Saturday, 10 September 2016

RETRO-MEASURE: 2006 Apple iPod "Classic" 60GB 5th Generation

Here's a very cool Steve Hoffman forum thread I came across a few months back which "inspired" this post.

I appreciate the iPod. It was quite the gadget back in the day, foreshadowing the rise of touchscreen "gesture" devices. I remember being impressed by the intuitive scroll wheel (introduced in late 2001) which over time transformed to the "touch wheel" by 2002 then the "click wheel" by 2003 and of course "multi-touch" by late 2007 as it joined the family of touch-screen devices. By the time it became the Touch the iPod basically got transformed into an "iPhone Lite".


Looking through my box of gadgets, I found my old iPod "Video" 5th Generation (model PA003LL) from back in July 2006 (I know this because I got free engraving on it). It's got a 60GB original hard drive installed. Color screen, 2.5", with 320x240 resolution. Firmware version on this baby is 1.3, presumably the latest version since iTunes didn't ask me to update. It still charges fine but there is noticeable degradation with the battery life. The central button feels a little "sticky" these days after many many hours of use over the years, but still functions well. This is the penultimate hard drive based iPod superseded by the iPod "Classic" (6th Gen) in 2007. It has been said that this 5G and "late 2006" 5.5G models with 80GB HD were perhaps the best sounding on account of these models being the last to use a Wolfson DAC internally.

Of interest I think is that Stereophile reviewed and measured an older version of the iPod back in 2003. That was the 3rd Generation they looked at with of course earlier DAC chips.

From a historical perspective, I think the iPod will always have a special place in the heart of those of us going through the dramatic transition between Walkman/Discman to having thousands of songs and hours of playback in our pockets. As much as I appreciated the Apple IIe as a child growing up and learning programming (I was personally a Commodore VIC-20/C64/Amiga fan), or impressed by the Macintosh line and the early GUI (again, the Amiga and later the Windows method of operation was always much more intuitive to me), it was the iPod that most left an impression for me of Apple as a company. The way it worked, the on-the-go convenience of iTunes as a purchasing platform and music library still has a special place in my heart as one of the most influential of the many ways Steve Jobs changed the man-machine technological interface.

With that, let's perform some measurements on this "retro" device and see how it compares to the audio devices from these days...

I. Digital Oscilloscope, Output Impedance, Impulse Response, Digital Filter Composite

First, let's have a look at the square wave on the digital oscilloscope, 0dBFS, at 1kHz:

I was suspicious of the very flat peaks that we may be looking at clipping up there... And indeed with a 1kHz sine wave, this is the case at peak volume:

The loudest volume to use is 3 "clicks" lower to prevent this clipping:

It is therefore at this volume setting that I will be measuring the iPod 5th Generation. We see that the oscilloscope waveform shows excellent channel balance with a peak of 1.4V without clipping.

Measured at 1kHz over a 20-ohm load indicates an approximate output impedance of 11-ohms. That's relatively high compared to many devices these days. Higher impedance headphones 90-ohms+ would sound best with this device. Subsequent Apple devices I've tested show lower output impedance results - iPhone 6 at 3.2-ohms, and the iPad Air 2 at 2.2-ohms using the same method.

I was curious about the output impedance curve, so I measured the impedance over 13 frequencies and this is what it looks like compared to the recent PonoPlayer:

Compared to the PonoPlayer, the old iPod clearly has a significantly higher output impedance overall. There's quite an increase in the output impedance in the lower bass frequencies (up to ~23-ohms at 20Hz). Higher impedance headphones as expected will yield a better bass response. Note that from around 50Hz and above, the impedance settles out between 9 to 11.5 ohms on this device. For more about output impedance and headphones, look here.

Here's the impulse response:

A typical symmetrical linear phase impulse response with polarity maintained. Given the amount of ringing, we can surmise that this is a reasonably sharp filter which is demonstrated by the "Digital Filter Composite" I've been using for awhile courtesy of ideas from Juergen Reis:

As anticipated, a steep filter is implemented as shown in the yellow "Wideband white noise" plot. There is a little bit of filter overloading hence the ultrasonic noise floor not reaching the level of digital silence. The green "19 & 20kHz sine" plot shows reasonably good suppression of intermodulation and aliasing products. A nice "orthodox" sharp digital filter setting. In comparison, it's not as clean as modern devices like the recent Focusrite Forte or Light Harmonic Geek Out V2 measurements.

II. RightMark Tests & Comparisons

Setup:
iPod --> 6' shielded phono-RCA cable --> Focusrite Forte --> Shielded USB --> Measurement Windows 10 computer
All audio test data uploaded to the iPod with iTunes as Apple Lossless (ALAC) files.

