I personally appreciate the information and opinions you’ve shared. Thank you.
2 Likes
Thanks! Just this makes the time spent worth it 
3 Likes
Can you explain how, or even if, amplifiers can change the sound/tonality of music?
Not quite immediate danger tho. It takes 120 dB_SPL for permanent hearing damage, you can listen to 96 dB just fine if it’s for 1h or less. But CD-res with additional noise-shaped dithering does go up to 112-120 dB of dynamic range, so in the end yeah, it covers everything you’re going to need in realistic situations, especially considering that you won’t normally have a noise floor of 0 dB in your environment anyway, so that already makes the bottom 20-30 dB of that range useless.
3 Likes
Yes, I didn’t go into dithering because that would make an extra long post and potentially even more confusing. But you are indeed correct, so I don’t think it’s ever possible for me to listen to the noise floor of 16 bits audio at the levels I always listen to (70-80 dB).
Even in the “8 vs 16 bits” test here The 16-bit v/s 8-bit Blind Listening Test, Part 2. I had to turn the volume quite a bit up to correctly identify the files.
Yes the amplifiers can indeed change the sound by altering the frequency response of the input signal. For example, tube amplifiers are popular since they cut some amplitude from the treble range and/or boost the mid-range.
I personally don’t see much value in this part of the audio chain altering the sound, since the output device will do so no matter what (no speaker/headphone is perfect).
But amplifiers can be made to output the input sound perfectly and since I value knowing what am I listening to, I would rather have just the headphone/speaker be the ones that do the coloring.
Maybe I am just biased against tonality changes since I have a preference for acoustic music, and you can instantly tell when an instrument doesn’t have the correct timbre. For other kinds of music or maybe if you don’t care too much about accuracy, sound-altering amplifiers can have their merits. Or maybe if you are bored and are swimming in audio gear?
But as I explained above, if a DAC alters the sound, on top of the amplifier and headphone/speaker, you will never make heads or tails of whatever you are listening to. I would stay away at all costs from these, since a DAC’s only purpose is to convert from digital to analog, perfectly. Fortunately, I know most modern decently-priced DACs output the sound exactly as it came in, but there are exceptions that try to make themselves sound “better”.
I wouldn’t have doubted it was “real” even for a moment, since I have seen even worse sh*t out there and the audiophile market is the worst offender of them all. But the quote from The Absolute Sound sounds funny so this may be for satire purposes.
1 Like
In theory, you’re right about how 44.1/16 is perfectly fine and enough. However, the theory is a perfect mathematical model. In practice, every part of the DA chain, from the recording, over the mastering, DAC, AMP, transducer right through your ear wax, can only ever degrade the signal.
The mathematically perfect model can only be approximated and the better the approximation should be the more processing power is needed. On top of that, their accuracy depends on accurate HF clocks beyond what a typical, cheap quartz or clock circuit can do. On top of that, every electronic circuit is prone to picking up noise or being disturbed by EFI, voltage variations, etc… Electronics are, at some point, analog devices, too… very small ones, yes, but that makes things worse, actualily, because they become more prone to environmental disturbance or quantum effects - stuff that needs to be corrected as well - in specialized hardware or codec software running within the ICs.
And that’s just a rather superficial look at why ADC / DAC devices do make a difference or why power “cleanliness” can have an influence if such devices use shitty PSUs. The story continues like that for AMPs, cables (in case of shitty cables, admittedly) and even hearing itself, which is about more than just the FR of your cochlea, as your brain also does all sorts of things with the auditory information it picks up before it enters your conciousness - btw: a critical part of the chain, we understand virtually nothing about. Virtually nothing, because, actually we’ve known for almost two decades, that high frequency sounds beyond the hearing threshold, while not percieved as sound, DO affect brain activity (although we don’t know what it means, as it seems). Google “Hypersonic effect”.
That said, I’m really tired of oh-so sciency, engineery people, who argue 44.1/16 is perfect and DACs make no difference, by over-simplifying critical parts of the audio chain that in reality aren’t at all simple but a complex, hot mess. Is it good enough for most mass-marketed music and most consumers? Yes, absolutely! Is it perfect? No, because reality only approximates the mathematical, perfect model and human auditory perception picks up all kinds of things that today we can’t even measure, let alone model accurately enough to draw conclusions about.
2 Likes
What you said can go both ways. I don’t think there are just a few DACs out there that don’t have the capacity to properly output hi-res. So the chance is not small to actually degrade the sound by feeding a DAC some experimental music format that it cannot properly process.
