Take Me to School (Understanding Audio Better)

Technical Discussion
41

Hello!

Well, it’s been a while since my last time here, I was trying somethings, some earbuds and a headphones, there where in the cheap end of things, I was just trying to listen to things and try to put them in context…

From the earbuds (there were a TWS earbuds from amazon, can get a link for the specific ones), since i still trying to get sound in context i don’t fell like doing a review of them, specifically 'cause it would not be that good and objective, however, i find my self in some trouble with those earbuds, the principal one is when i use them, i have to kind force them in my ear, but they slip out (or like they star slowly coming out), and it’s so uncomfortable and frustrating to have to put them in place like each song, and something that make it worse is the fact that when i touch them, they stop the music… :triumph: :rage: having to do that is so time wasting and so stressing that i have to stop using them within minutes…
Don’t know if changing the tips by other’s since the one it have are plastic, hope i can get some thought on that, other thing is that one of the pair don’t “fit properly” in my ear, i think it’s just my ear shape but still want some opinions on that, just to position my thought’s on earbuds.

The headphones in the other hand, i’ve been using a peohzarr model: AB078,

, there’s a like to the ones i talk about, there are close-back i think and they work perfect exposing the fact they told me of the sound close-back headphone generate itself, other than that in the sound quality they have way too much bass, like way too much, the vocals lose somethings in the loudy bass, i would like to get some thought if someone know those headphones (which don’t think would be the case) or some like it, that can get me some feedback on the sound quality and how i could place the sound it produce, i mean, like if the mids are bad o acceptable or if the highs are a joke or acceptable. They have what i describe like regular pads, they are kind comfortable but the material don’t feel durable and it look like it could get stinky with time, it give me something if been looking for a while (not in the way or quality i was thinking but it’s something…) which is isolation, if i’m in my room i can’t hear what’s happening in the living or the next room which is awesome. The arch or headband, not sure how it is call, well it’s just some aluminium wrapped with some foam-like material. The cable is made of a cloth-like material, was wondering if that’s durable… but that’s my thought, wish i could let more info on the sound but well…
I’m just trying to get some context of the sound i’m hearing and of course, i know those gears are not the best but i get something i can afford (at the moment) and something to star with.

I thank in advice whoever can give some thought/opinions and help to improve.

1 Like
42

Hi! This seems to be the right place to ask this question, and forgive me if it has already been asked: can (or should) I run a dac/amp dongle into another amp?

I do most of my music listening and movie watching in bed from my iPhone, and the only way to have peq from an iphone is the qudelix 5k (absolutely love this thing).
It of course powers any IEM I throw at it and the few headphones I’ve tried, which are all quite sensitive, and I’ve only owned budget gear, so haven’t really concerned myself with dac performance yet.

However, I’m looking to upgrade to an Audeze LCD-2F or LCD-X, and as for source equipment, what should I go with based purely off of sound quality?

1- Just qudelix 5k
to avoid amp into amp, which would have compounding noise and distortion.
but enough power?
According to asv, it can pull 200mw at 50ohms when using the 2.5mm balanced.

2- Qudelix 5k into Atom Amp+
for the extra juice and headroom.
but has two amps in signal chain with noticeable compounding noise and distortion?
The qudelix can go up to 2V SE output or 4V balanced output. Can you even run a balanced output into a single ended amp safely?

3- Neither.
Something like Atom+ stack or DX3 pro+
(using equalizor apo/peace on laptop)

And if there are noticeable sq differences
(which I’m assuming there probably are),
are we talking about tiny ones
(like between cables),
or night-day differences
(like a wide-Q 2dB eq boost or cut).
The reason why I ask is because if the difference in sq isn’t dramatic, I’d like to go with option 1 or 2, but if the differences would be clear, I’m willing to get used to using my laptop in bed.

Thanks!

43

A few years ago, @Torq answered this in response to a similar question I had. But I’m not certain where. It’s never a good idea to send an amp output into another amp - you will get noise increase. I was using a DFB at the time and you use a Qudelix, but the key here is that neither of them has a true line out mode. I just looked at Qudelix support and they don’t have a real line out mode.

