Computer Audio – SSD versus HDD

This may be hard to believe but the type of drive also has an influence on the sound

As an exception to my normal policy, in this article I will be publishing impressions from readers of this site as the main content in cursive font. Given my experience with computer audio thus far, I’m inclined to agree to all that is said. My own comment is in normal text.

READER REVIEW

There is a LOT of talk from the boys at Computer Audiophile about the superiority of the SSD over HDD for sound quality.

Well, there is NO doubt that the (Macintosh) G5 tower I have sounds fuller, more relaxed and with much greater sound stage width and depth compared to my Intel Mac Mini which sounded forward and very fatiguing. Both using Tiger, both using VLC as the player and both using a HiFace into my Purcell etc. I don’t know why, but the G5 sounds more correct, and that also matches up with your views on desktop vs laptop. BUT, with similar expectations (lower grain, etc) I carbon copied my G5’s HDD onto SSD and ran the G5 from the SSD but now my sound is thin, lean, no warmth, no body, but I have to say at least the treble is a tiny little bit better in that it seems to have a slight layer of hash removed. Booting back up from the HDD my warm mids and vocals have returned, stage depth, body etc and in actual fact, the treble has rolled off ever so slightly in a most pleasant (van den hul carbon cable?!) type of way.

This is the total opposite to what everyone is saying about SSD. For me, the HDD is far more musical! I won’t be going back to SSD. Ineteresting isn’t it, and if the technical engineering types were to be believed, there should be NO difference between cables, computers and storage devices!
I will continue my research.

Andrew Summerton

Even though I have not experienced this myself yet, this story matches up with my generic computer audio experience and my experience with CD transports in particular. It seems again and again that higher precision does not neccesarily translate to better sound, or a more musical delivery. On the contrary: it seems that the more precision you squeeze from an information carrier, the more analytical the sound becomes, laying bare the technical nature of the recordings. Of course, thinking along the lines of analog tape and its inherent losses this is theoretically a good thing because you want to retain all the information. But I think that it is sometimes taken too far. Maybe you need a bit of softening/rounding off to make the electronically stored music sound like real music. Remember: electronical/digital audio is not the real thing. It is a representation of it, played back through a series of conversions and amplifications. Therefore, if the replay chain becomes too accurate, you may start to hear the electronic process at work. This is just a theory of course but the more experience I gain, the more there seems to be some truth in it.

It seems logical to me that SSD’s could also sound different due to the different “DOS” used. Perhaps some stages of filtering/correction could be skipped. But this is all speculation. Extrapolating to CD playback: it may be one of the reasons that CD sounds better to me than a computer source for the same reason: error correction and conversion. I realise that I am making a whole lot of assumptions here so please take with a grain of salt. Think of it as food for thought if nothing else, but Andrew’s SSD story did fortify my ideas about this.

Until recently, I thought I even had a technical explanation. It was my understanding that normal hard drives make writing-errors every now and then and that they also have to work around bad sectors and such. As far as I understood, inside every hard drive there is a stage of error correction in order to work around this. Following that, it could well have been that SSD, for its accuracy, needs less error correction. Not entirely true, as it turns out.

Update 22-12-2013

ANOTHER READER REVIEW

I wanted to respond to the item about SSDs and hard disks, and your comment about error correction. As a B.Eng. in Electronic Engineering, and with a career in software and hardware interfacing, I have some familiarity and knowledge of this.

The summary: Hard drives – whether they be SSDs or hard disks – can afford no errors *whatsoever*. Even a single incorrect bit would completely destroy any program file. Error correction is applied at various levels (software and hardware) to allow for individual bits on the drive to fault, without affecting the data being read.

Hard drives will automatically move data around from what they consider to be ‘bad sectors’ – where errors have been detected – to other spare sectors on the drive. This is done automatically and invisibly while you are using your system. Note that typically data is not lost during these operations. The error correction used on the initial data recovers any errors found there, unless it is in a very bad state, in which case it is invariably detected as being in a very bad state!

A drive also does not know what data it is reading or writing. A single bit failure in a text file might have been acceptable for a user, but any bit failure in a program would mean unpredictable behaviour or failure of a program. It is simply not tolerated. Pretty much any unrecoverable error read by the system will render the file unusable, and the system will mark the whole file as such. Although unusual data failures very occasionally get through the error correction process, this is an *extremely rare* event. This is simply not something that affects regular audio data storage and playback.

I have not researched into why the audio file behaviour of SSDs and hard drives might be different, but will do so. It wouldn’t be due to the error correction though. Timing would be a much more likely factor, being the biggest difference between the drive systems…

Thanks so much for your articles, and for the time you put into it. I know you do it out of satisfaction of the listening itself, but you do help a lot of people with your articles!

Thanks!
Mark Gaz

Preliminary Conclusion

Because I have personally moved on to dedicated hardware for music replay I have not had the desire to experiment with this myself, but I do have an Aurender N10 that utilises an SSD to cache music on for playback. Compared to for example the AudioAanZee Reference Flow music server, the Aurender does indeed sound tighter and more articulate. But how much of this can be subscribed to the SSD inside remains anyone’s guess.

Based on my experiments with local HDD’s, USB drives and Network drives however I don’t have a doubt that music played from SSD sounds different from music played from a regular spinning hard drive. Whether you regard it as better or worse may be a matter of taste. Just be aware that there is indeed a difference in sound to be expected.

I have now received several more reports of people hearing differences. Some prefer HDD and some prefer SSD. Amusingly, one person preferred SSD at first, but was converted lateron. So I guess this is totally a matter of how the sound fits in with the rest of your system as well as a matter of taste. I have no doubt that SSD potentially makes for a subjectively more “accurate” sound, but that doesn’t mean that it is also perceived as more musical.