Question about MP3 Players with Roll Off
Nov 29, 2012 at 5:54 AM Thread Starter Post #1 of 35

H20Fidelity

Headphoneus Supremus
Joined
Feb 11, 2012
Posts
16,289
Likes
6,779
Can someone please tell me why some portable MP3 players have roll off? For example bass roll off / treble roll off. Surely company's like Cowon and Colorfly do test their players previous to launching, I'm sure with the technology they use can create a flat sounding / measuring player. Why do these company's make DAP's with roll off if they know us audiophiles don't like it?

I mean...surely then they make the roll off on purpose? For what reasons?

It's something I've been wondering lately. I must ask the community to clear my thoughts.
 
Nov 29, 2012 at 6:22 AM Post #2 of 35
I have wondered about this too......
popcorn.gif

 
Nov 29, 2012 at 6:23 AM Post #3 of 35
These days most recordings are over bright, and have unreal sub bass
 
this roll off just lets the music shine through.
 
blame the loudness wars
 
Nov 29, 2012 at 7:36 AM Post #4 of 35
Treble roll off often has to do with the low pass filter used after the DAC stage. All DAC stage has a low pass filter to cut off higher frequency (>22kHz) because those are where the noise is. How much and how fast the roll-off is depends on what kind of filter and how it is implemented. The most common filter that has a more obvious / wider roll off is the Butterworth filter, like the kind find on HM801, but almost known for being very smooth sounding. ODAC also has a Butterworth first order IIRC, thus it is very mildly roll-off after 18kHz. Another type of filter that often find on DAP is the Chebyshev filter type 1, know to have rippers on the treble region, followed by a very shape roll-off. You can often find this on SigmaTel chipset based DAP
 
Bass roll-off on the other hand often has to do with coupling caps for blocking harmful DC voltage. Capacitor followed by a resistor (the headphone) will form a high pass filter. When both the capacitance and resistance aren't large enough, the corner frequency (the point of roll-off) will become very high, and thus the bass will get cut off / roll off.
 
These are just some of the common reasons that cause roll-off. There might be others that I don't know.
 
Nov 29, 2012 at 9:43 AM Post #5 of 35
Hi ClieOS,

Thanks for explaining what causes the roll off.

But why are such audiophile focused players (Hifiman Cowon, Colorfly) designed to have roll off when other players like the Clip+ measure flat?

You'd think being mid / high end focused audiophile player would purposely be built to measure flat or have flat frequency response. but instead they seem to have roll off there for a reason?
 
Nov 29, 2012 at 10:07 AM Post #7 of 35
Hi ClieOS,
But why are such audiophile focused players (Hifiman Cowon, Colorfly) designed to have roll off when other players like the Clip+ measure flat?


Because they're careless and / or incompetent. Just look at the utterly schiitty firmwares they put on some of those "audiophile" DAPs. Even on mainstream / mass-produced DAPs which are a little more polished, gapless playback (which is a really basic feature) is missing most of the time, because whomever was in charge of designing and developing those DAPs didn't care, and / or the software engineers were too incompetent.

Schiitty firmwares, non-linear frequency response, high output impedance and unusually high distortion figures should all raise a red flag. Good, affordable DAPs exist. There's no excuse.

Edit: also, they probably want to get their products on the market as fast as possible, and they probably cut a lot of corners.
 
Nov 29, 2012 at 10:12 AM Post #8 of 35
Quote:
Because they're careless and / or incompetent. Just look at the utterly schiitty firmwares they put on some of those "audiophile" DAPs. Even on mainstream / mass-produced DAPs which are a little more polished, gapless playback (which is a really basic feature) is missing most of the time, because whomever was in charge of designing and developing those DAPs didn't care, and / or the software engineers were too incompetent.
Schiitty firmwares, non-linear frequency response, high output impedance and unusually high distortion figures should all raise a red flag. Good, affordable DAPs exist. There's no excuse.

