Is it possible to work out the FR of a DAP?
Oct 31, 2016 at 3:59 PM Thread Starter Post #1 of 13

CraftyClown

Headphoneus Supremus
Joined
Feb 19, 2013
Posts
1,622
Likes
201
Ok, so I regularly see comments and reviews that talk about observable colouration within DAPs, for example someone may comment that a DAP is slightly north of neutral in SQ. Now this makes sense to me as I imagine the amp section could be causing this colouration. What I don't quite understand is how one would measurably come to this conclusion? What would your reference of flat or neutral be in the first place?

Is this just more audiophoolery? Or do I need schooling?
 
Oct 31, 2016 at 5:30 PM Post #2 of 13
It depends mostly on the output impedance of the DAP, and is most relevant with IEM's like yours, which usually react very sensitive to a (too) high output impedance.
Your headphones(at least from your sig) are pretty unproblematic with this, and i doubt there is any DAP out there which could alter their frequency response only by 0.1dB.
 
look here: http://www.head-fi.org/t/800208/the-rmaa-rightmark-audio-analyzer-source-and-audio-device-measurement-thread-measurements-in-post-3-tutorial-in-post-2/60#post_12962566
 
there are many examples of how a DAP might alter the frequency response of sensitive IEM's.
 
Oct 31, 2016 at 6:07 PM Post #3 of 13
  It depends mostly on the output impedance of the DAP, and is most relevant with IEM's like yours, which usually react very sensitive to a (too) high output impedance.
Your headphones(at least from your sig) are pretty unproblematic with this, and i doubt there is any DAP out there which could alter their frequency response only by 0.1dB.
 
look here: http://www.head-fi.org/t/800208/the-rmaa-rightmark-audio-analyzer-source-and-audio-device-measurement-thread-measurements-in-post-3-tutorial-in-post-2/60#post_12962566
 
there are many examples of how a DAP might alter the frequency response of sensitive IEM's.

 
Thanks for that link. Seems I have some playing to do :)
 
So in that case it seems it is quite possible for people to measure the FR of a DAP and detect whether there will be some colouration.
 
Consider me schooled :D
 
Oct 31, 2016 at 8:24 PM Post #4 of 13
it's actually fairly easy to measure the frequency response of a DAP. the problem is that many pros and reviewers tend to still use stupid standards more fitting of DAC measurements. most DAPs show frequency response "unloaded", so into an impedance so big the result is close to having nothing plugged. it's an utterly meaningless measurement for DAPs IMO(cowon's favorite way for years). those measurements are real, but they do not show what happens when you plug a 10ohm IEM into the DAP. plenty of things could happen(or not).
also the IEM itself can be affected by the impedance of the DAP and get a different FR from that interaction(mostly BA drivers or multidrivers). so yeah RMAA with the IEM you use as a load is a more realistic way to get the actual FR. you could also use a microphone and measure the sound getting out of the IEM, but that's actually harder to do well and calibrate correctly. repeatability can be a major problem and it's hard to share results with other people doing measurements. while what Chris does in the link above, should give pretty close results if I was to do the same measurements myself with my own cables and soundcard(again we're talking frequency response only).
 
Nov 1, 2016 at 2:53 PM Post #5 of 13
  it's actually fairly easy to measure the frequency response of a DAP. the problem is that many pros and reviewers tend to still use stupid standards more fitting of DAC measurements. most DAPs show frequency response "unloaded", so into an impedance so big the result is close to having nothing plugged. it's an utterly meaningless measurement for DAPs IMO(cowon's favorite way for years). those measurements are real, but they do not show what happens when you plug a 10ohm IEM into the DAP. plenty of things could happen(or not).
also the IEM itself can be affected by the impedance of the DAP and get a different FR from that interaction(mostly BA drivers or multidrivers). so yeah RMAA with the IEM you use as a load is a more realistic way to get the actual FR. you could also use a microphone and measure the sound getting out of the IEM, but that's actually harder to do well and calibrate correctly. repeatability can be a major problem and it's hard to share results with other people doing measurements. while what Chris does in the link above, should give pretty close results if I was to do the same measurements myself with my own cables and soundcard(again we're talking frequency response only).

 
I completely disagree that it is a meaningless measurement.
 
You can say that it is not an entirely useful measurement by itself to the average consumer, and I would agree with you... 
 
The point of measuring the DAP is to do just that, measure the DAP. We're looking for the performance of the DAP, not a system with some arbitrary headphones with the DAP in the loop. It's a slippery slope that would lead to demands from DAP manufacturers to provide measurements for an endless list of loads... and that just doesn't make sense for a DAP manufacturer. And remember, 10Ohms is not just 10Ohms... it's 10Ohms at some nominal frequency... the impedance curve will change depending on headphone design. Two headphones with a 10Ohm spec could have a different impedance curve that happens to pass through the same point.  
 