As I mentioned in the last couple months, measurements will be using the Focusrite Forte as ADC from now on. Remember that results will be slightly different from the measurements using the old Creative E-MU 0404USB. When I post comparison tables, I will of course stay consistent with the same ADC device.

Testing the iPod is easy... This device was made before the day of all the various samplerates available with hi-res downloads. So we're basically dealing with 44kHz and possibly 48kHz files. We'll also see if the device benefits from 24-bit vs. plain 16-bits files.

First, here are the results at 44kHz:

As you can see, I'm comparing the iPod with the Light Harmonic Geek Out V2 USB DAC and TEAC UD-501 DAC as well as a modern PonoPlayer. Interestingly, the old iPod is able to play 24-bit files and this does seem to improve dynamic range but only marginally - less than 3dB difference. In any event, this is very much a 16-bit device and is outclassed by the other, newer devices.

A few graphs for your consideration:
Relative frequency responses from the different devices. The iPod is very flat all the way to 20kHz, very much like the TEAC UD-501 with sharp filter (the white and green iPod measurements are overlaid exactly on top of each other). As usual, the PonoPlayer at 44kHz has an earlier roll-off as previously discussed due to the filter setting used by Ayre.

Noise floor. Typical noise floors for 16-bit playback but the iPod is a little higher (worse) compared to others.

Stereo crosstalk. In my experience this is affected by the analogue cable used, so we must be careful in controlling this variable. The same cable was used in these tests and suggests the iPod was not as good as the newer products. Perfectly adequate in any event!

This is the IMD+N sweep measurement. iPod is a little higher than others with the PonoPlayer tending to be a bit higher than other modern DACs.

I was curious about the 48kHz samplerate with the iPod:
Alas I don't have data for the other devices at 48kHz since it's not a samplerate I measure regularly. In any event, it looks like the iPod 5G can handle 24/48... Again, not really able to benefit much from 24-bit audio data. Distortion values about the same as 44kHz.

Frequency response at 48kHz:

Noise floor at 24/48kHz:

III. Jitter

As usual, using the Dunn J-Test, here's what I see:


There appears to be a tiny spurious tone paired up with the 11kHz primary signal on the 16-bit test (about 100dB below the primary signal peak), but otherwise, no anomalies like the sidebands one anticipates with highly jitterish signals, and not the "skirting" one sometimes see with low level jitter. The 24-bit test with 12kHz primary signal looks clean.

Due to the iPod's higher noise floor plus the need to tone down the volume so as not to clip, jitter modulation tones are not visible in the 16-bit J-Test as they normally would be with other DACs. Bottom line in any event is that jitter does not appear to be a concern.

IV. Conclusions

There we go, another measurement of a "retro" device (last time was that old Sony laserdisc player). In summary, the 2006 iPod 5th Gen measures reasonably well overall. It has an output impedance around 11-ohms at 1kHz which suggests optimal sound quality can only be achieved with higher impedance headphones (like 90-ohms and above). This is especially important for the bass frequencies where impedance increases quite significantly starting at 50Hz, reaching above 20-ohms when measured at 20Hz! Observations around poor bass response have been reported over the years such as this post with measurements using different headphones. Clearly the flatter, more predictable curve for the PonoPlayer is much preferable. Remember though that there is a corresponding and more complex impedance curve for headphones (and speakers when we're dealing with amps), and this matching of impedance across the audio spectrum should be considered.

I haven't seen measurements of the Apple white earbuds originating from around 2006, but later Apple EarPods have impedance ~43-ohms according to InnerFidelity. Clearly these would not be ideal from an impedance match perspective (of course structurally, nobody would suggest these would be great headphones anyhow).

I thought I'd save it for here to show you a comparison with the lowest output impedance headphone DAC in my collection... The Light Harmonic Geek Out V2 (gray plot)!


The Geek Out V2 achieves a <0.5-ohm measured output impedance essentially across the frequency spectrum. This was measured at the lower power 100mW setting. What else can I say... Impressive!

At maximum volume, the iPod 5G will experience clipping so try to keep the volume control 3 units down from maximum. This clipping is not unique and was also found with the AudioEngine D3 reported in 2014 among others over the years.

The digital filter in this device is a typical sharp, linear phase setting with good antialiasing. Although this device was able to accept 24-bit audio data (unlike the Stereophile review's 3rd Generation model), it doesn't really benefit from the extra resolution. It does do a fine job with 48kHz material however if you have many of these; good for those 96kHz downsampled files I suppose.

