Head-Fi.org › Forums › Equipment Forums › Sound Science › If a DAP is described as "Warm" or having a frequency bump - why doesn't it show on Freq response curves?
New Posts  All Forums:Forum Nav:

If a DAP is described as "Warm" or having a frequency bump - why doesn't it show on Freq response curves? - Page 2

post #16 of 30
Quote:
Originally Posted by bigshot View Post

I think people are hearing things.

 

It happens most of the time, but that doesn't mean it's absolute in all cases. When people talk about "micro detail" differences I can usually get the same thing out of the supposedly "inferior" source by bumping up the volume on the amp in/after it slightly. Then sometimes the differences, and what is actually the "inferior" one, can be surprising...worse, what many people consider "inferior" can be even more baffling. A much loved CDP that I was excited to try out beause people said the soundstage was larger sent the drums all the way around my head, and I wonder how many people have seen Reed Richards sub as a drummer in concerts, because if we're talking about a natural or at least a believable reproduction, well, that doesn't happen until somebody gets the same effect out of exposure to gamma rays.


Edited by ProtegeManiac - 5/19/14 at 3:11am
post #17 of 30
Quote:
Originally Posted by ProtegeManiac View Post

It happens most of the time, but that doesn't mean it's absolute in all cases. When people talk about "micro detail" differences I can usually get the same thing out of the supposedly "inferior" source by bumping up the volume on the amp in/after it slightly. Then sometimes the differences, and what is actually the "inferior" one, can be surprising...worse, what many people consider "inferior" can be even more baffling. A much loved CDP that I was excited to try out beause people said the soundstage was larger sent the drums all the way around my head, and I wonder how many people have seen Reed Richards sub as a drummer in concerts, because if we're talking about a natural or at least a believable reproduction, well, that doesn't happen until somebody gets the same effect out of exposure to gamma rays.

Headfi in a nutshell
Good and Bad Change > Nothing
Edited by Hapster - 5/19/14 at 4:45am
post #18 of 30
Quote:
Originally Posted by Hapster View Post


Headfi in a nutshell
Good and Bad Change > Nothing

 

No not just Head-Fi, and certainly not just because it has a younger crowd than speaker forums - even the older demographic has the same thing going on. Many people tend to think stronger bass from upstream components always means "better," but usually ignore that it's pushing the bass instruments too far forward (that even depth-less recordings take on depth in the opposite direction and with just the bass drum). At least when it gets flabby there are more people who actually notice that happening.

post #19 of 30

if you look at those test response curves you see a very clear change in the frequency response when he connects a 32 ohm headphone as opposed to 200 kohm.  and you see it clearly shows a broad hump around 80hz and flatter highs.  this hump is the warmth you are hearing.

post #20 of 30
Quote:
Originally Posted by groovyd View Post
 

if you look at those test response curves you see a very clear change in the frequency response when he connects a 32 ohm headphone as opposed to 200 kohm.  and you see it clearly shows a broad hump around 80hz and flatter highs.  this hump is the warmth you are hearing.

Look at the scale. Driving the 200kOhm load, it's flat within 0.05dB from 10Hz to 10kHz, and it's less than a tenth of a dB down between 10kHz and 20kHz. For all intents and purposes, that's dead flat. You cannot hear a 0.05dB rolloff at 10kHz. As for the 32 ohm plot, it's elevated by about 0.15dB from 40-150hz, and it's down somewhere between 0.05 and 0.1dB above 6kHz. Again, this is completely inaudible (and it isn't flatter in the highs - it's actually slightly more attenuated than the 200kOhm plot, but the scales are different).

 

(Also, that very slight, and completely inaudible bass boost is due to the 4.5 ohm output impedance of the device combined with the Edition 8's impedance curve. That's not something that will always happen with 32 ohm devices - it'll depend on the impedance curve of the headphones, although I can't envision it being a problem unless you have headphones that drop below 20 ohms at some frequencies)


Edited by cjl - 5/19/14 at 1:16pm
post #21 of 30

we must be looking at different links... i see .2db hump and an upward curve in the highest frequencies.  I am not comparing absolute db in the highs but the shape of the response since ultimately you will compensate for this by turning the volume up.  in other words shift both curves to lie on 0db at their minimums and that is the response you will hear. This actually puts the mid bass hump up around .3db which you will definitely hear as warmth.  

 

your consideration of the result being the combination of source and load impedances is all the more true in this case.  We see the unit takes on a mid bass hump for one pairing in the 32 ohm range, it is possible it will take on an even larger hump with some other headphones.  These are curves for the amplifier output not the ear cup.  Maintaining a flat response is entirely the responsibility of the amplifier at this point regardless of the load impedance characteristics.  This is precisely what separates a good amplifier from a not so good one in maintaining a flat response regardless of the load.

post #22 of 30

Frequency response variations have to be a LOT bigger than .3 dB to make a difference. The threshold of perception for volume differences is between .5 and 1dB with tones. With music, it's higher. With music AND just a particular band of frequencies, even higher still. It depends on the frequency too. Above 10kHz, the imbalance can be even more without being noticeable. As a ballpark, I'd say 3-5dB is the threshold of perception.