Any DAC or USB DAC that DOES have a line out mode can be and should be used with an amp. The line out mode bypasses a lot of the internal circuitry that will produce noise, etc.

I didn’t really think this was true at first (but I’ve learned that @Torq is almost always right when it comes to stuff like this), and when I later got a device that had a line out mode the difference was obvious.

I still use an Apple camera adapter 3 with the iPhone so that I can charge and listen for extended periods. Possibly magnetic charger for newer iPhones would solve some of the issues. For very portable use, I go with a Dragonfly Cobalt (usually) which has OK power, but for more oomph I have my iFi xDSD. The new xDSD Gryphon would be preferable I think, at least the reviews and specs look nice.

There’s no point in running a balanced output into a single ended amp. The amp is expecting 2V input. And you really aren’t gaining any quality difference by using balanced in that way.

What kind of budget do you want to spend? I’m sure if you post this in the audio advice area you will find people willing to help you drain the excess in your wallet.

44

Absolutely agree with this.

I really wish the BTR5 had a line out mode so I could use my FiiO A3.
The FiiO performs great when using a proper DAC with line out.
But when I use the BTR5 as a “DAC”, though the A3, it just sounds soft and dull.

The gains in volume I’ve gotten have been far outweighed by the loss in sound quality.

1 Like
45

Let’s talk about line outputs and headphone outputs for a moment. What do we mean when we say that something has a “line out”?

Well, classically, we mean that it has a relatively high intended load impedance - that is, that it isn’t expecting to deal with a 32Ω headphone or a 4Ω speaker - and that the level isn’t expected to exceed “line level”. Well what’s line level? That’s a bit arbitrary - for audiophile headphone gear, 2V for single-ended and 4V for differential seems normal, but it can be much higher for some pro gear, and conversely the “standard” audio line level values are a bit lower (though nobody actually makes gear that’s intended to hit max output from -10dBV, at least not that I’m aware of).

By contrast, a power amplifier is expected to drive a load that’s lower impedance - and thus more challenging - and likely to output more than line level. 10x or more gain is typical of power amplifiers for speakers, and most headphone amplifiers have at least one non-unity gain setting. Since adding gain naturally amplifies upstream noise (and costs you some of your feedback, further worsening noise), it’s good policy to avoid attenuating the signal after amplifying it - and since volume controls are almost universally attenuators, this is why it’s ideal if we don’t have more gain than we need to hit line level before we input to our headphone amplifier.

Highly powered output stages can also be noisier, higher distortion, or otherwise not as nicely behaved as a line level output that’s never expected to output more than a few mA. The design priorities differ, at least in the case of speaker power amplifiers.

But what for headphone gear, it’s a bit less clear. After all, most amps do have a unity gain mode (so you’re not amplifying then attenuating), and the levels of noise and distortion in a high-quality headphone driving circuit can be just as low as in a line driving circuit. Indeed, because headphones require vastly less power, your average high quality headphone output looks an awful lot like a line output that happens to have perhaps a little more output current or somewhat lower output impedance than normal.

Now this isn’t to say that you should make a chain of 15 Schiit Magnis, each running into the next, but it does mean that there really isn’t that much to differentiate a unity-gain headphone output from a line output - some of them may distort or be noisy, but that’s also true of some line outputs, and a well-designed output intended to face a headphone load should perform just fine as a preamplifier or line stage.

This isn’t true! Differential connections actually make more sense as interconnects - the inputs of amplifiers are high impedance, and long cable runs tend to act line antennas for interference - with a differential connection, you get a significant rejection of all the common mode noise that’s picked up this way. I’ve never had a headphone or speaker pick up noise on its own, but I’ve had quite a few single-ended connections between stuff pick up noise of one sort or another, and balanced connections avoid that - that’s precisely why they’re ubiquitous in pro audio, where long cable runs and noisy environments are common.

