Take Me to School (Understanding Audio Better)

Oh yes, I think I must have blocked that out! @DarthPool inspired me to read up about laserdisc, and I learned that it’s actually an analog format on an optical medium! I never even considered that possibility.


Input. Take me to school.

Back in the day, consumer headphone amplifiers were sold to improve the sound coming from the portable devices of the day. Walkman. Mobile Phone. Discman. I recall a lot of DIY and homebrew amps being fit into Altoid tins. Most were powered by 2 or 3 AAA batteries. This is the era of the Airhead, and Electric Avenues, and a myriad of other little amps trying to be both small and better than stock sound. 1980-1995 era.

So, what are we supposed to do today with direct output from our phones? Many here point out that today’s better headphone amps expect to be driven from a “line out” of another device, and that “line out” ought to be at 2.0 volts. (I do have a DISCMAN around that has a Line Out, but I’ve never put a meter on it).

When I get my Lyr3, what can I expect connecting it to my iPhone audio out vs running the iPhone through a DAC with a 2.0 line out? (I also have a line out on my PC audio card, and the Mac Mini has a headphone out, but I don’t see a way to configure it as LINE OUT)

Take me to school.

Not reaching the 2V “standard” output from a phone, or DAC/amp output (like the DragonFly Black for example, which caps at 1.2V), means that you’ll be able to utilize less of the available power in the amplifier you’re feeding.

The same would be true if your “line out” didn’t reach 2V (some don’t).

Next, headphone outputs are intended to drive relatively low impedances (typically between 16 and 600 Ω, where as line-outputs are usually designed to expect much higher impedances (typically 10 kΩ or more). This is less of a problem than if it were the other way around, but isn’t ideal.

Most importantly, typically, is that “headphone outputs” are driven by an amplifier. ALL amplifiers generate noise and distortion. This is usually level-dependent. So as the level goes up, which it will to try and hit that 2V goal, so do the distortion figures as well as the noise floor. Which means when you feed that into another amplifier, it amplifies that incoming noise and distortion AND adds its own.

Generally line-outputs are buffers, rather than amplifiers, and have MUCH lower noise and distortion performance than an amplifier operating at half, or more, of its rated output voltage.

Again, using the DragonFly Black as an example. To get as close as possible to the 2V spec, it is running flat out, at the limit of its voltage rails, which is where it performs the most poorly (no amp running against its rails can be expected to be as clean as when it’s just “ticking over”). And in doing that, it’s hitting just 1.2V.

This means that a) you’re getting the highest noise and distortion from the DragonFly Black that it’ll ever generate and then b) you’re feeding it into another amplifier, still only giving it half what it really wants to see so that c) it can amplify that more-distorted-than-necessary-and-noisier-than-it-should-be signal to appropriate levels while d) still being unable to reach its rated output (which is typically specified for a nominal 2V input signal).

It won’t break anything doing this … but it will not sound as good as it could and compared to most other differences in an audio chain (e.g. say comparing an SDAC to A Modi 3), these issues can be relatively gross and easily audible - often unmissable.

TL;DR; Don’t feed amplifiers outputs into other amplifiers.


Thank you. The part I was missing was the amplifier vs buffer. I had erroneously assumed that the 2V line out was also amplified.

I would assume that this logic does not apply exactly the same way to pre-amplifiers that are specifically designed to drive amplifiers, or phono amplifiers designed to raise out put of cartridges (ignoring the RIAA curve for the purpose of this discussion, only talking about the gain).

One more point I’d like to ask - What is the difference between an amplifier being driven at the 2V line out and with the volume way down to one driven with a lower signal and the volume at perhaps 50-60%?

I find that I rarely have the need to really turn up the volume controls to near maximum levels, unless something in the chain is clearly mismatched.

1 Like

There are differences, yes.

For one a pre-amplifier may not actually amplify anything. Ignoring their input/output switching, they can be as simple as a passive attenuator. More commonly they’re an attenuator and a buffer, intended to feed into a high-impedance load (10 kΩ or more, usually).

Where a pre-amp does provide gain (amplification), it is usually a small amount, at VERY low power, and again into a much higher impedance load, and as such tends to be subject to much less noise and far lower levels of distortion than a comparable amp that has to drive headphones or speakers.

Phono stages have massive gain compared to pre, power and headphone amplifiers. This unavoidably adds noise and some level of distortion. But, again, they’re not delivering much in the way of power - the signals are still tiny, with little current requirement, and they’re feeding high-impedance loads, so they tend to be less of a problem.

Relatively speaking, and depending on the units in question, they may add more noise, by %, than a high-feedback, op-amp based, headphone amplifier (into a static load), but you have to have this gain stage - so that’s largely immaterial - and usually surface noise dwarfs the noise from the phono stage anyway.

But make no mistake, turn up the volume on a high-gain phono-stage, with no record playing, and you’ll hear more noise than you will with other sources.

