External hard drives and sound quality


I've just about filled up the internal hard drive on my Macbook with music files and am now looking at external hard drive options. Was wondering whether folks report any difference in sound quality when playing files from an external drive versus the internal?

I'm especially interested in hearing people's experiences using wireless hard drives. An Apple rep told me it would be no problem, as the hard drive wouldn't directly interface with the USB output, but I of course always like to be skeptical of anything an Apple rep says.
coverto

Showing 9 responses by kijanki

Blindjim -

Your internal drive could be maxed out, or heavily fragmented. There is so many other possibilities - for instance Itunes could be updated or options changed exactly at the time of switch to new drive.

Fat32 is garbage. Using non-journaled format is asking for a disaster.

If bits would "decay" then checksum errors would be reported all the time with hundreds of thousand of files on computer, compressed files (including jpegs) would be corrupted etc.

The fact that you need partitioning software to split drive or "redo" OS every year tells me that you have PC and Windows (I'm sorry).

Firewire does not load main processor (has own processing unit) the way USB does. It is also more elegant, in my opinion, with daisy chaining instead of hub (only one cable coming out of computer).

Richard_stacy - I'm not sure why you bother with NAS, unless you have network of people using it. It does nothing other than making hard disk slower. Why not just plug external drive into computer. External drives are also cheaper than NAS drives.
Larry_s - I don't understand it either. If data is there the only difference could be jitter and it doesn't apply here since data is synchronous (clocked). In addition it goes thru disk cache buffers and output FIFO buffers. At first I thought it was about audible noise of device itself (HD vs. SSD) - this would make sense.
Richard_stacy - I completely forgot about laptops. It's good to hear that it can be done successfully - I might need it one day (wife uses laptop).

Blindjim - the most important that you can manage it over long time. Some things are better in Windows some in Mac. Friend of mine uses Dragon Naturally Speaking (Windows program) on the MAC instead of Mac Speech that is based on the same engine, because it works better for her. Multiple Sclerosis affected vision and her hands are shaking a lot.
I use 1TB drive but with close to 1,000 CDs in ALAC it is only 300GB. I keep second identical backup drive at work but am still worrying about something going very wrong during backup.
Mapman - Name brand of the drive is also very important. Names starting with sibilant like Seagate or Sony will bring more sibilant sound (logical) while name brands containing word "digital" - like Western Digital should be avoided at all cost (we all know what digital does to music). Word of caution about Maxtor. On the surface it looks OK but after closer look we learn that Maxtor bought HD business from Quantum in 2000 that bought HD business from DIGITAL Equipment Corporation in 1994. Again digititis could attack our system. One cannot be too careful. It is all perfectly logical but if you don't believe me, let me quote Herman's post:

"but I am saying it is impossible to predict what effect something will have on a system without trying it."
Mapman wrote: "I have used internal, external USB, and wireless NAS storage and I have heard no differences. Nor do I hear a difference having used three different computers as music servers."

Mapman - you're not going to win this argument because "somebody else might hear it". To me this is utterly nonsense that is anti-scientific and brings voodoo-harm to this forum. If we really don't know anything and have to test everything than perhaps we should test if red car has better gas millage than blue car before buying - because: "would be just as silly as you stating they don't because neither of us has any proof".

If we would pay attention to every possibility we would not have computers today.
Herman - The key is to make some sorting. Most of items you mentioned are in analog domain where things are getting extremely complicated but "bits are bits" is relatively easy to explain. Data stored on HD is retrieved without error - (not even single bit) or computers would not work. Data is extracted first to HD cache buffers and then to computer memory to end-up in FIFO output buffers. Data leaves FIFO buffer bit by bit exactly same (no matter what drive it came from). Digital data can only sound different if jitter is involved (noise in time domain). It doesn't apply here since transmission between HD and computer is synchronous (clocked). Computer has to present on its output exactly same information from the same file on different HD otherwise networks would not operate properly.

Asynchronous S/Pdif data coming from CDP could be an example of place where bit are not just a bits because of jitter.

It is remotely possible that you might hear different sound from different HD because of ground loops created by disk that affects analog audio - but it has nothing to do with type of HD. That would be pretty good example of case where experiment is useless (brings false conclusion).

People claim that jitter rejecting Benchmark DAC1 is still sensitive to transports therefore not exactly jitter rejecting. Strange part is that cheap DVD players often sound better with DAC1 (according to their testimony) than expensive transports. How is it possible? Again - wrong conclusion. Jitter rejection of Benchmark DAC1 has nothing to do with it and it is simply ability of given player to retrieve data from CD (DVD players have good tracking). It could be issue of ground loops as well.

Negative feedback that you mentioned indeed lowers distortion and sounds better but only if is used intelligently. Good designer would design amp as linear as possible to get below about 5% THD and then apply negative feedback to knock down THD to about 0.5%. Now it becomes necessary to lower input bandwidth to one that amp had before feedback was applied (to prevent TIM distortions) and voila - we got great sounding amp that nobody will buy because it has limited bandwidth and 0.5% of THD.
Once people encounter something counter-intuitive like digital cable affecting the sound they tend to stretch this to every possible case. Different sound of NAS drives could be related to something else then drive itself - for instance music was ripped two times to drive and is not identical - even if this is the same song/piece, or system was just turned on while second drive was tested later when system was warm and tweeters (ferrofluid) warmed-up. In addition to this we have placebo effect. In coducted tests people often swear to hear big difference while in reality they listen to exactly same set-up.

I remember joke about Russian scientist that was removing flea legs one by one ordering it everytime to jump. When he removed last one and flea did not follow his order he wrote a note "Removing all legs makes flea deaf"
One has to be very careful with conclusions.
Herman - Russian fleas are very special.
I would go back to my example with Benchmark DAC1. Even if jitter rejection of this DAC would be perfect people would still claim that it is not, because different drives sound different (read info from CD differently). I don't have any problem with that as long as people don't make conclusion that Benchmark is jitter sensitive.

In your case when exactly same data lands inside of processor memory it has to be outputted same way. Sound might be different - I don't question your hearing but not because of the fact that data delivered to DAC is different.

So many other things might affect test. We often forget about noise created by dimmer switches or the fact that radio stations have to cut power in half around 6PM (FCC regulation) or about warm-up of the gear.

The only thing I know for sure is that data delivered to D/A converter will be exactly same with different drives if they use exactly same file. Other than that I agree that they might sound different for the reason of electrical noise or ground loops (or something else) that affect analog section of the system.
If you use computer for anything else (I do) - use Firewire interface instead of USB. USB is loading main processor while Firewire is a little more intelligent and has own processor that handles transfers - less load on main uP. I also think that daisy chaining in Firewire is cleaner/better than USB hub. Firewire cable cannot be longer than 15ft (10ft for USB). Extra 5ft might be important if you want to hide disk (noise).