Do You Need an External DAC/Headphone Amplifier?

[amirm]

@amirm
How is the jitter on the laptop? I am concerned that there will be a schiit modi-like situation with noise shifting with load, since it shares the board with other components.

also, thd+n vs frequency for the dac, to make sure 1khz isn't an outlier.

Here is jitter:

It is exceptionally clean other than a slight rise at very low frequencies.

BTW, my laptop that is being measured is also driving the Audio Precision analyzer which sucks up a full core by itself and is playing music over Topping DX7 as i type this.

 

[amirm]

Here is frequency response which likewise is excellent:

 

[bennetng]

I had a Dell Inspiron which had an "Creative Audigy 2" enhancement thingy. Something tells me it was a realtek chip though.

Some Realtek chips like ALC1150/1220 probably can outperform the original Creative Audigy 2 in terms of DAC performance. But those Realtek chips just don't have the DSP power of those Creative/EMU processors.

Things like Live!/Audigy/X-Fi and so on are just Creative's brand names, they have no correlation with DAC performances.

 

[Blumlein 88]

I posted several excerpts of music a couple years back from various sources. The digital original was presented along with 2 good DACs, a Lenovo T410 business laptop, and a smartphone. Asked people to rank them. The ranking result from the few who responded were just about equally split. FWIW I also had a listening test of digital original vs smartphone asking which is the smartphone. That particular phone was lower fidelity than the circa 2010 Lenovo. The results did not indicate people could hear it reliably.

I would note how actual listening tests, let your ears decide type tests, seem to be like kryptonite to subjectivist golden ears. Hundreds will download the files and occasionally more than one dozen, but less than two dozen will respond. Subjectivist sighted tests seem to involve influences from the eyes and from stories told about superior or inferior design. So few respond I'm loathe to post such files anymore.

FWIW the Lenovo above had SNR in the mid 80's decent, but not stellar distortion results. J-test signal had a widened base, but not much else. And unlike many older desktop sound cards it wasn't effected by the load on the machine. You didn't hear any mouse movements or clicks, no HD noise or other spurious noise leaked thru the outputs. I've had some laptops that did, but modern ones don't seem to be particularly afflicted by that problem.

EDIT after reading Jimster480's post below: I would note the listening tests had the headphone output feeding a high impedance ADC input. The lack of power and moderately high output impedance will mean using it with a headphone directly will show some deficiencies and frequency response changes that are audible compared to a quality headphone amp.

 

[RayDunzl]

but less than two dozen will respond

Sorry.

I usually check them (everybody's listening examples) out but am too deaf to make anything of them.

Sometimes it is fun to ignore the "Don't Look!" admonitions to see what's there.

 

[amirm]

Here is THD+N versus frequency. Graph in red is 22 kHz bandwidth (same as dashboard). For this graph using external DACs I usually use 90 kHz bandwidth so that the harmonics of 20 kHz are properly captured. The IDT DAC uses noise shaping and has good bit of high-frequency noise which upsets my measurements. So I selected 45 kHz to get reasonable results (in blue):

 

[Jimster480]

To avoid circular arguments, I'll let this one go . I'll agree that I know only what I have perceived. - While apparently the measurements appear to know even less of what I perceive. (Sorry, had to slip that one in there! )


I like that test. I don't think I've done that one. But I have performed (non blind) tests to try to identify between mp3 and wav. - Above 160 kbps at least
The answer it's very hard to distinguish between the two. I don't think I could reliably differentiate in a test like that especially because I couldn't differentiate in a sighted A/B comparison

The reality is that the biggest difference you hear, is likely in the Amp itself.
As Amir shows here, the power output is insanely anemic and most headphones will sound flat with onboard audio.
So when you use an Amp with any external DAC, it will likely sound great in comparison.

For me on my Laptop, I use my FiiO K1 along with whatever headphones I bring with me.
This provides me with a mostly great experience on the road, as I only have lower impedance planar headphones, or BA IEM's.
But the difference vs my Acer Nitro 5 or my XPS 13 (that I sold yesterday) is very distinguishable, and I would bet that most of the difference comes in the output impedance + power offered.

I can tell you that in some objective listening tests, its hard to tell the FiiO K1 apart from my DX7 + JDS O2 stack. It really depends on the track, and the headphones.
Testing vs the DX7 alone really does depend on the headphones since some headphones will have sliding frequency response that I can easily distinguish since I own this gear and listen to it every day

But I can tell you that if you piped some onboard audio via "line out" into the back of your Amp, and tested against your Gugnir Multibit.... with real objective listening (and not this bias you post on the forum).... its unlikely you will hear a difference where the Gugnir will outperform the onboard (unless you really have an older and very terrible onboard chip).

