Ah Geez - the Hard Drive Does Make a Difference

[khollister]

So I get my new 2010 Mini music server going (8 GB, 40 GB SSD) with a spare G-Tech 2.5" external drive for the iTunes library (USB bus powered - I have a FW Duet as a DAC). As an experiment I ordered an OWC Al-Pro mini enclosure with 500 GB drive since I know it uses the Oxford 934 bridge chip which I saw mentioned as audio friendly somewhere. I am using generic short USB cable. So I loaded up some files on it this morning and substituted it for the G-Tech ...

 

It sounds different (better I think - more open and lively). As an engineer, I'm having a big problem with this, especially as I am using PM in memory mode. As a former high-end audio geek I'm not totally unsurprised.

 

Keith

 

[AVSGSLIU]

Keith,

Try comparing sound quality between different locations of identical copies of a file; for instance copy on original hd vs copy on new ext hd vs copy on your ssd. I would love to hear about your findings from an engineer's perspective.

George

 

[Codifus]

...daisy chaining your firewire DAC and OWC drive? As far as I can tell, your OWC drive has USB and firewire interfaces. Leave USB out of it completely and see how that sounds.

It would have been nice if your Mac mini had more than 1 firewire port, but I guess Apple lost that battle to USB:(

 

 

CD

 

[khollister]

No - general consensus seemed to be to separate the DAC and the hard drive. Multiple FW connectors wouldn't help - even on my Mac Pro (which has 4 FW 800 connectors), everything is on a single port.

 

I'll give it a try and see what happens.

 

 

[Codifus]

....using the firewire connection?

 

CD

 

[khollister]

Sorry - got consumed with the PM vs. Amarra comparison and forgot to try it. I'll get there :-)

 

[ItemAudio]

Increasingly people are joining the dots . . .

 

[Encore]

I'm also pretty sure that my OWC HD was an improvement over my LaCie HD, which had its own power supply. A friend suggested that this might be due to the elimination of a switch-mode PSU in the setup. If two switch-mode PSU's are running at different frequencies, their differential may lie in the audio frequency range. I'm not in a position to judge this possible explanation. BTW, I'm using a compaq netbook running Foobar.

 

[nicedrivr]

I had bought an OWC Al-Pro 1-TB drive before finding CA, based soley on recommendations for durability, etc, and have found it to be far and away the best external drive I've ever tried (multiple failures of LaCies and the like on a TV project led me to it). Mine has the 924 chip since it has FW 800. Are you aware of any diff between that and the 934 in the 934 chip?

 

[sandyk]

George

You are a wicked man!You are going to upset a lot of people if you claim that "bits aren't bits"!

Chris and attendees at a Symposium a while back preferred the sound of SSD over HDD.

Someday, more people will realise that POWER SUPPLY is everything, whether for digital or analogue.

Even using a decent USB stick with a low noise, and low impedance +5V PSU instead of noisy +5V SMPS from the USB port gives an SQ improvement on playback of.wav files ripped directly to there,(assuming the use of EAC and a low noise , low vibration computer) AND the improvement is still there when uploaded to someone on the other side of the world if their equipment is high resolution enough. A NYC based member can verify this too, as he assisted me recently when developing such a low noise +5V supply to be used for .wav file comparisons. Initial experiments were when using an USB Isolator, where a plugpack supplied the power for a +5V regulator to provide more current than a single 500mA USB port is capable of. Some recent USB sticks like the Corsair Voyager GT are even more thirsty when it comes to power, and are now USB 3.0 capable too.

Quite a few members will now also be aware of the improvement when a HiFace is used to convert from USB to SPDIF, and others will be aware of a marked further improvement when the HiFace gets it's power from specialised Lithium batteries.

Even powering a WD TV Live HD media player with a +12V decent Linear PSU gives an audible improvement when playing .wav files from a plugged in USB stick via Toslink to a high quality DAC. NO, it is NOT a "Placebo" effect !

 

I will not be going further with this in open forum , as I have been there and done that, some time back, and was mercilessly attacked by several members for daring to claim such heresy.

 

SandyK

 

[ItemAudio]

Sandy: This forum (on average) is almost on the curve, but even here there has been acrimonious debate about what makes a computer sound good - many still taking issue with the concept of such a discussion.