46

Can you comment on the relative advantages of a differential cable over a shielded single-ended interconnect (as I look at my setup with some Worlds Best Cables RCA lines having directional floating shields)? The vendor reports excellent noise rejection numbers, but also states that shielding can warm the tone. I’d guess that differential better maintains the tone at the same noise level.

Confirmatory comment: I have an Android phone for gym/workout use. It has an integrated FM receiver and uses a single-ended earbud/headphone line as an antenna.

47

I mean, it’s a bit apples and pears there - both differential and single-ended cables can be shielded, with a conductive sleeve connected to ground around the signal line(s). Indeed, I can’t think of many differential interconnects that aren’t shielded.

All else equal, with a differential connection you’re reducing the common mode noise - the stuff that’s present on both the hot and cold signal lines - by your device’s CMRR. That can be a lot - tens of dB! But if your common mode noise was initially low - either because your run is short, you don’t have a very noisy space, or your shielding was sufficient - this doesn’t mean it’ll be an audible, or even measurable, difference.

If your shielding is changing anything to do with the tone, you’ve got an interesting problem on your hands, because it shouldn’t be interacting with the signal you’re carrying at all :rofl:

1 Like
48

Yes, I’m quite aware of this. And I didn’t ask about the length of the cable run. In my system and even in mobile applications, I find I am usually under 1meter, always under 2. I also tend to use reasonable quality connectors. So I was discounting the effects of longer cable runs.

But what about overloading an input? The balanced amp is still expecting 2V.
And does the Quedelix have a unity gain mode? When I was looking at the Dragonfly devices, I wrote to Audioquest, and they wrote back that running them at full volume is NOT, in fact a “line out” even though some of their literature states it is. I recall doing this during my discussions with @Torq .

49

FWIW, I’ve had significant common mode noise in single ended connections that were 1m or less - the world is a noisy place!

Input sensitivity isn’t quite so simple, because headphone amps are - generally - not fixed gain devices like power amps. A power amplifier will have a specified input sensitivity for its maximum output - say 500mV. Exceed that and, beyond probably going deaf, you’ll clip it.

For a device with adjustable gain, however, your maximum input depends on your maximum output, your level of gain, and any attenuator (e.g. a volume pot) that exists between your input and your most rail-limited portion of the circuit (which, since headphone amps are generally using a single set of voltage rails, will likely be the gain stage unless the output stage is quite far from rail to rail). Take the Atom Amp+ that @GabeM mentioned above - that will output 5.7Vrms or so at 32Ω and it has a unity gain setting, so the maximum input level at that load would be…5.7Vrms, because it’s applying no gain. If the potentiometer is at the input (not necessarily advisable policy, but not uncommon), it could take more than its maximum output voltage, so long as the pot wasn’t maxed, although this isn’t really what I’d do myself.

This all is a bit aside for @GabeM , however, as the Atom+ has only RCA inputs, and the Quedelix has a full scale voltage of 2Vrms from its single ended output. Well, unless he got, like, a balun, I guess, but that would be weird.

Based on their volume control page, it would appear that it only has a unity gain mode, with its max output being the buffered output of the DAC - which isn’t surprising, it’s a dongle.

I have no idea how Audioquest characterizes a “line out”, but I don’t recall any of the Dragonfly DACs having exceptionally high output - and being USB powered, that’s a bit of a pain to achieve, so it’s not surprising. I will note that unity gain isn’t required for a line output - given that it’s a DAC to begin with, do we even have a reason to care whether the voltage level comes from the full scale level at the DAC outputs or subsequent gain inside the device?

50

If I interpret their comments correctly, this mostly pertains to capacitance and the properties of multi-strand cable weaving and twisting methods. Gotham is said to be highly shielded but less bright than Mogami and Canare.

51

So there’s a bit of a “two truths and a lie” thing going on here - or, if not a lie, at least “something which isn’t in any way demonstrated by those two truths”. It’s true that that Gotham cable has a higher capacitance, about 103pF/meter (or 31.4/foot)
image
Vs the Mogami 2549 which he describes as “bright”, which is about 30% lower capacitance at 76pF/meter
image.

