Here are the RMAA results for the iPod Classic (2009) (http://outpost.fr/rmaa/iPod.htm), the Samsung Galaxy Nexus (http://outpost.fr/rmaa/GN.htm), and the FiiO E7 (http://outpost.fr/rmaa/E7.htm). The iPod was tested via it's Line Out Dock, as well as its headphone out, unloaded, and loaded with Sony MDR-XB700 headphones (24 ohms). The Galaxy Nexus and E7 were tested via their headphone out, loaded and unloaded. Here's (http://outpost.fr/rmaa/XB700.htm) a comparison between the 3, loaded with the headphones.
Audiophiles always say that the iPod's LOD outputs a "cleaner" signal than its headphone out; if I'm reading the values and graphs right, that seems correct. I couldn't say whether the difference is audible; one would have to also test the external headphone amplifier that would be connected to the LOD.
Overall the E7 seems superior, while the iPod is a pretty close second in some aspects, especially if you consider the LOD results. Of note is the significantly worse stereo crosstalk result from the Galaxy Nexus.
Edit: I used an old laptop for the recordings. I'm considering investing in a decent USB device. If I do get one, I'll run RMAA again.
You said in the other thread that you've got the volume cap on your ipod. If so, you need to either remove that so that you can use the full dynamic range of the ipod, or else set the level on the other players to be identical to the ipod. Otherwise you're basically handicapping the ipod by not letting the DAC use its full SNR.
The volume cap can't be removed. How does the volume of the recording affect the results? I didn't match the volumes. Are my results useless?
Here's some better results (http://sonove.angry.jp/RMAA/iPodClassic6_nano6_iPhone4.htm). Obviously my SNR and noise measurements are completely off. I'll run a corrected test later.
The volume cap can't be removed.
Sure it can. I know for a fact that we don't implement it in rockbox, since i reviewed some of the patches for the ipod 6G's sound driver
How does the volume of the recording affect the results? I didn't match the volumes. Are my results useless?
The frequency response will be unaffected, but it will probably make everything else look worse by about however many dB its capped. Both because they will be worse, and because your sound card will probably struggle to record such weak signals without introducing more noise.
That said,
Here's some better results (http://sonove.angry.jp/RMAA/iPodClassic6_nano6_iPhone4.htm).
.. those are a lot better. I tend to think that the system you recorded with actually has a higher noise floor then the devices you were testing so maybe the SNR results aren't going to be right regardless. Still, good to know all of the devices you tested have very high SNR
Thanks. I guess I'll just suck it up and buy a better recording device. I like the Edirol UA-4FX.
Well, I dusted off my old PC and tried three different PCI sound cards that were lying in a box. I got much better results (http://local.outpost.fr/rmaa/all.htm) with the Terratec Aureon Sky 5.1 (which wasn't the most expensive of the bunch). I measured a DR of about 92 dB. The iPod LOD is actually the winner with that setup, but all devices fared quite well (see individual results (http://local.outpost.fr/rmaa/)).
Well, I dusted off my old PC and tried three different PCI sound cards that were lying in a box. I got much better results (http://local.outpost.fr/rmaa/all.htm) with the Terratec Aureon Sky 5.1 (which wasn't the most expensive of the bunch). I measured a DR of about 92 dB. The iPod LOD is actually the winner with that setup, but all devices fared quite well (see individual results (http://local.outpost.fr/rmaa/)).
Which kinda proves a point I've been trying to make with some success for years, and that building a good potentially sonically transparent audio interface isn't exactly rocket science and doesn't take a NASA-sized budget.
On another forum, I recently saw a poster who just invested in this combo:
"I have just begun to convert some LPs to 192/24 digital files using PureVinyl and a TC Impact Twin." The interface runs close to $400 and the software runs about $130. I can see exactly what they offer as compared to simpler approaches, such as the ART USB Phono Plus Interface and Audacity, but the perceived need for all of these sound-modifying tools sort of shoots holes in the idea that the LP sounds superior all by itself. ;-)
And, lo and behold, my wonderful Asus N75SF (http://outpost.fr/rmaa/N75SF.htm) laptop with "Audio by Bang & Olufsen ICEpower". Seems legit!