Compared with DACs that are fed data through S/PDIF TosLink and coaxial interfaces, the iPod's jitter levels are trivial as one might suspect.

DAPs from about a decade apart!
Clearly the iPod performs quite well using lossless audio data for 16-bit DAC reproduction. Of course it's not up to the standard of 24-bit DAC's today... IMO, it sounds good and the improvements over the last decade are reflections of gradual evolutionary refinements in objective parameters, not massive audible differences. In fact, I believe that comparing the sound of the iPod with the Pono, the main difference is a result of the frequency response between the two devices rather than other parameters like the lower noise floor with the PonoPlayer. This is as a result of the digital filter used in the PonoPlayer which affects the high end plus the lower and more consistent output impedance, allowing better bass response across wider range of headphones. Furthermore the Pono doesn't clip at 100% headphone volume.

I see that one could open up these iPods and replace the HD with a new SSD or CF card using a ZIF converter board (something like this one, but check compatibility with your iPod)... That could be a fun hack although prying the case off doesn't look pleasant. This will provide easily more storage than the 60GB I have currently and improve battery life significantly. Maybe I'll give this a try some day... (Anyone given this a go, how did it work out?)

Also, don't forget software hacking with Rockbox! FLAC on an iPod anyone?

---------------------------

Well, in tech news this week we got the iPhone 7 unveil - water resistant, better camera, better battery life, stereo speakers, better screen with wider color gamut... Now available in black and jet black! Ooooohhhh, ahhhhhh... :-)

I don't know about you guys, but "Jony" Ive's voice and new-agey descriptions like "the most deliberate evolution of our original founding design... absolute unity and efficiency to the design... to define one truly uninterrupted form... single component dye absorbed by capillary effect... magnetized ultrafine iron particle bar used to polish... most singular, most evolved representation of this design" is, well, a little annoying (I'll reserve other adjectives at this time). I guess this kind of marketing impresses some folks who must find this "cool"... An appeal to the senses and subjectivity with a soothing voice, artistic visuals, and the voice of serenity :-). OK.

For audio lovers, as suspected (and it didn't take a genius to guess), Apple is aiming for less connectors and more wireless. No more 3.5mm jack built in, looks like a little Lightning-to-3.5mm DAC will be provided. The future as Apple sees it is the wireless AirPods headphones.  I must admit that although the technology looks interesting with IR sensors, beam-forming microphones, and motion/touch sensors, these AirPods just look ugly (my subjective opinion of course). Hope they sound better than they look (wonder if it'll be higher quality but still lossy Bluetooth CODEC like apt-X at least)! Still means yet another wireless device to keep charged. For those who need to look even cooler, I guess this is where the Beats line of upcoming wireless headphones might help (it's payback time for that $3B investment in Beats, right Apple?).

Audiophiles will have to ask themselves, at this point in history with Bluetooth technology as the likely wireless carrier, does this move in general bring us closer to higher fidelity, or are we moving further away? Note that this is a complex question. It might well be that a self-powered set of headphones will be able to achieve better frequency response by matching the built-in amplifier with the drivers. Plus "digital" headphones can implement DSP to smooth out irregularities. Even if the CODEC itself is lossy and the data resolution somewhat compromised, the overall sound from the transducers might be better. Remember my position on mobile audio - it's too noisy on-the-go for "high resolution" audio off a phone anyway, so I don't see technically pristine "perfectly" lossless audio output as a high priority in this context.

Interestingly, despite the loss of the headphone jack, the device is still 0.27"/0.28" in depth, same as the iPhone 6/6+, and no improvement in weight. The fact that there's no real space saving is a little disappointing I guess... It's not that I care about ever thinner phones, but I had thought the loss of this feature was supposed to allow an even smaller device, if even just nominally. I would not be surprised if Apple has underestimated the readiness of the public to let go of the analogue phono jack as a matter of convenience. We'll see with sales numbers and reviews in the months ahead.

Memory has doubled to 2GB which is good I guess, higher storage is also good, battery is better, cameras better (especially the iPhone 7+). No improved screen resolution (brighter and wider color gamut).

I'm sure this will sell, but I don't see it as being much of an improvement compared to the jump from iPhone 5 to iPhone 6 a couple years back especially since it doesn't look that different (of extreme importance when technology makes the jump from function to fashion)... I'm guessing my wife will be happy staying with her iPhone 6 and won't be asking for an upgrade this round. It'll be interesting to see how many of my Apple-centric friends upgrade. Maybe the Lightning-to-phono DAC adaptor is worth measuring out of curiosity.

