A small battery in form of large capacitors must be carried inside even the most modern class D amplifier for the simple reason that periodically, the voltage between the AC rails is zero, and such amplifier can’t get any power from the outlet. So it has to have tiny charge inside to still keep producing output during this dip. It’s just fact of life with AC.

The way this could be measured is via the voltage ripple, which is essentially the sag in output of the power supply when your energy draw during the dip exceeds the equipment’s store of electrical energy. So you can directly observe if your speaker load exceeds amplifier’s capabilities, by comparing the stability of the power supply’s output, I guess.

I’ve no idea why they think that having negative feedback is any issue, here. Seems completely unrelated to me. If anything, I’d say negative feedback is what makes your amplifier amplify correctly, at least to the limit that its other parameters allow.

Bass traps are most likely to reduce the height of the peaks, but do little to the valleys. The valleys are result of speaker-boundary interference, the reflected waveform from a wall – usually the front wall – mixing in with the speaker’s direct radiation, and canceling it. If you had very thick and large panels behind the speakers, that would reduce the level of this reflection. Dip of 20 dB suggests that the wall is bare and reflection unattenuated, and the frequency 100 Hz suggests that distance between speaker’s front baffle and the wall is about 86 cm. Halving that distance would move the cancellation up to 200 Hz which is far less objectionable.

If the measurements are the same, the performance is the same. If the measurements differ, then that difference can be audible. About 0.5 dB differences in frequency response can be discerned, and good rule of thumb is that it must have at least 100 dB SNR. Most DACs have performance that is considerably better than this relatively low bar. If you try to compare DACs that require superhuman senses and barely differ under measurement, chances are the effect is placebo or due to errors in test protocol such as not precisely matching the volume.

Yeah. Ultimately, when you have almost completely flat response, no harmonic distortion, accurate group delay, and small enough room reverb time, you’ve basically achieved sound perfection.

A sound system like that ultimately doesn’t sound like anything. It just sounds like whatever you play back on it, rendering it accurately. It is time to stop and just focus on the music at that point.

In principle you can hear everything that is going on in the music at that point. There is nothing left to discover other than what you already hear.

Audio hardware is not convenient at all. It is expensive, heavy, takes up space, and the more there is of it, the more likely something breaks or is not working correctly and damages the playback quality in some way that may be difficult to notice. It is also virtually impossible to guarantee that any change you make to the system will actually improve it, as that requires understanding where your most important problem point is and fixing that, rather than randomly changing some piece of equipment for another and hoping that it improves anything.

I find the inconveniences associated with equipment to be considerable, and so my tendency is to minimize it and stick to some known-good configuration that is vouched for by a reputable manufacturer, such as any active speaker that is a singular integrated box. For added peace of mind, you can just go ahead and measure the performance to observe that it delivers what is promised, so you know that not only did manufacturer promise a certain level of performance, what is actually delivered in your room by your total system is pretty much exactly what they said it would be.

This does not 100% eliminate the desire to “upgrade” your stuff, but a measurement/data-minded approach should focus your attention more towards where there might be actual problems. For instance, based on measurement, you may realize that your listening room is severely damaging the flatness of the audio spectrum in your listening seat, and that your room’s reverb time is excessive, and so you start to look into ways to add diffusion and absorption in the room, which are then genuine upgrades to your playback experience that don’t even involve changing anything in your audio hardware. You can also tell, in the end, if your realized in-room response is not appropriate, and you can fix it by simply equalizing it where there are problems, fix bass boominess issues, and even impart deliberate tonality changes such as some 5 dB extra bass below 100 Hz to make it sound warmer and nicer.

There is a risk that 18 ohm edition xs shows bad behavior with some amplifier outputs. If you have been using DT 770 Pro with 250 ohm impedance, there’s a good chance that the same headset amplifier, especially if it’s just some random laptop output jack, does not necessarily play as well against this much lower impedance headset. I recommend getting a low-cost amplifier like the 3.5 mm Apple usb-c headphone jack adapter which is decent amplifier with a known-good behavior, and in particular it has low output impedance, just to confirm that it still sounds the same. Sometimes, these headset jacks in laptops or computer monitors are just poor, for instance they might have some kind of DC voltage offset elimination capacitor that have too low values, or whatever, which could definitely create treble-heavy presentation if an unusually low impedance headset were plugged in them.