And, lo and behold, my wonderful Asus N75SF (http://outpost.fr/rmaa/N75SF.htm) laptop with "Audio by Bang & Olufsen ICEpower". Seems legit!
Looks like you should inspect that one for any misimprovements present in the processing chain. Its sound implementation does not seem to be that great, but there is definitely something else going on that dominates the results. I'd say there's EQ and dynamic compression in use, presumably for better speaker audio. ICEpower would be referring to Class-D speaker amps.
BTW, measuring DAPs at anything else than 44.1 kHz is a potentially risky business. At least keep in mind that sample rate does matter.
I still have two of the aforementioned Terratec cards in use. As you found out, they are quite flexible when it comes to recording levels. They apparently were a co-development with ESI/Audiotrak at the time. I'd need to reverse engineer the headphone amp at some point, which I am told uses a Class B buffer (I never liked the result with anything less than 300 ohms much, though according to measurements coupling caps may be the bigger issue).
Could you put those galaxy nexus results back up? I wanted to look at them again, they were pretty interesting.
I deleted the old results, there were obviously off. I also noticed that I fucked up the links in my previous post. Here's the directory (http://outpost.fr/rmaa/), with the Galaxy Nexus (http://outpost.fr/rmaa/GN-QC15.htm) and the comparison (http://outpost.fr/rmaa/all.htm) of all three devices.
What happened with the frequency response on the Galaxy Nexus? I thought it was flat the other day? Or am I confused?
Edit: Also, were the previous results for the QC15s as well? Those are amplified, so they're basically the same as having no load at all.
The GN always had a slight treble roll-off. I tested with the QC15 because that's what I use. I'll post results with regular headphones soon.
You must be right. I'm probably confusing things.
Results with Denon AH-D1100 (http://outpost.fr/rmaa/D1100.htm) headphones (32 ohms) and with Denon AH-D2000 (http://outpost.fr/rmaa/D2000.htm) headphones (25 ohms).
I installed Rockbox on my iPod last night. I updated the comparison (http://outpost.fr/rmaa/D2000.htm) with measurements for the rockboxed iPod (http://outpost.fr/rmaa/iPod-rb.htm). Without the volume cap, DR went up quite a bit!
Those measurements look a bit strange for the ipod (in headphones), I mean the lobes around the 1kHz sine wave. Were you using it plugged into a power adapter? A smps supply could cause that. They dissapear on the rockbox version which I think is interesting. It would be good to find out why.
All tests were run on battery. Note that the lobes disappear with the LOD (http://outpost.fr/rmaa/iPod-LOD.htm) and stock firmware too. The big difference is the maximum volume. The stock firmware suffers from a substancial volume cap (6-10 dB I think) on the headphone out, which neither the LOD with stock FW nor Rockbox have.
These lobes seem to be an artifact introduced by RMAA when input level is low. It's a little cranky at times (and sometimes, more than just a little).
Those measurements look a bit strange for the ipod (in headphones), I mean the lobes around the 1kHz sine wave. Were you using it plugged into a power adapter? A smps supply could cause that. They dissapear on the rockbox version which I think is interesting. It would be good to find out why.
Those kinds of "lobes" are usually windowing artifacts. IOW, the data and the time window that was applied to it were mismatched.
I got an EMU 0204 USB and made new measurements (http://outpost.fr/rmaa/). I think I need a better stereo splitter cable for loaded tests, because the stereo crosstalk values are suspiciously high. I even managed to squeeze out 7 more decibels in one case just by wiggling the jack :-/
I got an EMU 0204 USB and made new measurements (http://outpost.fr/rmaa/). I think I need a better stereo splitter cable for loaded tests, because the stereo crosstalk values are suspiciously high. I even managed to squeeze out 7 more decibels in one case just by wiggling the jack :-/
Wise man makes his own lab cables.