But regarding what you said about the hypersonic effect, I am actually quite familiar with the subject. Someone theorized it some time ago, but tests were made a few times on people and nothing perfectly conclusive was found out, so I am not about to count of this until proven otherwise.
However, what I will always take into account is that high-frequency sounds are always more fatiguing than low frequency sounds. And even the sounds you can’t hear can still damage your hearing or cause you headaches. And as someone that listens like 6 hours per day to music, it’s frankly unacceptable to get tired by noise I can’t even perceive.
Yes, often times just noise (not like anyone ever heard it at that frequency)
ALWAYS check your “hi-res” files. And people still won’t even believe the crap they are being fed. Just because a piece of equipment can technically record/output hi-res, that doesn’t mean they can record/output anything more than noise. Pretty much the case with vinyl as well.
1 Like
Dunno if your double negation was intended but what you said was that you think there are many DACs out there that have the capacity to properly output hi-res. Well people hardly have the gear to actually test that, so they have to trust the specs in that regards and from that point of view, you’d actually be hard pressed to find a commercial DAC in production today that doesn’t have the capacity to output hi-res. Also there’s nothing experimental about higher SRs or bit-depths.
The facts on the hypersonic effect are that a physiological reaction could clearly be measured but that there was no conscious reaction to it, i.e. we know it does something we can’t consciously perceive and we don’t know what other effects that something has or doesn’t have.
My argument for hi-res is mostly that 24 bit depth can be sensible for albums with very high dynamic range, where all tracks are aligned to the peak of the album, which might also be set below absolute maximum. In such cases, having a lower noise floor is beneficial. Also I’d rather have 24 bits than relying on dithering done right.
As for SR: At least in initial recording this should be as high as possible, because directly recording to 44.1 will result in noise >22khz being rendered into the audible range, as you know (I assume) and analog filters are, for several reasons, less desirable than digital ones for cutting of frequencies. If downsampled correctly, 44.1 is fine for most music, but still not perfect because we’re not listening to sine-waves but chaotic waveforms, which might have features in time-domains not covered by 44.1 but still perceived when listened to, nonetheless. Of course, this also necessitates transducers capable of acting just as fast, but these exist in the form of planars and electrostats.
Besides: Watch Z’s video on these weird sony speakers he released recently - they have super-sonic tweeters that are solely there for producing inaudible frequencies. He said he could perceive a difference of them being covered or not, so, yeah.
Sure, if fatigue is an issue, because you’re listening to music most of your wake time, then getting rid of potentially fatiguing components makes sense. But playing your own ball back: the transducers probably make a much bigger difference for that than the audio format you’re playing
still nice how you just didn’t, at all comment on my main argument which basically was “high end gear and sources aren’t there to try and make audio reproduction better, they are there to try and make audio reproduction less worse” 
1 Like
Well, I think you are smart enough to draw conclusions also based on reason, and not only on “feel”, “common sense” and “faith” . When it comes to me, I am just tired of being fed snake oil by the audio industry. Cables are one facet of the industry and I am convinced hi-res/DSD is just another one.
When it comes to most “audiophiles”, the flow is probably something like this:
=> They hear something that at that moment they think sounds better
=> so it’s definitely better.
When it comes to me, it’s more like this:
=> I hear something that I think sounds better
=> it’s possible my mood or my tiredness affects the sound
=> let’s inspect the waveform and spectogram
=> let’s do a blind test
=> let’s listen tomorrow as well
90% of the time it turns out my hearing sense was just very unreliable and prone to placebo as always. The rest of 10% of the time it turns out it was a different recording entirely 
So I just came from the forum on the r2r dac from massdrop. Someone exclaimed how wonderful it was, and how it brought up the mids. As a dac is this claim reasonable? It is based on an older ladder design. Does the ladder design give it any advantages compared to “normal” dacs?
Certain DACs colour sound, so to claim that an r2r DAC sounds better is not unreasonable. The r2r design was only replaced with delta-sigma as a cost saving measure.
If r2r is properly implemented, it can exhibit more precise digital to analog conversion than delta-sigmas.
1 Like
I’m pretty sure it’s child’s play to implement a cheap dac with a perfect frequency response these days. The only things they can improve with more expensive models is lower THD (which should be undetectable by humans at this point) and maybe better overall electronics inside to remove the electrical network obligatory noise.