Compared to Harman target, Edition XS should be overall closer than 770 Pro, from what I can see. I think there are only two tonality issues: there is not enough bass to sound natural, and there’s no 1-2 kHz region boost, which are fairly common features on open-back planar headsets from what I’ve seen, whether made by Audeze or Hifiman. I think you need to engineer some kind of acoustic resonance or resistance to alter the sound, or you just get the default, and open-backs are in poor position to manipulate the sound as the technology is very simple, basically a straight wire through a thin membrane with linear movement characteristic as function of voltage and no back-pressure given by a cup. I’ve heard that Dan Clarke has some kind of complex metamaterial providing a sort of analog equalization for his open back headsets, as an example.

Finally, when looking at headset measurements, I encourage disregarding any measured performance at about 8 kHz and above. The data simply isn’t reliable, as frequencies so short are similar to the distance between headset’s cup and the artificial head, so there will be distortion in the measurements for that reason. My guess is that this is also why sometimes changing the pads help, as pad thickness might vary, and that could well change the distance between the headset and the head, and perhaps changes the resonances experienced by user.

I did get alternative cable for Edition XS. I agree that its default cable is surprisingly crappy for what is a fairly premium headset.

In my opinion, they can be almost the same, but a 2-way is really pushing that design to as far as it can go, if the goal is to have a full-range speaker system. I have this 2-way speaker https://www.genelec.com/1032c#section-technical-specifications and you can see the fairly close to on-axis crossover holes around 2 kHz in the vertical directivity plot, and the promise that harmonic distortion says below 0.5 % only, which isn’t super great. Harmonic distortion vanishes as soon as the tweeter becomes active, at least as far as I can measure it.

Anechoic response starts to fall around 40 Hz, but the rule of thumb here is that response extends about 30 % below the anechoic limit, and I dialed in some extra digital emphasis for the 20-30 Hz region, and so I managed to get something very close from 20 to 20000 Hz response out of a 2-way system. Harmonic distortion less than 1 % is considered inaudible for complex signals.

I made a paradigm shift from passive system that was pushing 20+ years to an active system using modern technologies, so I essentially swapped out everything. Plus, I did go for maximum convenience with active speakers, so all that was needed beyond them was the signal source.

I want to hear the music as clearly as I can. To this end, I spent effort on identifying and setting up the gear that is capable of reproducing a flat sound spectrum with very little linear or nonlinear distortion. Then, I realized must install lots of acoustic paneling to control the reverb time and reflections. Once that got done, I realized I do have to deviate from flat sound spectrum, and put in about +5 dB bass boost to mimic equal loudness compensation contour for 10 phon below reference level. I found the curve I am using from some student’s University paper.

The journey has left me with sound that has what I think has lot of clarity at any listening level, and has a very pleasing tonality, with warm, mellow bass tone. The gear simply vanishes now from my mind, and I concentrate on the music.

Monty from Xiph is on record saying that even no dithering is perfectly fine for 16-bit audio. This is probably true. The quantization noise is really quiet, you simply aren’t going to turn your equipment up enough to hear noise at some low level like -96 dB, or wherever the quantization exactly ends up being. There is probably enough background noise in the signal to begin with, to make the whole question utterly irrelevant – if there is already a broadband hiss at, say, -80 dB, it hardly matters if you add tiny amount of additional quantization noise at some -96 dB level. It’s just going to vanish there and you can never tell it was present in the first place.