 

The quicker the self-appointed, half-informed IT 'cognoscenti' accept the difference between a printer and an amplifier, the sooner we will have wider comprehension of what matters in computer audio. But yes, yes and yes again: power supply, power supply, power supply.

 

If you don't think this is important, forget bits for a moment and reflect on how effectively you have joined your hard drive and PC mobo to the output transformer of your amp with all that efficiently conductive cable...

 

So keep plugging away about hard drives and USB and SATA cables, and software and whatever we will end up calling jitter before it officially becomes jitter: in time, this will all be axiomatic.

 

[bernardl]

Well, my question would be about the testing procedure leading to the conclusion that HDs do make a difference.

 

- How many tests were conducted?

- Were all the tests done blindly?

- Did they involve several people with no direct interest in the results?

 

Knowing how painful it can be to replace an HD in a mac mini for instance, I believe that many people claiming that there is a difference did not go through the trouble of doing all these tests. Typically, it would represent one or 2 days to do tests with real statistical significance.

 

Having replaced the stock HD of my mac mini by a high end OWC SSD, I want to think I hear a difference, but I cannot be sure because there was about one hour between the before and the after, I only did the test once and was alone when I did it. Subjectivity is so strong that these test conditions are meaningless.

 

Note that these comments are not related to the fact that I think the HD should make no difference at all if the communication between PC and Dac is asynchronous.

 

Cheers,

Bernard

 

 

[dmccombs]

Bernard,

 

We have done many tests comparing an internal stock hard drive and an internal SSD. We have the luxury of being able to A/B this since we have many Mac Minis on hand.

 

We have take a stock Mac Mini out of the box and one that we had just dropped in a SSD drive. We used the same cabling, music, etc, at the the same volume. The time to switch between units was just 2-3 minutes. The SSD advantage was obvious.

 

The SSD drives provides a lower noise floor, and all the good things that come with that.

 

[jhwalker]

WRT hard-drive vs. SSD-based playback: what if the player you are using pre-loads all the music into RAM before (or during) playback? Surely you're not suggesting music read from a hard-drive vs. an SSD sounds "different" once it's read? Bits are bits - I could understand (just) if you're saying that the signal path between the various storage devices might contribute to the overall noise of the system, but once the data is extracted from the storage device into RAM, there can be no difference in the sound.

 

[wgscott]

As the resident naysayer, skeptic and incorrigible vulture of ill-omen that I am, (and being also cheap and lazy), I have yet to replace my stock HD in my mini with an SSD. There is, however, one other very real reason why an SSD would be an improvement, even if you are playing from a memory chip:

 

The HD spins. The SSD does not.

 

Moving parts generate all sorts of noise (rf, inductance, physical vibrations, the sort of audible noise that accompanies disk thrashing, etc). So there is a non-voodoo reason why it could still be an improvement. I think in fact that is why people replace the internal drive with an SSD but still keep their music files on an external conventional drive.

 

[dmccombs]

JHWalker,

 

If we are talking about an internal drive, than the OS is still using the disk, so yes an SSD definitely sounds better than a conventional drive.

 

If you are talking about music stored on an external drive, then things are less clear cut. Which software player, how much memory you have for memory play, how many tracks you load, etc., can all have an impact. You may or may not here a difference with and external SSD.

 

[wgscott]

I replaced my HD in my macbook air with an SSD in July. Somehow two weeks ago the file system got corrupted, and i had to do a reinstall. The reinstall process, fortunately, was very conservative, and only replaced some system files, so after an update, I was immediately back to where I wanted to be. But it made me wonder if the file system is more prone to corruption with an SSD?

 

[Forehaven]

I missed this thread. I posted this yesterday:

 

"I'm using PM's memory and hog mode. In theory, since all music is put into RAM and plays from there, one should not be able to tell the difference b/w SSD or my ext. spinning drives. But I can, consistently, pick out the better sounding music off the SSD...it's more transparent, realistic, more details and longer decay just to mention a few.

 

I am surprised about this, and kind of bummed too. They don't make SSDs big enough, and too expensive. I know they only have so many read/writes before they die too, but boy, I wish this wasn't so. Spinning drives are so cheap and so big now. Just too bad 'memory mode' doesn't equal the playing field b/w the two types of drives. I may just copy files over to the SSD that I want to listen too, and delete them when I'm done. It's worth the effort, but who knows how long the drive will last doing this?

