Head-Fi.org › Forums › Equipment Forums › Sound Science › I hear no difference. What am I doing wrong?
New Posts  All Forums:Forum Nav:

I hear no difference. What am I doing wrong? - Page 4

post #46 of 67
Quote:
Originally Posted by SilverEars View Post

Another comment regarding output impedance.  Tyll calculated the output impedance of AK240 being 3.24ohm, but this was done using a tone.  AK240 is spec'd to be 2ohms, and likely the AK240 also has non-linear output impedance response.  I would think all sources have non-linear output impedance response to a degree.  Possibly Calyx off has bass roll-off with different loads because of the same reason.  We all know specs don't tell you the whole picture.


One of the graphs you posted shows that your statement regarding all sources having non-linear output impedance is incorrect, look at the graph for the X3 again.
It probable that the frequency response of anything hooked up to the devices you've shown graphs for and of sufficiently low enough impedance, will have falling bass response anyway. So the effect will make zero audible difference, hence PDVJAM's statement that he hears no difference between DX50 and DX90 in the bass is probably correct.
post #47 of 67
Quote:
Originally Posted by SilverEars View Post
 

Below shows that output impedance(was spec to be <.3ohms, never trust specs!!!) is not linear with DX50 and there will most likely be bass roll-off with iems.  Anybody know what the audible affects of subbass roll-off?

 

 

 

The problem of quoting some guy from the internet is that he might not actually be correct in the first place. In this case, I am pretty sure the DX50 graph above is wrong on the output impedance. My own measurement puts it at 0.6ohm with a 1kHz test tone, which is more inline with iBasso's own number. Definitely not around 2.5~3ohm as the above graph suggested. That makes me wonder how accurate the rest of the graphs are.

 

 

Quote:
Originally Posted by SilverEars View Post
 

Another comment regarding output impedance.  Tyll calculated the output impedance of AK240 being 3.24ohm, but this was done using a tone.  AK240 is spec'd to be 2ohms, and likely the AK240 also has non-linear output impedance response.  I would think all sources have non-linear output impedance response to a degree.  Possibly Calyx off has bass roll-off with different loads because of the same reason.  We all know specs don't tell you the whole picture.

 

 

Not really. The graph above just tells you that Calyx M Coffer has coupling caps on the output... and you don't need coupling caps on the output if your amp section doesn't have DC offset to worry about, which is the case for FiiO's DAP. It all depends on the implementation.


Edited by ClieOS - 8/22/14 at 6:26pm
post #48 of 67
Quote:
Originally Posted by PDVJAM View Post
 

Two days ago I did listening tests once more. Checked AP100, DX50, DX90 and Fiio X5. Used more uncompressed songs with big dynamic diapason (checked by http://dr.loudness-war.info/) . Listened them for about 3 hours and decided to buy DX90 (AP100 and X5 have no rockbox): in my opinion, it does not have much difference with the younger brother, but he sounds a bit more gently. I didn't do ABX, but tried to compare them on same volume (DX90 with a but louder, so volume 200 on DX50 same as ~190 on DX90) as fast as I could (by switching SD card). Well, I'm happy with it:) Maybe I'm wrong and this "gentlyness" because of price difference.

 

Cool! I'm planning on getting the DX90 too...unless I go for the Calyx M instead.

post #49 of 67
Quote:
Originally Posted by Roly1650 View Post


One of the graphs you posted shows that your statement regarding all sources having non-linear output impedance is incorrect, look at the graph for the X3 again.
It probable that the frequency response of anything hooked up to the devices you've shown graphs for and of sufficiently low enough impedance, will have falling bass response anyway. So the effect will make zero audible difference, hence PDVJAM's statement that he hears no difference between DX50 and DX90 in the bass is probably correct.