That being said, I do recommend use of dither noise, and triangular dither is not bad. These settings you have there control the noise shaping, which can somewhat extend the retention of signal at hearing-criticial 1-2 kHz region at expense of more noise in less-critical bands. I think any choice, including “None” for flat dither noise, is perfectly fine. U shape dither is likely technically the best choice, suggesting by name that it places dither noise at low and high frequencies roughly following ear’s sensitivity curves. If you had some signal near, -96 dB to preserve, then you would likely want to make this choice to ensure that as much of the audio in the hearing-critical bands would be preserved as possible.

I couldn’t hear any difference personally. Granted, I didn’t spend a whole lot of time on either sample, but to me they sounded exactly the same.

My system only has 48 kHz digital audio link from an external soundcard to the Genelec speakers that I use, and it is possible that a 22 kHz signal would already be largely filtered out in the playback chain because it is above 20 kHz and if a sample rate conversion to 48 kHz is needed, some low-pass filtering will be applied.

Testing stuff like this would take some care to ensure that the test is even valid – I’d need to use a microphone to make sure that any ultrasonic sound is even there.

I think almost anyone who spends effort to put vinyl on equivalent digital footing would approach this quesiton like this:

• the equivalent bit depth of vinyl “samples” is not even 16 bits – the noise floor is higher on vinyl than on a redbook CD. A vinyl “sample” might be 14 bits, at best.
• The sampling rate is not infinite, either, that which you referred to as “infinite bitrate because it is analog”. The reason is that your stylus is not one atom wide point sampler, but rather a kind of gigantic pyramid in atom scale, and the material and its cutting process results in grooves of specific minimum size that is practically achievable before the features are either too thin, or the recording head too brittle, or the material heating too much due to friction of the playback, or whatever. This actually affects the maximum frequency that can be stored and played back, which can be used to compute its sampling rate. I have heard that vinyl might go up to 50 kHz, but that just means that digital sampling rate 100 kHz and above is superior, then.

If you understand digital audio, these are actually 100% valid arguments that put vinyl on equal footing with digital, and suggest that any 16-bit hi-res file is already likely to be technically superior. There remains the question of whether it matters that you get ultrasonic content to play back, because you can’t hear it, our amplifiers may have trouble amplifying it, and our speakers are not likely to play those frequencies back well in any case due to being inaudible to humans.

Well, there exists formats like SACD that technically have no DAC, in a manner of speaking. The “garbage” 0s and 1s there run at such a high sample rate that they average out to the correct signal with a tiny bit of low-pass filtering. This guy is the god of explaining precisely and correctly how digital audio works, Monty from Xiph: https://www.youtube.com/watch?v=cIQ9IXSUzuM

Audiophiles are conservative, with a small c. The way things were done some half-century ago, with glowing vacuum tubes, ancient horn technology to match the low efficiency of such amplifiers, and analog recording mediums appear to have seeded a kind of culture which changes only as the generations themselves change.

As you correctly observe, the source of music production is almost invariably digital. Writing it into analog medium for playback makes relatively little sense compared to just offering the digital data.

I like the ambience of the artist called Bluetech. For instance, I often listen to this Phoenix Rising album, which had a 2021 remaster. It is less dubby than most of his work, and it has many musical elements like melodies and harmony rather than being some kind of experimental noise.

Also, for music I actually like, it is important that it doesn’t have a steady rhythm. It can be danceable in some sense, but it better be doing something more interesting than just go boom-boom-boom-boom throughout.

I’ve used 8330A+7350A with GLM. The setup is good for flat response, for getting an easy to setup room correction, and perfect sub integration with mains, and it is relatively cheap system in total, considering what all you get. I think it is a winner.

However, I ended up selling the setup with preference for bigger monitors with larger drivers and no sub. I’m now using 2x 1032C in my living room which comes with 10" woofers and which I found used in a local audio shop, so not much price difference compared to brand new 8330A pair and 7350A, actually. I decided I wanted two sources of bass in the room to get room modes better under control, as singular sub didn’t seem to work ideally in any location that was practical in my living room. I also tended to pull the subwoofer’s level up by some 3 dB above calibrated level to make the bass sound more natural, as flat response is slightly thin-sounding unless you listen at reference levels where ear’s bass behavior is roughly flat also. I think Genelec’s Sound Character Profiler technology does not apply the frequency response correction curve to the sub, only to the main monitors, and this was annoying because it meant that I couldn’t adjust the response correctly from a single UI, but would have to combine the SCP filters with a manual level adjustment of the sub.

