[Argyris]

Quote:

Quote:

... Hence the second half of my post, if you read it. Quoting me on a partial post out of context proves nothing.

I read your entire post, including the second part.  Your premise misses the concept that being recessed or pronounced in the frequency response is relative.  Look at how HeadRoom presents their frequency response graphs.  All of the graphs are adjusted so that 1000 Hz is at the 0 dB level.  All of the graphs are relative to having 1000 Hz at 0 dB.  There is a reason why they do that.
 
Adjust all of your comparisons so that the mids are always "5".  Instead of "high 5, mid 3, low 5" you would have "high 7, mid 5, low 7".  In a relative context they are exactly the same.

Ham's got the right idea. I didn't end up putting this in my post, so I'm glad it got brought up. The HR graphs are indeed calibrated to the 1khz @ 0dB standard. This is so comparisons are as apples-to-apples as possible. You could shift the entire graph up or down the chart and the relative difference in amplitude between frequency zones would be preserved. If a midrange of "5" is considered neutral, I would plot the DT880 as follows:
 
low 4, mid 5, high 7
 
My original point, which I'll illustrate again in this format, was that "mid 5" isn't between two peaks--it's adjacent to a higher treble figure, yes, but it's also adjacent on the other side to an equal or lower bass figure. I've read quite often that the DT880 has recessed midrange which, assuming my assessment of the balance is accepted (and it may not be), is impossible.
 
To address Ham's point, the following--
 
low 2, mid 3, high 5
 
--would be an identical assessment, as the difference was applied to each value equally.

 

[Riku540]

Right, but rating 4/5/7 opposed to 2/3/5, though proportionately equal, can suggest two entirely different things.
 
4/5/7 suggests that the bass is a little on the light end, with average/neutral mids and a bit of extra treble.
 
2/3/5 suggests that the bass is almost non-existent, with severely lacking mids, and average treble.
 
If a characteristic of a headphone's sound signature is average, there is no logical reason to tip the scale in either low or high extremes.
 
That would mean 8/8/8 is neutral, which it is, but then V-shaped headphones suddenly become 12/8/12? Yes it's the same but it's just ridiculous to rate that way.
 
This can and only will make headphone comparisons more confusing for potential buyers.

 

[Argyris]

This is the very reason HeadRoom chooses to standardize on 1kHz. This frequency is a common test tone and is usually used to determine sensitivity ratings, and it also usually produces graphs that look proportional to the chart where they're plotted. There are very few headphones (the W5000 is one of them) that don't quite fit with this calibration because very few headphones have a huge peak at 1kHz. I see your point here, except that a rating of 4/5/7 is proportionately identical to 2/3/5. There is no objective standard that will change that fact.
 
The reason 4/5/7 looks better to us is that, at least in the western world, we usually think in base ten. In this instance, we would automatically assume that the scale is from 0-10 because this is the most common way that data is presented to us. The number "5" is right in the middle of that plot, therefore it is equidistant from the two extremes. If we shifted the entire plot down by two, our lower extreme would be -2 and our upper extreme would be 8. We don't like this arrangement because it creates an illusion of less leeway below the midpoint than above it (even if we cognitively understand that our lower boundary is below 0). Look at the W5000 chart:
 

 
That weird bump in the midrange offsets the plot--it's a feature unique to this response. (NOTE: I'm not discussing the validity or lack of validity of the curve, just its appearance on the chart). Using our system, I would classify this one as 3/5/5. This is a rare exception where using "5" as the midpoint creates an odd result of anemic-looking bass. If we moved the response up by 5dB it would give a better conceptual picture of this curve: boosted midrange and treble. Let's call it 5/7/7. There is no difference between these two assessments, as they are both ways of describing the relative relationship between the three frequency zones. However, the imagined latter example would "look better", as the flat portion of the graph would hug the zero line, which it does on most of the other headphone charts. Interestingly, 3/5/5 "looks better" in number form, but the associated chart "looks better" as the equivalent 5/7/7.
 
You are arguing for an objective standard where "5" is always neutral, and in a qualitative sense at that. In a system of relative relationships, that objectivity is purely aesthetic, as shown by the discrepency between the numeral and graphic depictions of the W5000 plot. On the other hand, I agree with you 100% that shifting the results up or down by excessive margins would create unusual and confusing results. I'm not arguing for which is "better", merely that both assessments are technically equivalent. What matters in a comparison of different frequency zones in a headphone is how they relate to each other, not objectively which values are used.
 
EDIT: I emphasize for those reading that this is all academic, as I've seldom seen anybody boil a sonic signature down by numbers. It might be useful to do so, though--a little more objective than word descriptions, for one thing.

 

[joelpearce]

Well, that semantic argument probably really clarified everything for our confused newcomer.

 

[Amatsu]

Is it possible to fix the recessed mids with some equalization without messing up the sound too much? Just wondering.

 

[joelpearce]

Quote:

Is it possible to fix the recessed mids with some equalization without messing up the sound too much? Just wondering.

EQ seems highly dependent on the particular pair of headphones and the quality of the equalizer.  Personally, I've tried a bunch of equalizer setups, and I've never found one that I've kept and used consistently.  When you find the right pair of headphones, they don't need it.

 

[Ham Sandwich]

Quote:

Is it possible to fix the recessed mids with some equalization without messing up the sound too much? Just wondering.

I EQ my Denon D2000 to fix their recessed midrange.  A picture of the EQ curve I use is linked in my sig.  I like the results and I don't find the EQ to mess up the sound too much.  For me the EQ solution gives far more pros than cons in terms of fixing the sound.
 
But you need to use a parametric EQ for things like this.  A 5 or 6 band or even 10 band graphic EQ isn't going to work well, and would likely do more harm than good.  A parametric EQ allows you to select the frequencies you want to adjust and gives you much more control.  Fortunately if you do computer as source it is easy to add a good software based parametric EQ. 

 

[ikamusume]

Someone said to just turn up the volume to get from 2 2 2 to 5 5 5, but to my ears that never sounds good. When I crank the volume to get inefficient speakers to finally make some sounds, it just takes all the joy out of the music and it's like I'm trying to listen through a pillow. Sure I can turn up the volume until I can hear it through the pillow, but it's just not the same as I would have gotten if I could have listened at a lower volume without the pillow, instead it's flat and lifeless. It's like a photo that was taken with too small a camera lens and then photoshopped to be bright enough, the result is not dynamic.