It’s also true that a parallel capacitance to ground, like that between the signal conductor and the shield, will form a low-pass filter when combined with the resistive component of a device’s output impedance. If we had completely zero ohm sources, this wouldn’t be true, but we’d also see a lot of sparks when we unplugged cables, so tradeoffs I suppose…

The untruth is that these effects will produce anything significant in the audio band. This is a simulation of the frequency response of a 1st order lowpass formed between a 75 ohm output resistance (equivalent to a Schiit Modi, and a very standard line output Z) and the capacitance to ground from the Gotham’s shielding for a 3 meter run.

image
image2544×1179 151 KB

The bandwidth - defined by the -3dB point - is 20 megahertz.
Even with a 100m cable’s equivalent capacitance to the shield (and yes, since it’s RCA, we should really be considering the conductor and shield capacitance in parallel to ground, but that’s a bit aside to the shielding), we have a dip of .04dB at 20khz, the top of the audible band
image
image2546×1180 155 KB

It’s only with the equivalent conductor to shield capacitance of 1KM of the Gotham that we see our bandwidth drop to 20khz (and being 3dB down at 20khz, odds are that the older among us would still barely hear it)
image
image2545×1186 155 KB

By (absurd) comparison, the Mogami 2549 would only be a bit under 2dB down at 20khz for a 1km run
image
image2551×1180 156 KB

So the general message I would take away here is “don’t connect your Modi to something over 100m away” - but you’ll have noise issues (or even meaningful signal level drop from the simple resistance of the conductors) before cable capacitance is an issue here.

5 Likes
52

Wow. Thanks guys for the thoughts and info, definitely learned something today.

Big question here, maybe a rookie one:
I know it’s a bad idea to plug a single-ended connection, say a headphone, into a balanced/differential output on an amp because in single-ended/unbalanced, the grounds are tied and it will short out the differential amp.
However, what about going from a balanced/differential source, such as a DAC, into a single-ended input, such as on an amp? For example, going from a balanced xlr out to an unbalanced rca in. Will the same thing happen, or is it completely different? I’ve worked in pro audio, and there, most people say that it is totally fine (as in nothing will blow up), the whole connection just becomes unbalanced.
And what about going from a differential/balanced headphone output into an unbalanced 3.5mm or rca input?

Thanks guys!

1 Like
54

So there’s a few things here.

For a two wire differential connection - like the one you get for each driver with a 4-pin XLR connector on a headphone amplifier - there is no ground connection. The + output is a non-inverted version of your signal, and the - output is an inverted copy. The difference in voltage between the two terminals of the driver causes a current to flow through the coil, and boom, you get sound. If you swap one of the two output phases with a connection to ground, you cut the output in half - because +1 - -1 = 2, but +1 - 0 = 1 - but otherwise everything proceeds as normal.

With the 4-pin XLR or TRRS connections, however, there is no ground - when you short the negative pins, you’re summing the inverted copies of your right and left channels, and then the difference in voltage you get across your coils (which drives your transducers, don’tcha know) is relative to that sum, rather than zero. Not what you’d call optimal signal fidelity, in my view.

Now, there’s another, more pressing reason to not short them: headphone outputs are low impedance, meaning that a lot of current flows when there’s even a relatively small voltage difference. This is why the pro audio folks aren’t too concerned about shorting one of the connections on a TRS balanced signal line (the third contact is ground, incidentally) - the output impedance there should be in the high two to low three digits. An RME ADI-2 Pro can output +19dBu, which is 6.9Vrms, on its single-ended line outputs, which have a rated output impedance of 100Ω. Because I = V/R, this means that even directly shorting those to ground, your worst-case output current is 69mA, and since power = voltage*current, that’s about half a watt - which I’d hazard that the ADI-2 Pro can sustain indefinitely, if the line outputs are just series resistances with its high current output stage.