On second thought, I'm not sure my stereo splitter cable is to blame. When connected only to the EMU's line-in, without headphones attached, I get crosstalk values below -90 dB. I've also found other (http://rmaa.elektrokrishna.com/Comparisons/Sansa%20Clip%2B%20-%20Impedances.htm) measurements (http://www.markuskraus.com/RMAA/clipplus/data.htm) of the Clip+ on the web that match mine (http://outpost.fr/rmaa/Comparison_Clip.htm), so I guess that high stereo crosstalk value is pretty much accurate. It's ABXable (http://pastebin.com/raw.php?i=f5A0JG5H) on material with content in one channel and silence in the other.
Crosstalk depends on current, so if you drive a high impedance source you won't see much, if any, crosstalk. But once the current ramps up you'll see a lot more of it.
Skamp,
Interesting measurements. I'm very surprised how EMU 0204 performs well comparing to O2 DAC without load, but not so great with your low impedance D2000 (25 Ω). EMU performs weird and that isn't surpise as it wasn't designed for low impedance phones. Its output impedance is 22 Ω. However it should performs well with high impedance phones.
I have similar EMU Pre Tracker USB and use it with HD650 and HD800 ( variable impedance, >325 Ω; (http://graphs.headphone.com/graphCompare.php?graphType=7&graphID%5b%5d=863) ) . Considering the main rule of thumb (nwavguy's) Zphones/8 EMU USB interfaces (22 Ω, Zo) should drive 250 Ω and higher impedance phones without any issue.
Shortly, it will be interesting to see your measurements with O2 and EMU 0204 with 300 Ω phones.
I'm not sure it will be enough correct to use variable resistor (potentiometer) as dummy load because an impedance of high-z phones is actually variable per frequency (see the previos graph). One way around is to use quite simple filters RLC for more realistic simulation of phones as a load, that shouldn't be difficult. Or just get a real 250-300 Ω phones.
BTW, as I have already mention to You on irc channel about my impressions of E17. Mmm, I prefer EMU Pre Tracker after all. Simply, it's cleaner. So I suspect that EMU interfaces could be comparable with state of art O2 DAC/AMP when used with high-z phones.
Well, the EMU shows a 0.5 dB bump at 30 Hz (http://outpost.fr/rmaa/EMU_0204_USB-D2000.htm), but that's really nothing. I don't know how much the dampening factor is affected, and I'm not sure I could hear it. A comparison of the EMU to the O2 for that wouldn't be easy since I would have to physically plug and unplug my headphones from one to the other instead of just recording outputs and ABXing that.
I followed these instructions (http://www.head-fi.org/t/603269/little-dot-mk-v-output-impedance#post_8265817) and I measured the output impedance of the O2 at 0.5Ω, the Clip+ at 0.5Ω, the iPod Classic at 5Ω, the Galaxy Nexus at 12Ω, the EMU 0204 USB at 23Ω and my laptop at 71Ω.
This thread is really interesting for me.
As a user of an iPod Classic 7G (MC297LL) bought in New York back in 2009 now it's being about time to think about the battery. Being in Europe at this time if i go the Apple route they will provide me with new iPod (EU capped). If i go my way i will have to open the iPod and risk the whole unit working condition.
Since i only listen thru multidriver balanced armatures (Westone UM3x and Shure SE420) when i want to do critical listenings i always amp the line out with an iBasso T3 (To get rid of the 5ohms of output impedance of the hp out).
Hence my main concern would be if the line out port is capped. It does not seem likely judging from the rmaa.
My secondary concern would be if from a subjetive point of view is there any QUALITY difference that is audible in the output of the stock 7G versus the Rockboxed 7g.
Since the Westone UM3x are really sensitive i don't think the volume cap will affect me even with soft recordings. And the SE420 at high volumes are not enjoyable from the hp out because the output impedance brings more forward the "already mid forwarded" sound of the SE420 transducers. But if quality is worse then i might go the risky DIY route to replace the battery.
BTW your last iPod Classic loaded RMAA was made with the unit running stock OS or Rockboxed?
Thanks a lot for all your work Skamp. Really appreciated.
Hence my main concern would be if the line out port is capped. It does not seem likely judging from the rmaa.
My former measurements indicate that it is not.