The X3 shows flat FR, and the dips and peaks you see of the impedance graph is not linear to a degree(never said completely non-linear), but it's low enough that doesn't really matter.  Maybe you're mixing up the FR with the impedance response barely visible at the bottom.  If that measurement for the DX50 was correct(Clie has said he measured it to be not the case) and given that every headphone have different impedance characteristic, how a headphone react to output impedance is dependents on how relatively close the headphone impedance is to the output impedance.  For example some iems dip as low as 10ohms in the bass region, and spec'd as 16ohms at 1k.  If it was a much higher impedance phone, the difference will not be audible than if the iems were at the load.  Anyway all this is mute since Clie says the measurements are wrong.

 

I have not seen a output impedance sweep until this one, are there others?

Quote:
Originally Posted by ClieOS View Post
 

 

The problem of quoting some guy from the internet is that he might not actually be correct in the first place. In this case, I am pretty sure the DX50 graph above is wrong on the output impedance. My own measurement puts it at 0.6ohm with a 1kHz test tone, which is more inline with iBasso's own number. Definitely not around 2.5~3ohm as the above graph suggested. That makes me wonder how accurate the rest of the graphs are.

 

 

 

Not really. The graph above just tells you that Calyx M has coupling caps on the output... and you don't need coupling caps on the output if your amp section doesn't have DC offset to worry about, which is the case for FiiO's DAP. It all depends on the implementation.

Yeah, when I saw the THD imbalance, I didn't have much faith in the measurements.  Really?? Coupling caps at the headphone out?  It's a usb DAC/Amp.  Why would they add coupling caps to a headphone out?  If there is indeed a cap at the output as you say, cap is still an output impedance which is affecting the FR with load.  Given you say there is the cap at the output, it's probably creating a high pass filter with the resistor.

 

 

 

I'm curious if there are significant capacitance output by sources?  Any graphs of them? :tongue_smile: I know DACs have coupling caps and the input impedance to the amp is typically very high so that shouldn't roll off to the amp, correct?  Or is there a possibily for the roll-off to take place between the DAC and the amp?


Edited by SilverEars - 8/22/14 at 5:01pm
post #50 of 67
Quote:
Originally Posted by SilverEars View Post
 

Yeah, when I saw the THD imbalance, I didn't have much faith in the measurements.  Really?? Coupling caps at the headphone out?  It's a usb DAC/Amp.  Why would they add coupling caps to a headphone out?  If there is indeed a cap at the output as you say, cap is still an output impedance which is affecting the FR with load.  Given you say there is the cap at the output, it's probably creating a high pass filter with the resistor.

 

 

 

I'm curious if there are significant capacitance output by sources?  Any graphs of them? :tongue_smile: I know DACs have coupling caps and the input impedance to the amp is typically very high so that shouldn't roll off to the amp, correct?  Or is there a possibily for the roll-off to take place between the DAC and the amp?

 

Just output caps. There is no need to add resistor as the load itself will be enough to create a high pass filter.

 

To avoid DC offset on the output, you will need to use a direct coupled circuit, which usually requires bi-polar power section for the analog stage. As USB only comes in +5V + Ground, you need an extra circuit to create a -5V rail to have bi-polar power, that means extra circuit, extra cost, and extra care to avoid adding more noise to the mix (or they can just go with a chip amp that has internal voltage pump to avoid the need of bi-polar power). Sometime caps are just added as a extra safety precaution; sometime they are added purely for 'flavour'. In this case, I think it is mostly for flavour. since the ES9023 used in Coffee doesn't have DC offset issue at all (and ESS is promoting it as such - "eliminating the need for output dc-blocking capacitors"), and there is no amp section since Coffee's output comes directly from the DAC. (see my next post).

 

I found some picture of the Coffee and you can see the big but somewhat pointless caps: http://diy-audio-blog.blogspot.com/2012/02/calyx-coffee-inside-story.html

 

To be fair to Calyx however, most amp's input impedance will be in the thousands Ohm range so the roll-off will be well under 20Hz and insignificant to the measurement. The 16 to 250ohm load used in the measurement are not a representation of the real world usage. Again, this is another case where graph doesn't tell you the whole story.