 

You're talking about an imbalance that is ten times or more beneath the level of audibility.


Edited by bigshot - 5/19/14 at 2:49pm
post #23 of 30
Quote:
Originally Posted by bigshot View Post
 

Frequency response variations have to be a LOT bigger than .3 dB to make a difference. The threshold of perception for volume differences is between .5 and 1dB with tones. With music, it's higher. With music AND just a particular band of frequencies, even higher still. It depends on the frequency too. Above 10kHz, the imbalance can be even more without being noticeable. As a ballpark, I'd say 3-5dB is the threshold of perception.

 

You're talking about an imbalance that is ten times or more beneath the level of audibility.

I wouldn't say 3-5dB is the threshold of perception for something like a midbass boost. Playing around with a good parametric EQ for a bit shows that a 5dB boost in that frequency range is fairly clearly audible. If I had to guess at the real threshold, I'd say it's more like 1-2dB for most frequencies (although as you said, for extremely high frequencies, the threshold will be much higher). I do agree though that 0.2-0.3 dB is definitely too small to matter.

post #24 of 30

My general rule of thumb is that if you can get within 3-5dB of being flat coming out of the transducers, you are doing pretty doggone good. I'll tweak it a dB here and there, but differently recorded music provides enough variety of sound to make it not really much more critical than that. Obviously the most important range to get right is 1kHz to 4kHz, because our ears are particularly sensitive around there. If you don't have any sub bass, there's nothing wrong with a little bit of a mid bass hump.

 

Of course, perfectly flat is ideal. I'm talking "in practice" here. But there's no reason that solid state CD players, DACs, DAPs or amps should have variation more than +/- .5dB to 1dB nowadays.

post #25 of 30
Quote:
Originally Posted by bigshot View Post
 

My general rule of thumb is that if you can get within 3-5dB of being flat coming out of the transducers, you are doing pretty doggone good. I'll tweak it a dB here and there, but differently recorded music provides enough variety of sound to make it not really much more critical than that.

That would definitely be pretty darn good. Your amp should be capable of significantly better than that though, since nearly all of your variability will come from the transducers and (if you're using speakers) room. Any modern amp or player should be capable of a flat response from 20-20k with well under 0.5dB of variation (and realistically, it should be capable of better than 0.1dB without much trouble). Any device that doesn't achieve this is a device that I would definitely look at carefully before buying, not because I think that level of variance is audible, but because it is so easy to achieve that kind of flat response with modern technology that you almost have to be ignoring good design principles to get anything else (and if your device is flawed enough to have a >0.5dB variation in frequency response, it might have other flaws as well).


Edited by cjl - 5/19/14 at 6:48pm
post #26 of 30
Thread Starter 

Thanks for all the updates guys - It's been good to follow your reasoning.  So at this stage I'm still thinking that the "warmth" that the poster mentioned is probably more expectation bias than real. Either that or he has a hearing threshold an order of magnitude better than mine (not discounting this - but think it unlikely).

post #27 of 30
Quote:
Originally Posted by Brooko View Post
 

Thanks for all the updates guys - It's been good to follow your reasoning.  So at this stage I'm still thinking that the "warmth" that the poster mentioned is probably more expectation bias than real. Either that or he has a hearing threshold an order of magnitude better than mine (not discounting this - but think it unlikely).

 

Perhaps he has loudness or eq turned on the iPhone without even knowing it?  That would more easily explain his reaction.

post #28 of 30

That is a lot more common than any of us think.

post #29 of 30
Thread Starter 

Could be.  I guess what it's pointing out to me though is that (as measured) the iP5 (like most of the iPhone series from 4 up) is essentially flat from 20-20K, and actually performs pretty well. 

post #30 of 30

Also worth noting, sometimes the obverse is what causes a perception.  A good example is early lower priced Etymotics were accused of having weak bass.  They had a new revision.  It sounded like it had better bass than the previous version mostly ameliorating the issue.  How did Etymotic achieve this?  Simple, they rolled off the uppers 3 octaves or so gently.  Instant warmth with no change in low end response. 

 

I have a bit more experience with doing room EQ for loudspeakers for a few people I know.  I ask how they wish it were different and make changes.  Often the changes are not where they would expect it to achieve their desires.  But usually more often than not it is successful.  Perceived FR is a case of balance depending on what is available to work with.  A bit of study of the psychoacoustics aids one in figuring out the basics. 

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › If a DAP is described as "Warm" or having a frequency bump - why doesn't it show on Freq response curves?