Excepting non-linearities in the volume control within the amplifier, and assuming the same power output into the load, then if it is competently designed then the noise and distortion output of the amplifier being driven will be the same.

Typical headphone amplifiers are implemented as (greatly simplified):

Input -> Attenuation -> Gain Stage

Gain stages are generally fixed-level. They may be switchable to different fixed levels, but aren’t otherwise variable (typically). Output level is fundamentally controlled by how much signal is fed to this gain stage, which is what the attenuator does … lets you vary (reduce) signal level reaching the gain stage from the input.

Whether you feed this amplifier 1V at zero attenuation (i.e. it’s volume at max) or 2V at 50% attenuation, the gain stage will be delivering the same power (voltage and current), and will have the same noise and distortion performance in both cases.


Again … the problem with using amplifier outputs to drive other amplifiers, is the noise and distortion from the driving amplifier being re-amplified, with more distortion in the driven amplifier.

This is exacerbated if the driving amplifier is working at, or near, it’s limits, which is generally where it performs at its worst.

It can be preferable, still, if the output from your source is only available as a headphone output, and it doesn’t have enough power to drive the headphones in question properly. But it won’t be as good as a proper line-out.


For portables especially, many share the same op-amps etc for both line out and headphone out to conserve space and power. In those cases, wouldn’t the line out be exactly the same as the volume matched headphone out? I think the statement that Line out is better, assumes distinct circuitry from that driving the headphone output, but I’ll be the first to admit I still have much to learn.



I have the 2014 Mac Mini. Just for kicks, I plugged in the STAX headphone amp to the Mac Mini Headphone Out.

Looked through the Sound and the Midi control panels. It appears there is no way to configure this as a line out. The Midi panel defaults to 2 channel 24 bit integer 44.1khz, but allows selection of 16 bit integer at 44.1-96.0 khz and 24 or 32 bit floats at the same range of kHz.

I’m not at all sure that that represents in the context of the midi control panel as I have no midi instruments. (once I really wanted a Yamaha WX-7 or WX-9). But it doesn’t look like I can do line out.

There’s a reason I say “proper” line-out.

If it’s just the headphone output, via an amplifier, configured to run at 2V instead of being variable, then it gets you nothing special.

Better, dedicated, players, have a simple switch that takes the output off the buffers, bypassing the amplifier stage (which is usually a monolithic, anyway), for “Line Out” mode. (This only requires two extra components, which are sonically neutral and TINY).

But the point remains … “proper” line-out will sound better feeding an amplifier than a “headphone” output.

1 Like

The Mac Mini doesn’t have a true or “proper” line-out.

What you set the output bit-depth and bit-rate to is up to you.

However, if it doesn’t match the source material being played, then Core Sound will re-sample it to whatever you’ve set it to in the Audio-Midi Utility settings.

Proper music-playing applications, like Roon or Audirvana, will reset those settings on the fly to match the actual music file being played - thereby avoiding the OS’s built-in resampling - and output bit-perfect data to the DAC.

Note that choosing 24 or 32 bit output for 16-bit material just pads the sample data with zeros and is benign.


I should add … not all DACs use amplifiers to get variable output.

Some simply take the direct output off the conversion stage, add a current buffer, and output that both to their RCA and XLR outputs, as well as their headphone outputs.

Chord do this, so do Soekris, among others, and the result is basically the same as a true, but variable, line-out. This approach does, however, require that the output stage from the digital to analog conversion has enough voltage swing to provide the necessary output levels - which is why not ALL DACs can do it this way.

1 Like

Oops, another question. Have Tidal, and use it on iOs and Android (mostly iOs). Also have Roon Remote on my iPads.

I think I’ve been under the impression that for MQA from iPad, I need to go direct from Tidal. Not sure if that’s also the case with a general high-resolution file.

And that Roon Remote, if I use it on the iPad, trying to play MQA or High Res file can only get 44.1??? (Using xDSD and/or Dragonfly and camera adapter…)

Both Roon on iOS, and the iOS TIDAL app, should be able to output MQA and, in the case of Roon, hi-res files, at their native bit-rates, up to the limit of the external DAC.

For the DragonFly that’s 24/96, for the iFi xDSD even higher.

If that’s not working, check the Audio Settings and make sure that the Max Sample rate (etc.) is set correctly for each device.

For direct output from the built-in headphone outputs or the Apple 3.5mm dongles, the limit is 24/48.

Amen to that. I would much rather listen to a well recorded 320 kpbs mp3 than a poorly recorded (not terrible) high resolution or lossless WAV, FLAC, SACD, etc. regardless of the quality of the sound system I am using. This is why I was so sceptical, disappointed and even angry with Neil Young’s so called battle with the music industry when he launched Pono Music and declared, “My goal is to try and rescue the art form that I’ve been practising for the past 50 years.”

In some cases the 24/192 KHz master he offered might have been a different and better master than what was originally released to the consumer but for the most part, I don’t think it was.

1 Like