 
The quote in my signature seems relevant here.....
wink_face.gif

 
Nov 29, 2012 at 11:48 AM Post #9 of 35
Quote:
Hi ClieOS,
Thanks for explaining what causes the roll off.
But why are such audiophile focused players (Hifiman Cowon, Colorfly) designed to have roll off when other players like the Clip+ measure flat?
You'd think being mid / high end focused audiophile player would purposely be built to measure flat or have flat frequency response. but instead they seem to have roll off there for a reason?

 
The answer is multi-dimensional.
 
First, almost none of the audiophile DAP maker has access to either a software or hardware team that is capable of making anything remotely half decent of a SoC solution. Beside Apple (which is the only one with enough money to ask others to do the engineering for them), everyone else is buying their solution from off-the-shelf solution provided by another company. Sandisk has bag of money, but not nearly as much as Apple, so they get their solution (at least on Fuze and Clip) from AMS. Try and go to AMS to order a few hundreds to a few thousands of their SoC unit and they will laugh at you. We are talking about ten of thousands of unit here (if not 100x more), and that's the level the big boys are playing. Companies like HifiMan or ColorFly, which only expect to sell a few hundreds unit, to a thousand if they are extremely lucky (and I do mean EXTREMELY), will never get entertained by an decent SoC provider. Instead, they need to rely on the small SoC provider that ain't known for great engineering nor great software support - but they are the only people around that will take your 'tiny' order, so what can you do? You just have to suck it up and deal with them. These SoC provider not just here to give you a chip, they are the solution provider - that means you tell them what kind of hardware you want to add on top of their existing SoC design, and they modify it just enough to make it work, then try to modify the stock firmware so it will play nice with everything, or at least not crash in the first few seconds. At 99.95% of the time, the audiophile brand themselves has little to no control over firmware, as software engineer is way to expensive to hire by a small firm, and the SoC provider will NEVER release their source code to you. Remember, most of these SoC provider is only doing your business because large company usually don't want to deal with them, yet your hundreds unit order isn't important enough for them to dedicate too much resource for you - so here you are, stuck in a loop that is audiophile DAP manufacturing with all the difficulty in the world. If you know how much investment you need to put into an audiophile DAP to turn it into real product, you will likely not going to do it in the first place. That is one of the reason why the few audiophile DAP that do come to live want to charge you for big money - because that's the only way they will even make any money out of a few hundreds unit of DAP they are going to sell. It is the power of mass production that makes thing cheap and good, Apple is the best example. For HifiMan or ColorFly, they are like the medieval of an electronic manufacturing, so their cost always skyrocketed and the finishing will never be the same level to Apple or Sandisk.
 
With the first handicap in mind, now comes the part of trying to piece together a design that works - I am not too familiar with others so I'll stick to HM801 in this case - first, the choice of Butterworth filter is intentional. It is how Fang like it and it is a well regarded design in the past (you can find the same filter in WOO's DAC, plus quite a few others). Even top ESS chip has a Butterworth filter built in, they just also have a few others option for you choose which to use. Fostex HP-P1 has a slow roll-off filter too, yet you rarely see people complain it being not HiFi enough - the true is, it is all about turning to the desired sound, it isn't like Butterworth filter has no benifit either - it has no group delay or rippe when compared to the sharp Chebyshev Type 1, but how many times do you see anyone complaining about not having those? It is really a silly mentality among audiophile that choose to only look for a certain fact while totally ignore the others - but as human, that's what many of us do. On the case of Clip - which I have no idea what filter it used (could very well be a high order Butterworth) but it has the benifit of big manufacturing and miniturization, so even very complex filter design (like a fifth order Butterworth) that won't fit on HM801 PCB can be shrinked down to a tiny part of the SoC. Again, we go back to the first point where audiophile DAP maker is limited by what that is availble to them.
 
