Slashdot Mirror
Sony Thinks You'll Pay $1200 For a Digital Walkman
An anonymous reader writes: The Walkman is one of the most recognizable pieces of technology from the 1980s. Unfortunately for Sony, it didn't survive the switch to digital, and they discontinued it in 2010. Last year, they quietly reintroduced the Walkman brand as a "high-resolution audio player," supporting lossless codecs and better audio-related hardware. At $300, it seemed a bit pricey. But now, at the Consumer Electronics Show, Sony has loudly introduced its high-end digital Walkman, and somehow decided to price it at an astronomical $1,200.
What will all that money get you? 128GB of onboard storage and a microSD slot to go with it. There's a large touchscreen, and the device runs Android — but it uses version 4.2 Jelly Bean, which came out in 2012. It also supports Bluetooth and NFC. Sony claims the device has 33 hours of battery life when playing FLAC files, and 60 hours when playing MP3s. They appear to be targeting audiophiles — their press release includes phrasing about how pedestrian MP3 encoding will "compromise the purity of the original signal."
What will all that money get you? 128GB of onboard storage and a microSD slot to go with it. There's a large touchscreen, and the device runs Android — but it uses version 4.2 Jelly Bean, which came out in 2012. It also supports Bluetooth and NFC. Sony claims the device has 33 hours of battery life when playing FLAC files, and 60 hours when playing MP3s. They appear to be targeting audiophiles — their press release includes phrasing about how pedestrian MP3 encoding will "compromise the purity of the original signal."
I have one, and its technology sucks balls.
It's got a great DAC - an ESS SABRE 9016 - that powers many modern A/V receivers. Point there.
The problem is the amplifiers suck.
Ayre amps supposedly have no feedback, and that makes it "good". I suppose it is given they sell amps for $20,000 that are handmade in Colorado. However, just because you can hand make something doesn't translate into a mass-manufactured product. First off, the amp in the Pono is fully discrete (transistors, no op-amps). This is fine, if you manage to match all the transistors in each stage properly. Also fine in a $20,000 handmade product where you can go through and characterize every transistor and find matching pairs so they behave identically. But in a mass manufactured product, they probably are grabbing transistors off a reel, which means instant mismatches since they're within their specs, but will deviate due to manufacturing issues.
So a discrete amp already is at a disadvantage because without taking time to characterize every part, you're going to get an amp that behaves differently between channels and between units.
Yes, integrated units are better - best are dual units because matching within a die is far better (under 1% difference) that matching between dice (over 10-20%). IC designers know this, and they know that manufacturing can trim the differences down to practically nil within a die (in IC manufacturing, everything is based on ratios - you cannot say you want a 1K resistor because you'll get 1K +/- 30% tolerance. But you can design two transistors that will be well within 1% of each other, even if you need a 2:1 or 3:1 ratio or more - so designers work on ratios rather than absolute values). It's why you have dual DAC and dual op-amp or even more (6 channel DACs are common too) in a single package - the matching between the parts will be remarkably close, brought in closer because they can be laser trimmed during fab.
The next problem is lack of feedback causing a REALLY HIGH output impedance - about 5 ohms. If you don't know, this causes EQ because headphones with 8 ohm impedance can really vary between 1-12+ ohms over the audio range. This causes EQ (equalization) which means the amplifier actually produces different gains at different frequencies, a la a graphic equalizer. You can use an EQ to reverse this trend (that's what they're actually for - to equalize the response), but that's a bunch of processing. I've seen comments that say you should go for 8 times the output impedance at a minimum - so 40 ohm headphones or higher to minimize the EQ (at 8 times, the variance is around 0.5db).
Again, Ayre amps may do this because you're going to pair it up with good speakers that already will have higher impedances so you won't notice. But Joe Average will be using jellybean 8/16/32 ohm headphones (most common impedances).
The problem with Pono is that it hits EVERY audiophile rumor out there. Discrete good, op-amp bad (true back in the 70s with early opamps, but since the 80s we've had great audio op-amps that have excellent transfer characteristics). Feedback is bad (because feeding back a "time delayed" signal just ruins the audio purity - never mind that we're talking nanoseconds here) - even though using it lets you have lower output impedances. And that high output impedance means EQ up the hell.
And let's not say about the claim from Ayre themselves saying it's 80-90% as good as their $20,000 amp. That's just wrong on so many levels - are you saying that the amp is overpriced? Or to go the extra mile costs an extra $19,600?
Hell, I'm surprised they stuck with 3.5mm jacks given all the design work - 3.5mm jacks while convenient, do have limitations w.r.t. cross talk and other parameters.
And the hardware's kinda crappy - underpowered SoC running Android AOSP 2.2. yes, 2.2. it's sluggish all around.
I've actually never wanted to back out of a kickstarter as much as I have with Pono.