Edited by ClieOS - 8/22/14 at 8:47pm
post #51 of 67

All these technical experts...if I ever develop audio products, I will probably consult with you guys!

post #52 of 67
Quote:
Originally Posted by ClieOS View Post
 

 

Just output caps. There is no need to add resistor as the load itself will be enough to create a high pass filter.

 

To avoid DC offset on the output, you will need to use a direct coupled circuit, which usually requires bi-polar power section for the analog stage. As USB only comes in +5V + Ground, you need an extra circuit to create a -5V rail to have bi-polar power, that means extra circuit, extra cost, and extra care to avoid adding more noise to the mix (or they can just go with a chip amp that has internal voltage pump to avoid the need of bi-polar power). Sometime caps are just added as a extra safety precaution; sometime they are added purely for 'flavour'. In this case, I think it is mostly for flavour. since the ES9023 used in Coffee doesn't have DC offset issue at all (and ESS is promoting it as such - "eliminating the need for output dc-blocking capacitors"), and there is no amp section since Coffee's output comes directly from the DAC.

 

I found some picture of the Coffee and you can see the big but somewhat pointless caps: http://diy-audio-blog.blogspot.com/2012/02/calyx-coffee-inside-story.html

 

To be fair to Calyx however, most amp's input impedance will be in the thousands Ohm range so the roll-off will be well under 20Hz and insignificant to the measurement. The 16 to 250ohm load used in the measurement are not a representation of the real world usage. Again, this is another case where graph doesn't tell you the whole story.

It doesn't tell you the whole story, but it tells you more than other measurements I've seen where output is tapped without load.  It tells you there is something wrong at the output.  Yes, the resistors are not representative of real world usage if you exclude Orthos and dynamic iems.  Much higher impedance dynamic diver cans such as Sennheisers or BA drivers will have a difference response than a resistor and the caps will add to that.


Edited by SilverEars - 8/22/14 at 7:49pm
post #53 of 67
Quote:
Originally Posted by SilverEars View Post
 

It doesn't tell you the whole story, but it tells you more than other measurements I've seen where output is tapped without load.  It tells you there is something wrong at the output.  Yes, the resistors are not representative of real world usage if you exclude Orthos and dynamic iems.  Much higher impedance dynamic diver cans such as Sennheisers or BA drivers will have a difference response than a resistor and the caps will add to that.

 

I think I just stumble on something else that can shed some light to Calyx Coffer. The previous reason to think there might not be an analog section is because there isn't any noticeable opamp or chip that will suggest there is something between the DAC output and the jack. As I inspected the pictures from the blog I linked, I notice there is a ST Microelectronics' DIP8 chip that is marked '4821 - 9219'. I searched over ST's database and didn't find anything under the marking, then I look into ST's product line and found that it is actually a TS482 - the assumed "1" is actually the letter 'I" (also explains why no one is able to identify it previously). TS482 is indeed a headphone driver with single power supply rail and thus needs coupling caps for blocking DC offset. Now that'll explain why those big caps are there.


Edited by ClieOS - 8/22/14 at 8:49pm
post #54 of 67
Regarding the discussion about caps at the output for DC block.  I recall GeekWave was offered a special package.  It's called:
 

Specail IEM Package
IEM's, or In-Ear Monitors, are a special kind of headphones that most people would recognize as those thingies that recording artists have in their ears when they're on stage.

IEM's are special because they're very sensitive and, unlike most headphones, they don't require the same amount of power to make them work well. What they do need, though, is very low background noise. If you choose this upgrade, we'll customize your Geek Wave to have a very low noise floor. We'll also add a high-precision DC offset amplifier and tune down the power to 100 mW (which will be plenty). Finally, we'll replace 8 key output bypass capacitors and will dial the bias current down.

This means, for iem use, the caps should not be there. But, that is also saying that for non-iem usage, the caps are there.  I wonder how significant it would be for low impedance portable headphones with the cap there?  I guess because they assume low impedance for iems, and high impedance for non-iems(which is not always the case)  LCD-X is 20ohms only, and it probably wouldn't like the caps there. 