Have a great week ahead guys & gals! As usual, hope you're all enjoying the music. And you can look forward to playing Mario on the iPhone, plus Pokemon Go on an Apple Watch 2 soon... Yeeehaw, fun times in the fall! :-)

Anders: You were looking for the LP Blind Test files from 2014... Here they are:
http://www.megafileupload.com/7v8x/Archimago_s_LP_Test.zip

Result of the blind test here.

11 comments:

  1. "Interestingly, despite the loss of the headphone jack, the device is still 0.27"/0.28" in depth, same as the iPhone 6/6+, and no improvement in weight. The fact that there's no real space saving is a little disappointing I guess... It's not that I care about ever thinner phones, but I had thought the loss of this feature was supposed to allow an even smaller device, if even just nominally."

    No, that's just what clueless rumor-mongers speculated. All you need to do is look at the iPod touch, which has a headphone jack and is substantially thinner than the iPhone, to realize that removing the headphone jack won't affect thickness unless the iPhone is considerably thinner. And I can't imagine a larger device like the iPhone 7 or 7 plus at that thickness would survive very well in some people's pockets.

    ReplyDelete
    Replies
    1. Good point Kirk...
      I remember the issues when the iPhone 6+ was released and the early complaints of bent phones, etc. from the thinness relative to size.

      Delete
  2. Those of us who use the audioquest dragonfly dac have already connected their headphones through the lightning part via the camera adapter. Same with the dragonfly and the old 30 pin connector. Old news, Apple!

    As to why Apple eliminated the headphone jack but didn't save any space? Waterproofing? Selling beats bluetooth headphones? Having future plans for this space? It is easier purchasing an adapter for your old headphone than lightning adapter headphones that you cannot connect to a non-apple device.

    ReplyDelete
    Replies
    1. And because of the latter, I will never purchase a lightning headphone.

      Delete
    2. Right Brause,
      I agree... I cannot see myself purchasing a Lightning-specific headphone. Unless it becomes widely accepted, I'll just leave that for Apple fans.

      Don't know about the Dragonfly. I hope the new models measure better than the V1.2 I tested awhile back.

      Good point about the waterproofing. Of course we've seen water resistance from other thin phones also - like the Galaxy S7 or maybe a Sony model.

      One feature I would have liked to see is wireless charging in the iPhone 7 since that feature seems well overdue and has been great in my experience.

      Delete
  3. I don't know whether the dragonfly 1.5 is worth it. I use the fly and the schiit fulla. While the fly is more versatile, the fulla is restricted to my MacBook Air. In direct comparison, the fulla sounds much better. And paired with the audioquest jitterbug, is sounds even better.

    My iPhone 5S does not sound much different with or without the dragonfly 1.5. In a blind test, I would not be able to tell the difference. Diminishing return, I'd say. Considering that the adapter is $35 CAD, this setup is cost prohibitive. The comparable 30 pin adapter for my old iPad is only $4 CAD online.

    The real difference in performance is between the fulla and the fly. The fulla is quite a step up from the dragonfly 1.5.

    ReplyDelete
  4. And I just listened to my Macbook Air setup with the fulla and with/without audioquest jitterbug: miraculously the sound is cleaner with the bug. Maybe you could investigate this little rascal.

    ReplyDelete
    Replies
    1. Hi Brause,
      Interesting comment on the Dragonfly vs. Fulla... The Fulla's AKM4396 DAC is certainly very capable although AKM has a newer generation of DACs out there.

      Yeah, overall unless headphones run out of steam with the iPhone output, I suspect there won't be huge sound differences. No matter what, to run the phone with the CCK and DAC hanging off it will be a hassle and it's just one of those things where the likelihood is very low that many actual commuters will bother with such a thing.

      Yeah... I'll see if I run into the Jitterbug whether there's anything to investigate. The Stereophile measurements last year showed no difference. I think already the interest in these kinds of devices are waning.

      Delete
  5. In another effort, I compared the iPhone 5S/dragonfly to the iPod Classic 160GB/Fiio E12 Montblanc amp. The latter setup had more oomph and warmth and analog feeling, the former may have had a slightly better soundstage. But: both setups were enjoyable and I could live with either of them.

    ReplyDelete
  6. My faithful 30Gb 5th gen iPod from 2006 developed battery issues (discharges very rapidly), then hard drive issues (clicking noise, iTunes can't see it)...all within the last few months :<. I considered repairs, but realized I could get far more storage than that using an SD card on my Android phone, for less money. And it can even run foobar2k. :>

    ReplyDelete
  7. Hey man, you should point out that the raising output impedance is caused by coupling capacitors at the output.

    ReplyDelete