You might think audiophile DAP maker should have all that they need to build the perfect DAP - yet the real world is quite the opposite. It is the big companies that has every resource available, yet they are also the least interested in making a Hi-end DAP with the best sound parts and best measured spec, becuase those are not something they can sell by truck load. Yet, the few companies that do want to make audiophile DAP is lacking all the resource to compete. As audiophile, we are stuck between them.
 
Nov 29, 2012 at 12:03 PM Post #10 of 35
It is the big companies that has every resource available, yet they are also the least interested in making a Hi-end DAP with the best sound parts and best measured spec, becuase those are not something they can sell by truck load.


And yet, iPods and Clips measure better than most audiophile DAPs. From what you've said, "audiophile" companies can't compete either on price nor on quality, so how do their products make any sense at all?
 
Nov 29, 2012 at 12:05 PM Post #11 of 35
Quote:
And yet, iPods and Clips measure better than most audiophile DAPs. From what you've said, "audiophile" companies can't compete either on price nor on quality, so how do their products make any sense at all?

 
I have iPod, Clip+ and HM801, and yet the two that are better measured isn't as good sounding as the last one - you tell me.
 
Nov 29, 2012 at 12:18 PM Post #12 of 35
I have iPod, Clip+ and HM801, and yet the two that are better measured isn't as good sounding as the last one - you tell me.


Well, unless you can come up with a better explanation, I'd say either you like the non-linear frequency response / coloring (which I would totally understand, since I use an EQ myself), or you're the victim of expectation bias (which I am too, except I know it!).

In any case, I'm fairly certain that no magic is involved! If they're too incompetent and / or don't have the means to make a device that measures well (or even just decently), how could they be so much more competent than Apple's or Sandisk's army of engineers, or have the means to make a device that sounds significantly better? It doesn't make any sense.
 
Nov 29, 2012 at 12:23 PM Post #13 of 35
Yep, I see that argument coming with whole bias blah blah blah - not interested in arguing with you on that. So be happy.
 
Nov 29, 2012 at 5:59 PM Post #14 of 35
You might think audiophile DAP maker should have all that they need to build the perfect DAP - yet the real world is quite the opposite. It is the big companies that has every resource available, yet they are also the least interested in making a Hi-end DAP with the best sound parts and best measured spec, becuase those are not something they can sell by truck load. Yet, the few companies that do want to make audiophile DAP is lacking all the resource to compete. As audiophile, we are stuck between them.


Wow ClieOS! you really explained that well. I understand completely now, Thanks a lot for your in-depth detailed answer.

Thanks heaps dude!

:beerchug:
 
Nov 29, 2012 at 6:48 PM Post #15 of 35
Quote:
Well, unless you can come up with a better explanation, I'd say either you like the non-linear frequency response / coloring (which I would totally understand, since I use an EQ myself), or you're the victim of expectation bias (which I am too, except I know it!).
In any case, I'm fairly certain that no magic is involved! If they're too incompetent and / or don't have the means to make a device that measures well (or even just decently), how could they be so much more competent than Apple's or Sandisk's army of engineers, or have the means to make a device that sounds significantly better? It doesn't make any sense.

 
Different people have different preferences. If the world worked the way that most commenters on the internet wanted it to, it would be very boring indeed. The general idea is: My set of preferences is right. How could you possibly not agree with me on what products to buy?
 
to Clie's point, Tegra 2 is a great example, it only cost $20/unit, but the minimum order was 100,000 units. 
 
In regards to the firmware, though, I just disagree. Why Hifiman, or Colorfly, or SOMEONE wouldn't just pick up Rockbox, fergodssake. A dude in his basement can get Android up and running on a decent piece of hardware, look how easy it was to implement Rockbox by some guy reverse engineering everything on to the 60X and 801. That is the one thing I don't get. You don't need some super computer engineer to do it. Hire some software programmer to do it in their spare time. Done and done. 
 

Users who are viewing this thread

Back
Top