 

Anyone else notice this?"

 

Well, I guess is the answer is yes

 

 

All the reasons suggested so far, only the power supply seems like it'd make a difference if using RAM with PM or Amarra. Mechanical vibrations shouldn't make a difference if using RAM. Lowered noise of SSDs also shouldn't make a diff. if using RAM.

 

To be clear, this is with a mbp, internal OWC 200G drive, 4G RAM, compared to a ext. Hitachi Deskstar in an Oyen digital fanless box.

 

Interesting topic. And I agree with Darrell, the differences were obvious.

 

[khollister]

So my take on all of this is that while there is good empirical evidence that things like SSD vs HDD and memory play make an audible difference, I think people need to be very careful at the theory they attach to those observations.

 

I absolutely reject the notion that the data pulled of a SSD is different than the data pulled off a HDD in block mode. What I can accept as hypothesis is that the electromagnetic environment (radiated or conducted) is different between the two and that resultant EM environment ultimately pollutes the attached or nearby analog devices (again either via conduction or radiation).

 

As an engineer I can easily accept that we may not be able (either due to technology or knowledge) to directly measure a specific effect, but we should, at least, be able to theorize why/how a given audible effect might happen.

 

I am most curious about the obvious audible effects of different software, all of which claim to be bit perfect. Something is clearly going on, but none of the developers are talking. In fact, I wonder if the audible changes that we comment on in different versions of a given program are even intentional. My guess is that a lot of this stuff might be accidental because we do not understand the mechanism by which different software (which should all result in the same binary data) obviously either alters the data or generates a different noise spectrum that couples into the analog portion of the system.

 

It is all a little too much like PFM for my tastes, but I certainly won't say it is all BS.

 

[Codifus]

I remember how the Macintosh OS9 and below etc, before OS X, a user could completely turn of virtual memory and literally have a program like Photoshop reside in RAM. With operating systems like Windows NT and later as well as OS X, it is now as if there is no longer any control or capability by software as how to whether they load into RAM exclusively or not. The OS decides, and it never gives RAM completely to an application.

 

Could that be the case here? That could explain why SSD may be better. SSD is RAM pretending to be a hard drive.

 

CD

 

[jhwalker]

is the same as yours, khollister. Bits are bits, and data is data - there is no difference in the data coming off an SSD or off a rotating hard drive, so the sound *cannot* be different due to that difference. But there are certainly other factors that can come into play - underlying electronics, the relative isolation (or not) of the electromagnetic environment, etc.

 

I'm all for optimizing systems, but we need to be sure we're looking at real effects and real *possibilities*, not somehow entertaining the idea, for example, that data coming off an SSD is different from the same data coming off a rotating hard drive.

 

[The Computer Audiophile]

Hi jhwalker - Very well said.

 

[dmccombs]

JHWalker,

 

Yes, obviously the bits are the same between SSD and conventional drives. It is good you articulated that for the discussion though as it is of major importance. The difference in sound quality seems to be due to electrical noise related issues (and maybe speed for an internal drive).

 

So in addition to looking at SSD vs. Conventional drives, people may want to consider weather their external drive is plugged into the wall, or drawing current down the same cable as their audio signal. But, maybe this consideration branches from the "Hard Drive Makes A Difference" conversation a bit.

 

[khollister]

My point, perhaps too thinly disguised, was that I believe that most of the digital domain effects are either

 

A) Jitter in the case of synchronous streaming

B) Noise induced by either radiation or conduction.

 

Even if the digital data were somehow modified, it is impossible that the altered bits would be correlated in such a way as to have the subtle audible effects (too hard, no depth, etc) - it is far more likely that the random corruption would result in clear ticks, pops, etc.

 

I am much more likely to believe that RAM-cached vs disk streaming playback sound differences have everything to do with different noise spectra due to altered EM environments (HDD motors, rotating magnetic platters, increased current draw on the SMPS, etc).

 

 

[dsnyder]

Hello,

 

I agree with khollister's assertion that most difital domain affects are due to changes in jitter levels or noise leackage through electrical, RF, or EM. Fix the noise issue or provide better isolation for the devices that are affected by it and everything should sound better. Sometimes this is easier said than done.

 

-- David