Nov 1, 2016 at 3:04 PM Post #6 of 13
   
I completely disagree that it is a meaningless measurement.
 
You can say that it is not an entirely useful measurement by itself to the average consumer, and I would agree with you... 
 
The point of measuring the DAP is to do just that, measure the DAP. We're looking for the performance of the DAP, not a system with some arbitrary headphones with the DAP in the loop. It's a slippery slope that would lead to demands from DAP manufacturers to provide measurements for an endless list of loads... and that just doesn't make sense for a DAP manufacturer. And remember, 10Ohms is not just 10Ohms... it's 10Ohms at some nominal frequency... the impedance curve will change depending on headphone design. Two headphones with a 10Ohm spec could have a different impedance curve that happens to pass through the same point.  


​I read Castleofargh's comment to mean the measurement is fairly useless on it's own. Taking the two measurements, with and without load as Chris does, provides far more useful information.
 
Nov 1, 2016 at 3:18 PM Post #7 of 13
 
​I read Castleofargh's comment to mean the measurement is fairly useless on it's own. Taking the two measurements, with and without load as Chris does, provides far more useful information.

 
Sure. It's more helpful in making a decision about whether or not a DAP is a good fit for the equipment you have.
 
It's equally helpful to have complete measurements for just the DAP into a high impedance load and separate complete measurements for the IEMs or headphones. This way one has all the information one needs to make a determination, and the DAP manufacturer is not burdened by having to perform tests with an endless list of headphones. 
 
Nov 2, 2016 at 9:08 AM Post #8 of 13
  I completely disagree that it is a meaningless measurement.
 
You can say that it is not an entirely useful measurement by itself to the average consumer, and I would agree with you... 
 
The point of measuring the DAP is to do just that, measure the DAP. We're looking for the performance of the DAP, not a system with some arbitrary headphones with the DAP in the loop. It's a slippery slope that would lead to demands from DAP manufacturers to provide measurements for an endless list of loads... and that just doesn't make sense for a DAP manufacturer. And remember, 10Ohms is not just 10Ohms... it's 10Ohms at some nominal frequency... the impedance curve will change depending on headphone design. Two headphones with a 10Ohm spec could have a different impedance curve that happens to pass through the same point.  

what I find meaningless is doing those unloaded. pretty much any load would make more sense for a DAP than 10kohm. my DAP is super flat when I play music without plugging anything in it!!! yeah!!! ^_^ that's something.
even if they're super lazy, they could take a standard value and do only it, 16ohm has been some kind of reference for years, why not that. or 30ohm if the marketing guys start pissing themselves because 15ohm would show how crappy the amp section really is. I don't care, anything would be more relevant than unloaded to me.
 
ideally, here is my carebear specs from manufacturers. first it would all be done by an independent group, like Audio Precision's lab or whatever, for consistency. 
then I'd like to have the voltage and power value reaching 1% THD into low impedance loads. and then measurements done into one of those loads (10 or 15ohm would be nice I guess), with voltage a little below into the safe zone to see how the DAP behaves without clipping, but still under a difficult load like it will most likely be when using my IEMs.
that would IMO have value, I could know what can and what can't work for nominal use. what load and how loud I can expect to go with a given pair of IEM before I buy the DAP. and if the DAP is flat when driving small loads.
impedance over frequency for the DAP would be cool too,  but I don't hold my breath. marketing leads things and marketing cares about specs only when they're shinny.
 
Nov 2, 2016 at 2:08 PM Post #9 of 13
   
I completely disagree that it is a meaningless measurement.
 
You can say that it is not an entirely useful measurement by itself to the average consumer, and I would agree with you... 
 
The point of measuring the DAP is to do just that, measure the DAP. We're looking for the performance of the DAP, not a system with some arbitrary headphones with the DAP in the loop. It's a slippery slope that would lead to demands from DAP manufacturers to provide measurements for an endless list of loads... and that just doesn't make sense for a DAP manufacturer. And remember, 10Ohms is not just 10Ohms... it's 10Ohms at some nominal frequency... the impedance curve will change depending on headphone design. Two headphones with a 10Ohm spec could have a different impedance curve that happens to pass through the same point.  

 
Actually 10ohm is 10ohm when you are doing measurement on DAP.
 
Almost all credible measurement on source are (and should be) done with 'ideal load', meaning the load is almost purely resistance, a la a pair of resistors.. A 'real world' load, i.e. headphone, will have impedance as well as induction, which will contribute to those problem you mentioned and thus they are not generally used. The idea is that if the source can't handle an ideal load, then it certainly can't handle a real world load -  so we can establish a pretty good, though not perfect baseline that will give repeatable result with every measurement by anyone with good enough tools.
 

