers to send the members their products for measurements to be published.
That's ironic. It appears to be a site by the same Amir in the thread I linked you before from many years ago. Did you read it? https://www.avsforum.com/threads/establishing-differences-by-the-10-volume-method.1136745/
He did the near same exact thing you did. Why would there be any objection there to what you did??
If ASR normally include objective measurements, wouldn't you also need to provide those for your experiment? You appear to have done listening tests. Only site I know of where that is the objective standard, is Hydrogen Audio
There, you would certainly have to had used controls on your listening, volume level and blinding mandatory for any validity. Perhaps unsurprisingly I don't see any such requirement on ASR.
 
That's ironic. It appears to be a site by the same Amir in the thread I linked you before from many years ago. Did you read it? https://www.avsforum.com/threads/establishing-differences-by-the-10-volume-method.1136745/
He did the near same exact thing you did. Why would there be any objection there to what you did??

The objection is not from him, I haven't seen any post by Amir personally. But the thread you referred to is quite insightful, thanks a lot, I'd never have chanced upon it.
 
@tcpip Sounds to me like there are differences in the final analog output stage of each CD player that's affecting the sound - however slight, or dramatic it might be. No way to know for sure. The full details of the support circuitry for each DAC chip, and how each was implemented could very well affect the sound. Not sure what you mean when you say 'etched' sound, and since you didn't say how long you listened, I would have guessed you experienced a fatiguing characteristic to the music. But, this usually 'comes over you' after listening for while. That's the way all CDs sounded to me in the 80-90s. Of course back then my system was of lower quality too, and no where near the fully resolving audio chain I get to enjoy now. Today, however, we're treated to much greater dynamic range, and better S/N ratio with DACs and amplifiers - even with little inexpensive chifi gear. IMO, that's a good thing that helps to reveal a greater level of detail, and thus the full impact of the sound. There is a 'high-definition' sound to most well-implemented Sigma-Delta DACs today. I actually prefer the precise articulation of each and every note of the music on a well recorded track, but others prefer a 'softer' sound. Respectfully, I'm not a fan of the laid back softer sound. My .02 cents.
 
i found in a benchmark dac for 1000euros an 78xx regulator without the C's between the legs.

Adding these made a big sonic difference.

Similar things happen also in other expensive gear so its obvious that differences between exist.
 
I actually prefer the precise articulation of each and every note of the music on a well recorded track, but others prefer a 'softer' sound. Respectfully, I'm not a fan of the laid back softer sound. My .02 cents.
Neither am I a fan of this laid-back sound. And when I hear such sound, typically I hear less texture and detail.

Anyway, I think I've heard from all of the members everything which I could have expected to hear. I'll just record the analog output, publish the files, and we'll all discuss after that.
 
also on cheaper cd players the C's for regulators are often lacking, letting the more expensive models sound better who have them

this is all datasheet knowledge which was not implemented.

since i worked with dsp i have the theory that a tilt of the f response of a half db already changes the sound because it spreads over several octaves.