[Kaffeemann]

I'm not the most intelligent person but not that stupid 

Of course I adjusted the volume.
30 Ohm changes the damping factor from 63 to 0.63.

 

[castleofargh]

  My 30 Ohm resistor finally arrived and Iistened to my iBasso DX90 (0.3 Ohm) and this IEM with and without the resistor.
The bass does sound different with the resistor. Softer, more blurred, it's harder to hear single bass notes.

I would love to see how the amp part behavior changes here.  crosstalk, FR, THD, everything.
 
anyway with my lowskill lazy approximations, if I take a 1ohm variation(not sure there is that much on the graph) between 20hz and 1khz I get a whopping 0.5db difference.

 

[Kaffeemann]

Yes, I don't think the FR changes noticeably. It would be great if someone could perform a RMAA test of this sort. Unfortunately my notebook only has a combined audio jack.
My guess is that distortion and time domain characteristics in the bass are affected.
Edit: And probably crosstalk.

 

[SilverEars]

Well, looking through Rin's graphs, I noticed he adds impedance to BA based iems, but haven't seen dynamic driver based iems being tested with added impedance.  Not saying Rin is the authority on what happens to iems, but he does try to explain what he is doing and what is going on with the results.

  Yes, I don't think the FR changes noticeably. It would be great if someone could perform a RMAA test of this sort. Unfortunately my notebook only has a combined audio jack.
My guess is that distortion and time domain characteristics in the bass are affected.
Edit: And probably crosstalk.

Maybe if you look around there is something out there that show a headphone with flat impedance graph being affected somehow with significant enough impedance value.  

 

[stv014]

  I'm not the most intelligent person but not that stupid 

Of course I adjusted the volume.

 
How did you adjust it exactly ? It is not very easy (but it is possible) to perform a fast switching level matched blind test between different output impedances, as both the level and the impedance need to be changed at the same time. If you just make the switch, then adjust the volume while the music is playing to compensate, that is not good enough.

 

[stv014]

  Yes, I don't think the FR changes noticeably. It would be great if someone could perform a RMAA test of this sort. Unfortunately my notebook only has a combined audio jack.
My guess is that distortion and time domain characteristics in the bass are affected.

 
Distortion could change, at least it measurably (but not necessarily audibly) does in the case of full size headphones. Time domain characteristics (CSD) are mostly just the frequency response (including the phase response) visualized in a different way. For example, a narrow resonance on the frequency response translates to long ringing in the time domain.

 

[stv014]

  I would love to see how the amp part behavior changes here.  crosstalk, FR, THD, everything.

 
In theory, the amp part (that sees the combination of resistor+driver as a higher impedance load) is likely to perform better, at least as long as the increased output level that is required to compensate for the power lost on the resistor does not result in clipping. Other than that, the higher and more resistive load impedance is easier to drive.

 

[stv014]

  We are talking about a 0.7 dB increase in bass response. That's never enough to explain the huge rise in distortion.

 
Indeed, as these graphs show, in a level matched test, where the low impedance frequency response is compensated to have the same amount of bass resonance, there is still increased distortion with a higher source impedance, and it appears on the headphone driver (and its acoustic output as well), rather than the amplifier output before the serial resistor.

 

[SilverEars]

   
Distortion could change, at least it measurably (but not necessarily audibly) does in the case of full size headphones. Time domain characteristics (CSD) are mostly just the frequency response (including the phase response) visualized in a different way. For example, a narrow resonance on the frequency response translates to long ringing in the time domain.

Why would the distortion change with added resistance?  NM, I saw the link you posted above.

 

[Kaffeemann]

   
Indeed, as these graphs show, in a level matched test, where the low impedance frequency response is compensated to have the same amount of bass resonance, there is still increased distortion with a higher source impedance, and it appears on the headphone driver (and its acoustic output as well), rather than the amplifier output before the serial resistor.

 
stv014, can you perform similar measurements with an IEM that has a flat impedance response?

  Why would the distortion change with added resistance?  NM, I saw the link you posted above.

 
Also, take a look at this.

 

[SilverEars]

   
stv014, can you perform similar measurements with an IEM that has a flat impedance response?
 
Also, take a look at this.

Yes, I would like to see this also.  

  stv states that he would like to see how orthos are affected, maybe that would give us some insights.
 
I looked at the bechmark test and the resonance area is affected significantly which is predictable.  What's interesting is the area that is nominal which is closer to being resistive(although if you look at the phase that area is still reactive).  Sony is nominal around 80ohms and the 650 is nominal around 300.  The 650 has less distortion in the nominal area.  Yours is 18 ohms.  Quite possibly, the distortion could show up more with the flat iems from what I'm seeing of the changes in distortion from 300ohms to 80ohms.  

 

[castleofargh]

 

  I would love to see how the amp part behavior changes here.  crosstalk, FR, THD, everything.

 
In theory, the amp part (that sees the combination of resistor+driver as a higher impedance load) is likely to perform better, at least as long as the increased output level that is required to compensate for the power lost on the resistor does not result in clipping. Other than that, the higher and more resistive load impedance is easier to drive.

yup that's what seems to be the most common situation. I'm a little obsessed with crosstalk because of a few DAP measurements I've seen. like this about the ak100 on goldenears:

 
so I wonder if at some point a well thought added resistor could not "help" a given amp enough to compensate what will be lost on the headphone side from that added impedance?
because we now have a bunch of ultra sensitive, ultra low impedance IEMs going as low as 8ohm. on those with fluctuating impedance response over frequency it's not interesting as obviously the FR change will matter more than the rest, but on something pretty flat like kaffeemann's or my 16ohm IE80, those measurements made me very curious indeed.
 
edit: or if we should simply avoid IEMs with such small impedances? ^_^ because I see people sold on some DAPs because they have 2 DAC chips to gain some meaningless 2 or 3db noise level and stuff, so it would seem a little weird to wilingly lose all those "improvements" with a 12ohm IEM because the amp section doesn't like it much.

 

[SilverEars]

   
Distortion could change, at least it measurably (but not necessarily audibly) does in the case of full size headphones. Time domain characteristics (CSD) are mostly just the frequency response (including the phase response) visualized in a different way. For example, a narrow resonance on the frequency response translates to long ringing in the time domain.

Yeah, you see it on CSD plots like this one.  Looks to be for narrower peak areas you can predict the ringing, but on the low end it's different.  Bigshot brought up whether the decay duration is audible or not, anybody have information on ringing and audibility?
 

 

[stv014]

  stv014, can you perform similar measurements with an IEM that has a flat impedance response?

 
I do not have any to test. Someone else with suitable equipment (I did not use anything expensive myself, but it did require some basic DIY tools for the microphone and the loopback measurements with the resistor that I switched with a relay) could try measuring it, however.

 

[Kaffeemann]

  Yeah, you see it on CSD plots like this one.  Looks to be for narrower peak areas you can predict the ringing, but on the low end it's different.  Bigshot brought up whether the decay duration is audible or not, anybody have information on ringing and audibility?
 

 
If you look at other IEM measurements you will see that narrow peaks don't automatically imply that there will be ringing.
IMO ringing in the treble is easily audible. If the treble of an IEM is harsh despite good EQ settings ringing is probably the reason.
Overly long decay of bass frequencies will make it harder to hear single bass notes.