 

What are these DC offset amplifiers?  Aren't they typically used?  Fiio and ibasso output high power and also drive iems, so I assume they use the DC offset amplifiers because they are run by battery.  Are there DAPs that use dual rails?

post #55 of 67
Quote:
Originally Posted by SilverEars View Post
 

 

What are these DC offset amplifiers?  Aren't they typically used?  Fiio and ibasso output high power and also drive iems, so I assume they use the DC offset amplifiers because they are run by battery.  Are there DAPs that use dual rails?

 

I am not an EE but 'DC offset amplifier' actually doesn't make much sense to me. I would think they might be referring to high precision "low" DC offset amplifier, which is just a common way for your chip maker to refer to their better-than-average chip. All chips have some DC offset. But in general, better the chip, lesser the DC offset. There are also different implementation to avoid or eliminate DC offset as well, not just relying the DC offset parameter of the chip.

 

All FiiO DAP are running on dual rail in the amp section. HM901 / HM801 as well. I am sure there are more out there.

post #56 of 67
Quote:
Originally Posted by SilverEars View Post
 

Also, it would be enlightening to look at the frequency spectrum of the uncompressed music to see the output response and how well they interact with the headphone type. My guess is that certain headphone type can work well with certain genres because of the spectrum they may output.  There are also other gear related factors that goes into it.  Like how the headphones responds to amp's output from the DAC.  There are lots of claims of some gears causing changes to the sound.

Ok, so I've looked into this in Audacity and looked into the spectrographs, and these show why the sound is harsher for pop music(top two) and audiophile recordings(bottom two).  Notice the red in the treble area is more prominant for the pop music vs the audiophile recording.  The redder the color, the waveforms have higher peak.

 

 

 

 

 

FR of HD800 (notice the treble peak region).  Now, why would people say the 800 sounds harsh?  :o Also explains genre listening preference of this type of headphone, and why only certain genres are preferred.

 


Edited by SilverEars - 8/22/14 at 10:45pm
post #57 of 67

Music that is slammed into the wall like that would sound harsh with just about any headphones. I've never understood why people who like music that is poorly mastered bother spending a lot of money on audiophile sound equipment. 

post #58 of 67
Quote:
Originally Posted by bigshot View Post
 

Music that is slammed into the wall like that would sound harsh with just about any headphones. I've never understood why people who like music that is poorly mastered bother spending a lot of money on audiophile sound equipment. 

Because we still like the music and still want it to sound as good as it can, short of completely remastering it.

 

Thanks to my listening habits, I have a harder time understanding why someone could only listen to one or two genres, even on the best gear. Or worse, only listen to it for an hour a day!

 

But these are questions for another thread.

post #59 of 67

The best equipment in the world won't improve the sound of massive clipping. Once the music has been hit with a sledgehammer of overdriving and compression, it is what it is. It'll never be any better.

 

I agree with you on the limited musical frame of reference most people have. I ask people what kind of music they like, and they answer "All kinds." So I ask them what their favorite opera or country artist or string quartet is... blank stares. It usually turns out that "all kinds" really means "all kinds" of just one kind of music... rock.


Edited by bigshot - 8/23/14 at 5:59pm
post #60 of 67
Quote:
Originally Posted by bigshot View Post
 

The best equipment in the world won't improve the sound of massive clipping. Once the music has been hit with a sledgehammer of overdriving and compression, it is what it is. It'll never be any better.

 

I agree with you on the limited musical frame of reference most people have. I ask people what kind of music they like, and they answer "All kinds." So I ask them what their favorite opera or country artist is... blank stares. It usually turns out that "all kinds" really means "all kinds" of just one kind of music... rock.

There's always more to the music than the clipping. Good gear won't cure clipping, but it still makes the rest sound better.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › I hear no difference. What am I doing wrong?