So, if someone claims a DAC sounds better than the next already expensive DAC (and we are to assume it’s not just placebo), then the only explanation is that the DAC alters the frequency response of the input sound. A different FR is never inherently “better”, just something one might prefer over some different sound. I would say the only objective "best’ sound is that of a flat line, since then you can trust the source exactly and therefore you can alter it with the amp and speakers/headphones in any way you want.
And when you increase and decrease parts of the FR, there will always be music that sounds better and music that sounds worse on it. There is no uneven FR that works better universally. It’s always give and take.
One other troubling thing I heard is that some DACs like Chord Hugo output a higher voltage of sound into the amplifier, so the final volume would be higher than what you are used to. And what do we know about higher volumes? We automatically tend to think it sounds better. That’s a nasty pitfall to fall into.
3 Likes
Isn’t the real difference between dacs that audibly sound different the implementation of the dac by the manufacturer? I’m sure there are certain imperceptible differences between the actual dacs themselves, but aren’t the real sound differences are created by the designer working to implement the chip into the device?
That would be an amplifier thing, to change sound depending on the implementation. I have noticed people have trouble grasping the role of the DAC and what it should do different from an amplifier.
Allow me to re-explain. Just like the first video I posted explains, a digital signal can only have ONE possible analogue sound wave solution, and that is the original recording perfectly reproduced, no imperfections, no room for even more “resolution”. A DAC’s singular purpose is just that: to give you that single unique signal that corresponds to your sound file. After that, it passes the analogue signal to the amplifier, integrated or external.
If a DAC that is not built extremely crappy sounds different from another, then it means it has been tampered with to make it sound “better”. And that is never a good sign. Ideally you want to know exactly what you are listening to, so a DAC without any extras is the best.
Afterwards, a perfectly uncolored sound can be colored in any way you want with the amp and headphone/speakers.
1 Like
Gotcha, just wanted to make sure that the real coloration comes from the rest of the system, not the dac.
Except ofc that it’s not their job either - the amp or the headphones - to color the sound intentionally, they just do it because “they can’t help themselves” (a.k.a. technological limitations). Strictly speaking only an EQ is meant to color the sound to a user’s liking, everything else does it because of poor implementation (or amp implementation by people with poor understanding of what an amp should do 
).
2 Likes
Agree, I prefer the sound to be as close to the original source as possible, that is why I use a neutral amp and tend to use as close to neutral output devices as possible. But here is already the domain of preference, so I can’t go tell people how they should listen to their music.
That’s how it should be at least, because it seems that DACs that color the sound do exist. I have a feeling it’s those very expensive ones that people praise they make a night-and-day difference, but I haven’t heard one that sounds clearly different so far, unlike with AMPs.
Or AVRs, probably. My old Denon AVR-1312 seems to be “U-Shaped” a bit (just a +3dB sub-bass boost and +3dB… somewhere over 10khz, compared to my other DACs). No difference between “direct mode” or “stereo” (Stereo just activates upsampling).
I can easily A/B the DACs using the Analog in, bypassing the DAC inside the AVR, using it only as an amp.
Not the R-2R design per se but rather the optional NOS (non-oversampling) and filterless operation mode that some R-2Rs come with leads to substantial differences in sound, primarily:
- pro: time-domain behavior is near-perfect, unlike in typical OS (oversampling) DACs these days, which slow down all the transients and add pre-ringing and/or post-ringing to the impulse response
- con: frequency response is worse than in OS DACs, exhibiting severe treble roll-off starting around 2 kHz and going down to almost -4 dB at 20 kHz (in the Massdrop x Airist you were looking at this is solved with a filter, so it’s not a pure filterless NOS design).
I’ve been reading up on old-school NOS DACs since I found out my upcoming iFi Micro iDSD also has a NOS (and filterless) operating mode, and I’m getting the impression we’re still not at the point where “the DAC problem” has been solved perfectly in every cheap device. You need something like NOS/filterless operation to get perfect time-domain reproduction, but you need OS or hybrid digital solutions to get perfect frequency-domain reproduction. Maybe there are expensive digital hybrids out there that can achieve a near-perfect compromise between these two design goals, but I don’t know about them and probably can’t afford them. Just the fact that recent highly popular designs like the Micro iDSD and the R2R-11 include both operation modes and a switch to choose which one you want already tells me that this problem has probably not been resolved yet within a mass-market friendly budget.
2 Likes