I do not see anything wrong with monitors for casual listening. Genelec speakers are highly accurate, and the DSP is good, and you can adjust the character of your sound output straight into the speakers, like dial in some few dB worth of bass boost, which I like to do. The bass boost, even if it is only a few dB, is a huge deal in making the sound wonderfully warm and powerful. These are flexible monitors that are surprisingly small thanks to the metal enclosure, have wide dispersion angle, low enough harmonic distortion to be essentially perfect if sub is crossed over at around 100 Hz, and you also get basically flat group delay so that all frequencies hit the listener simultaneously. The combination makes for accurate, revealing sound which is really unlike anything I heard before. The only limiting factor in performance is likely to be your room, so budget for panels, you will want them once you see the GRADE reports from GLM.

I am surprised at the uneven response achieved. Technically the equalized response looks somewhat alright to me, except that it clearly lacks for about 5 dB worth around 100 Hz, which is quite critical for the feeling of punch and warmth in the music.

There is supposed to be a general decreasing slope in the in-room response measurements, worth about 5 dB across the entire frequency response, but also an additional loudness correction bass boost is acceptable unless you regularly use very loud listening levels which do not need it. I think another +5 dB tuned around 100 Hz should be added, using broad filters. With enough bass, the speakers would sound much warmer, I think.

In my opinion, the uncorrected response below 1 kHz looks alright except for that big peak around 37 Hz that should be pulled down by some 5-6 dB, maybe. That would be the only correction I would want done, personally. Above 1 kHz, assuming the tonality corrections are very broad and smooth with low-Q (resonance factor) filters, they are probably an improvement also, as the uncorrected response there is very strange-looking. These speakers are not neutral at all, for whatever reason.

My opinion is: wide dispersion cone from the speakers, with frequency response that extends to 20 Hz, is the big sound. I have got a pair of Genelecs that do it and orchestral pieces are just massive sounding, for instance someone hammering the timpani makes a punch you will feel in your body. It very inherently sounds “large”.

The wide dispersion angle’s effect is more of a conjecture. Whatever kit you buy, it has a dispersion angle inherent to its design, and that’s what you are stuck with, so it is not easy to show that this one matters. It’s also a parameter that rarely is measured and shown in speaker specifications, but sometimes you see off-axis measurements from the manufacturer and you can estimate how much the level is reduced at, say, 45 degrees off axis horizontally. If it is about 6 dB or less, that is definitely a wide dispersion angle speaker. In any case, if much of the sound energy is radiated also on the sides of the speaker, it will reflect from the room’s geometry and that usually gives it a spacious, expansive feeling.

The encoders are a moving target. The output bitrate is a hard limit on the output rate of the encoding, but there are many choices that encoders can make on how to represent the audio at that rate. For instance, there is compression involved, which is sensitive to minute details of the data encoding and trivial changes to data can compress differently. There’s a byte reservoir that allows use of more bytes temporarily if particular part of music needs it, but you have to pay that “loan” back by using less bytes later. The psychoacoustic models may have become better, and increased CPU power can also be used to perform more exhaustive search in the encoded audio space for the best representation.

I guess I’d say growing acceptance of class D amplifiers, built-in DSP, and similar digital response linearization techniques are new. Especially studio monitor space is nowadays looking a lot like it’s just a computer with transducers built-in into its box.

The epitome of that design is the Genelec The Ones series which are a concentric point source technology, just completely uniform directivity whether vertical or horizontal. Completely digital inside, analog is now an afterthought and no longer preferred way to input audio to these speakers.

Figuring out how to make designs properly work where the tweeter is inside the midrange is also somewhat new, concentric drivers tend to have discontinuity and waveguide’s motion related irregularities which cause warbling or modulation distortion. But 3-way concentrics could have done that for some time.