My secondary concern would be if from a subjetive point of view is there any QUALITY difference that is audible in the output of the stock 7G versus the Rockboxed 7g.
There is none, except for a proper EQ with Rockbox, as opposed to EQ presets on the original firmware, that sometimes induce heavy clipping.
BTW your last iPod Classic loaded RMAA was made with the unit running stock OS or Rockboxed?
Rockboxed.
Hence my main concern would be if the line out port is capped. It does not seem likely judging from the rmaa.
My former measurements indicate that it is not.
My secondary concern would be if from a subjetive point of view is there any QUALITY difference that is audible in the output of the stock 7G versus the Rockboxed 7g.
There is none, except for a proper EQ with Rockbox, as opposed to EQ presets on the original firmware, that sometimes induce heavy clipping.
BTW your last iPod Classic loaded RMAA was made with the unit running stock OS or Rockboxed?
Rockboxed.
Thanks a lot for all the input. I guess i could live quite well with a capped unit.
If you find the time to measure with the current gear the stock output it would be the ice on the cake.
And in any event thanks a lot and merry xmas!
Thanks a lot for all the input. I guess i could live quite well with a capped unit.
To be clear, Rockbox removes the volume limit.
If you find the time to measure with the current gear the stock output it would be the ice on the cake.
I will if I ever switch back to the original firmware again. The only difference is probably just ~6 dB less dynamic range.
I just got a Fuze+, it's pretty nice. I didn't bother with the original firmware at all, I immediately installed the latest Rockbox build, so I can't comment on it.
There's a slight hiss with my Shure IEMs (not with my Denons), but only when there's no music playing.
I measured the output impedance at 0.9Ω. Here are the RMAA measurements (loaded (http://outpost.fr/rmaa/Fuze-D2000.htm) / unloaded (http://outpost.fr/rmaa/Fuze.htm)). The frequency response looks kinda funky, but with 0.5 dB variations (at most), it doesn't matter.
I understand why some people hate the touchpad, but with Rockbox it's quite usable (sometimes even practical). The portrait-mode display is very efficient, as the album artwork can be displayed full screen with metadata below it. It's also very convenient for browsing large libraries. With the font I use (Helvetica, 15pt), it displays 20 lines, much more than my iPod Classic, and of course my Clip+.
Thanks for posting that.
Regarding
There's a slight hiss with my Shure IEMs (not with my Denons), but only when there's no music playing.
Do you mean while the music is stopped there is a slight hiss, but then when you start playing a file of digital silence it goes away?
Or the general background hiss of common music masks the slight hiss you hear while it's in play mode?
I realize that statement was ambiguous. Actually the hiss gets a bit more pronounced while playing 16 bit / 44.1 kHz silence. Only with the Shures though: with the Denons (25Ω), I hear zero hiss.
I didn't really hear any hiss with the Clip+, but I did hear some electronic noise (as reported by Rockbox users). The iPod Classic is 100% silent in all circumstances (including with the Shures).
It's strange that the noise and dynamic range values in the RMAA measurements (including those with the Shures attached) don't seem to be affected by it.
I realize that statement was ambiguous. Actually the hiss gets a bit more pronounced while playing 16 bit / 44.1 kHz silence. Only with the Shures though: with the Denons (25?), I hear zero hiss.
Dynamic range is generally independent of load unless the load is so heavy that it causes clipping.
If you hear noise with one set of IEMs and not the other then the difference is likely due to efficiency of the respective IEM
I didn't really hear any hiss with the Clip+, but I did hear some electronic noise (as reported by Rockbox users). The iPod Classic is 100% silent in all circumstances (including with the Shures).
Now you've lost me. Hiss
is electronic noise.
Now you've lost me. Hiss is electronic noise.
Pardon my layman terminology. The noise with the Clip+ is not hiss. The Rockbox website speaks of "CPU noise" (with certain hardware variants). It's the same kind of noise that I used to hear with my old HTC Legend smartphone.
Depending on what you are doing the voltage and clocks on the fuze plus change to conserve power. It's possible some settings have different noise. Power management was only recently implemented and not yet fully tested.