 
On a side note and addition to @castleofargh suggestion, actually a 0.1% THD might also be more meaningful for DAP as I know quite a few of the big-brand-name offerings can't output enough to reach 1% THD. One of the thing I really like about the review done by the admin on erji.net (*the Chinese equal to HF) is that he always includes AP's measurement as part of his review (with ideal load as suggested by AP). Those data generally won't tell you how good a DAP might sound to the ears, but it will certainly point out those with design flaw / weakness easily.
 
Nov 2, 2016 at 2:24 PM Post #10 of 13
  what I find meaningless is doing those unloaded. pretty much any load would make more sense for a DAP than 10kohm. my DAP is super flat when I play music without plugging anything in it!!! yeah!!! ^_^ that's something.
even if they're super lazy, they could take a standard value and do only it, 16ohm has been some kind of reference for years, why not that. or 30ohm if the marketing guys start pissing themselves because 15ohm would show how crappy the amp section really is. I don't care, anything would be more relevant than unloaded to me.
 
ideally, here is my carebear specs from manufacturers. first it would all be done by an independent group, like Audio Precision's lab or whatever, for consistency. 
then I'd like to have the voltage and power value reaching 1% THD into low impedance loads. and then measurements done into one of those loads (10 or 15ohm would be nice I guess), with voltage a little below into the safe zone to see how the DAP behaves without clipping, but still under a difficult load like it will most likely be when using my IEMs.
that would IMO have value, I could know what can and what can't work for nominal use. what load and how loud I can expect to go with a given pair of IEM before I buy the DAP. and if the DAP is flat when driving small loads.
impedance over frequency for the DAP would be cool too,  but I don't hold my breath. marketing leads things and marketing cares about specs only when they're shinny.

 
 
   
Actually 10ohm is 10ohm when you are doing measurement on DAP.
 
Almost all credible measurement on source are (and should be) done with 'ideal load', meaning the load is almost purely resistance, a la a pair of resistors.. A 'real world' load, i.e. headphone, will have impedance as well as induction, which will contribute to those problem you mentioned and thus they are not generally used. The idea is that if the source can't handle an ideal load, then it certainly can't handle a real world load -  so we can establish a pretty good, though not perfect baseline that will give repeatable result with every measurement by anyone with good enough tools.
 

 
On a side note and addition to @castleofargh suggestion, actually a 0.1% THD might also be more meaningful for DAP as I know quite a few of the big-brand-name offerings can't output enough to reach 1% THD. One of the thing I really like about the review done by the admin on erji.net (*the Chinese equal to HF) is that he always includes AP's measurement as part of his review (with ideal load as suggested by AP). Those data generally won't tell you how good a DAP might sound to the ears, but it will certainly point out those with design flaw / weakness easily.

 
I'm talking about using IEMs or headphones as a load coupled to a mic or some other measurement system as I think is being suggested as a better measurement. The impedance curve of this load will vary by frequency... and that variation is not generalized across all IEMs or headphones, that is to say it will vary by design. As such there would be endless configurations for the test setup and that is simply not practical. All this to say, a 10Ohm load in the context of which it was being described is in fact not a 10Ohm load. 
 
You could certainly use a purely resistive 10Ohm load for your test... but what does it accomplish? This is simply fixed gain/attenuation relative to the output impedance of the DAP, which is stable across the band of interest. Okay, so shift your high impedance FR up or down accordingly... why does the manufacturer need to test with that load? Am I missing or overlooking something? 
 
Nov 2, 2016 at 2:35 PM Post #11 of 13
   
 
I'm talking about using IEMs or headphones as a load coupled to a mic or some other measurement system as I think is being suggested as a better measurement. The impedance curve of this load will vary by frequency... and that variation is not generalized across all IEMs or headphones, that is to say it will vary by design. As such there would be endless configurations for the test setup and that is simply not practical. All this to say, a 10Ohm load in the context of which it was being described is in fact not a 10Ohm load. 
 
You could certainly use a purely resistive 10Ohm load for your test... but what does it accomplish? This is simply fixed gain/attenuation relative to the output impedance of the DAP, which is stable across the band of interest. Okay, so shift your high impedance FR up or down accordingly... why does the manufacturer need to test with that load? Am I missing or overlooking something? 

 
Even with a headphone as load, you won't go through a mic. The actual setup is the measurement input (i.e sound card or audio analyzer) is in parallel with the load. Since the measurement input will have much much higher input impedance than then load, it will be 'invisible' to the source. Using a headphone as load will be more realistic and thus the result will be closer to what you will see in the real world. Of course, if we were to standardize to using headphone as the load, we will have to specify which headphone to use first. The difficult part is to ask anyone to agree to which headphone to use. I am sure many manufacturers will have their own interest to protect and thus not to agree to anything.
 