 

[amirm]

And finally, intermodulation distortion (again at 45 kHz bandwidth):

 

[amirm]

Linearity of course is pretty awful given the low output level:

 

[garbulky]

The reality is that the biggest difference you hear, is likely in the Amp itself.
As Amir shows here, the power output is insanely anemic and most headphones will sound flat with onboard audio.
So when you use an Amp with any external DAC, it will likely sound great in comparison.

I remember running headphones without an amp (an Senn HD600). I don't miss those days!

But I can tell you that if you piped some onboard audio via "line out" into the back of your Amp, and tested against your Gugnir Multibit.... with real objective listening (and not this bias you post on the forum).... its unlikely you will hear a difference where the Gugnir will outperform the onboard (unless you really have an older and very terrible onboard chip).

To be up front, I have not ever owned any Schiit products.
I tell people this. If your experience shows you little difference between electronics (that aren't trash to begin with), then GOOD NEWS! Stop right now. You have just saved yourself a bunch of money. You'll be just as happy and your wallet that much fatter.

 

[Timbo2]

So to answer the question posed in the title, “no, but you’d probably be well served with a straight amplifier for headphone use.”

Thanks for the analysis!

 

[garbulky]

Linearity of course is pretty awful given the low output level:

View attachment 14865

Is there a particular voltage one has to reach to consider it 0db on a test?

 

[amirm]

Is there a particular voltage one has to reach to consider it 0db on a test?

No, this test is relative.

 

[amirm]

I tell people this. If your experience shows you little difference between electronics (that aren't trash to begin with), then GOOD NEWS! Stop right now. You have just saved yourself a bunch of money. You'll be just as happy and your wallet that much fatter.

Well, even BETTER NEWS would be if you believed in audio science just the same and didn't waste your money on things that don't change the sound. Instead you could put that money to good causes like music, food, vacation, charity, etc.

 

[Xyrium]

Speaking of bad computer sound, I am currently using the headphone jack in front of my Intel NUC to feed my Topping PA3 amplifier and Paradigm Atom small monitors - and the sound is pretty bad. As soon as my new SMSL DAC arrives, I expect a big improvement in sound quality.

In my low-budget retirement setting, I actually now own four low-power chip amps. Three are decent quality - Trends, SMSL and Topping - and one is a cheapo junker - Lepai, which I designate as "For Emergency Use Only!!!"

Man, don't those NUC's at least have a line out? Ugh...The rest of the chain seems respectable though!

 

[garbulky]

Well, even BETTER NEWS would be if you believed in audio science just the same and didn't waste your money on things that don't change the sound. Instead you could put that money to good causes like music, food, vacation, charity, etc.

I'm always confused when people think I don't believe in science.

I argue that I waste very little money. Well okay....that last part is not actually true...according to my wife.

 

[garbulky]

Does anybody remember listening to sound using the computers built in speaker?

 

[bennetng]

I posted several excerpts of music a couple years back from various sources. The digital original was presented along with 2 good DACs, a Lenovo T410 business laptop, and a smartphone. Asked people to rank them. The ranking result from the few who responded were just about equally split. FWIW I also had a listening test of digital original vs smartphone asking which is the smartphone. That particular phone was lower fidelity than the circa 2010 Lenovo. The results did not indicate people could hear it reliably.

Not completely on topic but people who are interested can give it a try. The "90dB" thing refers to dynamic range rather than SINAD.
https://hydrogenaud.io/index.php/topic,116189.0.html
There are some test files in Reply #15

Since I already revealed the answers I am not expecting members here to post something like ABX test logs and so on.

 

[Music1969]

Hi @amirm

How does the Topping do into an 8 ohm load?

The reason I ask is because my MrSpeakers AEON closed are ~14 ohm headphones.

In future, are you able to test all the way down to 8 ohms, since many head-fiers use low impedance planars and lower impedance IEMS.

The Chord Mojo is rated at 8 ohms.

Many thanks

https://www.innerfidelity.com/images/MrSpeakersAeonBook.pdf

 

[amirm]

Hi @amirm

How does the Topping do into an 8 ohm load?

The reason I ask is because my MrSpeakers AEON closed are ~14 ohm headphones.

In future, are you able to test all the way down to 8 ohms, since many head-fiers use low impedance planars and lower impedance IEMS.

It is time consuming to test at many impedances. I hope to automate the process but have not had the time to build the fixture. Trade off between building tools and doing reviews without!