By comparison, the headphone output on the ADI-2 Pro has a .1Ω output impedance - if you shorted that to ground directly, with 6.9V, you’d need to push 69 ampere to ground, for almost 500W of power dissipation. That even the ADI-2 Pro can’t handle (nor can a lot of power amps…), so we can see why we’d rather not short a low impedance source to ground. In practice amplifiers using IC opamps for their output stages should go into overcurrent protection rather than simply exploding - and a well-designed discrete amplifier should also have overcurrent protections - but uh…it’s not going to be a happy time for the device, and your best case is that it does nothing while burning up a ton of power needlessly.

The Quedelix’s output impedance is less than 1Ω, so you really don’t want to short the pins to ground - it will probably not hurt the device, but it going into overcurrent protection seems likely, and the better solution is to just use the single-ended output if you need a single-ended connection.

7 Likes
55

Wow! I think that is one of the best, most in-depth answers I’ve ever gotten and a fantastic explanation. Thanks so much!

1 Like
56

I keep hearing things about how EQ, especially software EQ does one or more of the following: Blunts transients, worsens transients, or distorts transients. This made me realize that I don’t really know what people mean when they talk about transients. I had an idea that I assumed was correct but now I don’t know. When I use EQ, I can’t really tell that what I thought of as transients being changed at all.

What’re transients and does EQ have a meaningful impact on them?

2 Likes
57

Found this generalized definition on sweetwater’s site that might help…

A non-repeating waveform, usually of much higher level than the surrounding sounds or average level. Good examples of transients include the attack of many percussion instruments, the “pluck” or attack part of a guitar note, consonants in human speech (i.e. “T”), and so on. Due to their higher-than-average level and fleeting nature, transients are difficult to record and reproduce, eating up precious headroom, and often resulting in overload distortion. Careful use of compression can help tame transients and raise average level, although over-compression will result in a dull, squashed, flat sound to the signal.

I generally think of it as the initial strike or more appropriately, the attack (or maybe the ‘maximum threshold’ is another way of looking at it).

I personally don’t use EQ so maybe someone with more experience in this area can chime in. But if I had to guess, it will if you use it incorrectly.

58

I agree with what @hifiDJ posted: I liken transients to the initial attack of the note, whether its a guitar pluck, a cymbal strike or a drum hit; I sometimes see transients described as “the leading edge.”

Determining whether EQ worsens transients is probably widely debated, but here’s my take: even if it does negatively impact transients, the improvements in sound achieved by fixing the frequency response using EQ outweigh the potential negative impact on transients, if there are actually any negative impacts at all.

Download something like Equalizer APO and Peace, then experiment to see if you can hear a difference in transients when toggling EQ on and off. Turn your main sound down about 10-15 dB before doing so - that way you won’t introduce clipping because you weren’t adjusting the “Pre Amplifying” option in Peace. Then it’s just a toggle of the switch to listen for the differences yourself.

1 Like
59

I’ve been using Peace for a couple years now. It is very hard to test though because I have no way to volume match the frequency I am testing between EQ on and off.

60

Adjust said frequency’s volume using the Pre Amplifying gain in the upper left corner of Peace. If you boost a large range by say 3dB, perhaps lower the Pre Amplifying gain some to compensate for the volume change. If you mess with it enough, you’ll figure it out.

1 Like
61

I would like to ask about the materials science, theory, and actual studies done on cryogenic processing of audio components. I say components because I was going to say cables, and then I saw this from a purveyor of cryogenic treatments of many different materials.

I was idly wondering about what I have read regarding transmission of current or signals in wire, and the effects of various dialectric materials. The treatises I read always talk about the cryogenic treatment of cables and talk about the copper or other metal.

What effect can extreme cold have on the insulating material? Could that effect be as great as the effect on a copper wire?

I still have trouble getting my head around the idea that making something temporarily cold would have a lasting effect, then I think about the process of quenching steel. But that may be entirely different.

Take me to school. (but not beyond Bachelor’s degree level, I’m not ready for a Master’s)

1 Like