Actually I have seen very bad Sony and AK result (on THD and power) with a 16ohm ideal load, so it does accomplish something - it tells us just how bad those DAP's amp section are. If they can't even handle a simple load, what can we expect them to do well with an actual 16ohm headphone with the extra impedance and induction?
 
Nov 2, 2016 at 3:59 PM Post #12 of 13
@grump
RMAAs with an IEM as load don't have the same purpose, they initially were a simple method to assess if the impedance of the DAP was high or not. like the guys at ABI were doing some years back using the triple fi on purpose because it would make a mess of the FR if the source wasn't low impedance. the IEM had a huge impedance swing going as low at 8ohm. and that was exactly the reason why it was used. incidentally it also often would show if capacitors were at the output and creating a high pass with the IEM. so it felt like it answered several questions in a simple free test.
 
now to actually measure the DAP itself, instead of the impact on the FR with a specific IEM, I totally agree that a resistor is what should be used. BTW question to anybody who knows. I have plenty of cheap crap resistors(you know the "buy one, get 1000free" kind of crap), but I get in trouble with some when I run something several times for averaging. some resistors clearly change behavior when they get hot. is it because I really use the crappiest crap that exists, or are there resistors better than others for such tests? if they're expensive stuff, forget it I'm a cheap bstard when it comes to measurement rig. but if not, I could most certainly improve the accuracy of some of my pikatchu tests. with one or 2 proper loads.
 
what interests me with low impedance loads is that instead of looking at fullscale voltage and almost no current, we see how the amp behaves when it's actually difficult. higher current flow, worst crosstalk, things getting hot, more noise, can it even provide the current for the load? all of it becomes fun into low impedance. that's where we really find the limit that interests IEMs users because indeed IEMs nowadays often have very low impedance at least at some frequencies.
 
 
and about microphone, maybe it's me that confused everybody, I mentioned it to say it's not a good idea. it's obviously not the best idea for people used to doing those stuff, but it's usually the first idea people who know nothing about measurements will have. you want to know the FR of a system, measure the sound getting out of the entire system with microphone. intuitively it seems to be the right move. I've had a few people asking me about that in PM over the years. and I tried myself when I just started being too curious for my own good. in fact, in the beginning I didn't have much of anything and I devised a way to find the impedance of a DAP by measuring the variation between 200hz and 2khz on my jh13 because those were the only known impedances I had ^_^. I did this with a headset microphone stuck to the nozzle
biggrin.gif
. #shamescience
 
Nov 2, 2016 at 4:06 PM Post #13 of 13
  @grump
RMAAs with an IEM as load don't have the same purpose, they initially were a simple method to assess if the impedance of the DAP was high or not. like the guys at ABI were doing some years back using the triple fi on purpose because it would make a mess of the FR if the source wasn't low impedance. the IEM had a huge impedance swing going as low at 8ohm. and that was exactly the reason why it was used. incidentally it also often would show if capacitors were at the output and creating a high pass with the IEM. so it felt like it answered several questions in a simple free test.
 
now to actually measure the DAP itself, instead of the impact on the FR with a specific IEM, I totally agree that a resistor is what should be used. BTW question to anybody who knows. I have plenty of cheap crap resistors(you know the "buy one, get 1000free" kind of crap), but I get in trouble with some when I run something several times for averaging. some resistors clearly change behavior when they get hot. is it because I really use the crappiest crap that exists, or are there resistors better than others for such tests? if they're expensive stuff, forget it I'm a cheap bstard when it comes to measurement rig. but if not, I could most certainly improve the accuracy of some of my pikatchu tests. with one or 2 proper loads.
 
what interests me with low impedance loads is that instead of looking at fullscale voltage and almost no current, we see how the amp behaves when it's actually difficult. higher current flow, worst crosstalk, things getting hot, more noise, can it even provide the current for the load? all of it becomes fun into low impedance. that's where we really find the limit that interests IEMs users because indeed IEMs nowadays often have very low impedance at least at some frequencies.
 
 
and about microphone, maybe it's me that confused everybody, I mentioned it to say it's not a good idea. it's obviously not the best idea for people used to doing those stuff, but it's usually the first idea people who know nothing about measurements will have. you want to know the FR of a system, measure the sound getting out of the entire system with microphone. intuitively it seems to be the right move. I've had a few people asking me about that in PM over the years. and I tried myself when I just started being too curious for my own good. in fact, in the beginning I didn't have much of anything and I devised a way to find the impedance of a DAP by measuring the variation between 200hz and 2khz on my jh13 because those were the only known impedances I had ^_^. I did this with a headset microphone stuck to the nozzle
biggrin.gif
. #shamescience

 
Some RF designs use (synthetic) diamond filament resistors specifically because they are very stable over temperature. As you might expect, they are also very expensive... 
 
You might want to look at Vishay foil resistors... they may have better temperature performance at a more reasonable price. 
 

Users who are viewing this thread

Back
Top