HDD or SSHD for music storage only

[hicr49]

I had only recently explored the debate between SSD vs HDD and there are people who take stand on both sides. Well, what about HDD vs SSHD (what some called Hybrid Hard Drive)? Will people think SSHD not be as good for music storage only purpose when no OS or program is running from it? Will there be the more noise from the SSD portion while the gained speed is not relevant with music reading?

 

Sorry if this topic had been brought up before. I ran a search but didn't came across any thread.

[davide256]

The answer is, it depends. In the specific case of a music server where the renderer and drives are an integrated unit, it will matter a great deal because HD's are both electrically and mechanically (fans, disc spin) noisy and require a higher voltage than an SSD. There are fewer SQ design problems to solve with an SSD. If the drives are network attached it won't matter much because the drives are network isolated from the renderer; the file server application and the SATA/ IDE controller solution are the limiting factors for NAS audio performance.

 

SSHD is no better than HD as an audio storage solution... the SSD part is small, used to boost frequently used files stored on the SSD part but most of the time you would be pulling data off the HD portion. Its a better fit for a laptop where the OS frequently used files can fit on the SSD portion. Keep in mind that its likely to die sooner because SSD has a shorter lifetime than HD and the SSD section gets used hard... mine died after 3 years

[botrytis]

I think an SSD would be a waste for music. I am currently using a USB3.0 HS (spinning) and there is no issues at all and that includes playing DSF files. On my main machine, it also uses and spinning HD for storing music. I only use the SSD for the OS and cache files.

 

SSD's are noisy also, so I wouldn't worry about the noise.

[DavidZ]

>>>>SSD's are noisy also, so I wouldn't worry about the noise.

 

 

I have 3 in my server and they are dead silent. Do the memory chips in your computer make noise too?

[botrytis]

On ‎3‎/‎13‎/‎2017 at 2:15 PM, DavidZ said:

>>>>SSD's are noisy also, so I wouldn't worry about the noise.

 

 

I have 3 in my server and they are dead silent. Do the memory chips in your computer make noise too?

The point is there are MORE causes of noise in a computer than the storage medium. The CPU, Video card, fans, and Power Supply all are sources of more noise than either a spinning HD or SSD. Once you crush all the former noises then maybe you will be able to hear the storage medium problem.

[DavidZ]

1 hour ago, botrytis said:

The point is there are MORE causes of noise in a computer than the storage medium. The CPU, Video card, fans, and Power Supply all are sources of more noise than either a spinning HD or SSD. Once you crush all the former noises then maybe you will be able to hear the storage medium problem.

Fair enough -- I can't dispute what you hear. For me, spinning drives and fans make noise. The Zuma I built following Chris' design has neither, and to my ears is silent. 

[botrytis]

I think that is more of your brain filling in noise because of intern all bias thinking about fn noise. Until you have done a blind test, than you cannot say that.

[Ralf11]

What is the source of electrical noise on a HDD?  Is it the stepper motors?

I am not clear how mechanical noise would affect the bit stream, and some sources of noise should be far away from the data cables, tho I guess we may only be dealing with a quasi-linear spatial effect as the cables could constitute a line source at an angle, not a point source...

 

[davide256]

Yes, motors generate electrical noise on any power bus. Talk to your power company sometime, you will be surprised how well they can enumerate the motor and compressor devices drawing power at your home from their load signature. It's been a well identified conclusion on this site in numerous threads that clean power and elimination of electrical noise issues improves audio performance. 

[Yucca06]

You can just power your HDD/SSD with an external PSU, so it will not charge your usb bus, or your MB.

No internal power : no (bad electrical) noise.

From a relative low cost iFi power (+ Sata adapter) to a Bakoon PSU, you have some choice.

(I connected my OS SSD to a Bakoon battery, and I have to say I heard some improvement when I did this, don't know if it's worth the price, though)

[botrytis]

3 hours ago, Yucca06 said:

You can just power your HDD/SSD with an external PSU, so it will not charge your usb bus, or your MB.

No internal power : no (bad electrical) noise.

From a relative low cost iFi power (+ Sata adapter) to a Bakoon PSU, you have some choice.

(I connected my OS SSD to a Bakoon battery, and I have to say I heard some improvement when I did this, don't know if it's worth the price, though)

Any power source will Give you noise.

[Yucca06]

Certainly, but you can get less noise (ripple) with a "better" PSU or a battery.  At least better than the crappy USB/Sata power you have with most of non modified  PCs today...

[Ralf11]

the motors that run a HDD are not quite as powerful as the ones a utility might detect on an AC power line

speculation is fine for forming a hypothesis, but I am interested in any tests of the hypothesis....

 

[botrytis]

5 hours ago, Ralf11 said:

the motors that run a HDD are not quite as powerful as the ones a utility might detect on an AC power line

speculation is fine for forming a hypothesis, but I am interested in any tests of the hypothesis....

 

I totally agree. In fact, there are 2.5" drives that run on very little power.

[hicr49]

I agree there are different electronic noise associate with different type of hard drives and power supplies. Beyond that part of the evil, I recently switched a portion of my music library from a smaller 250GB Samsung SSD to a larger Curial SSD and found there are good amount of auditable different between the two SSD in my system. I prefer the old Samsung SSD quite a bit better. Thou not sure if there's such thing as burning in a SSD. Will try compare again in a couple weeks.

[davide256]

On 3/24/2017 at 2:03 PM, hicr49 said:

I agree there are different electronic noise associate with different type of hard drives and power supplies. Beyond that part of the evil, I recently switched a portion of my music library from a smaller 250GB Samsung SSD to a larger Curial SSD and found there are good amount of auditable different between the two SSD in my system. I prefer the old Samsung SSD quite a bit better. Thou not sure if there's such thing as burning in a SSD. Will try compare again in a couple weeks.

I suspect the older Samsung had a better controller and less read latency...

[Ralf11]

do you have a way to measure jitter?

[audiventory]

In my opinion, into server must be operation system and other software only on SSD for lesser acoustical noise.

 

Music may be placed in HDD array in other room and readed via network.

 

Jitter, latency HDD and network for playback audio is not matter. Because there several intermediate buffers. Clock audio defined in DAC only.

[Ralf11]

if jitter won't matter then can there be any real difference in SQ between the two?

[botrytis]

6 hours ago, Ralf11 said:

if jitter won't matter then can there be any real difference in SQ between the two?

 

I agree here. It is all in your mind. Do a blind test, where you don't know which one is being used. This will tell you what is actually going on.

[davide256]

4 minutes ago, botrytis said:

 

I agree here. It is all in your mind. Do a blind test, where you don't know which one is being used. This will tell you what is actually going on.

I agree that one should empirically test  to see if you can hear  differences, with the caveat that a difference  you can't hear may reflect on the system limitations and not whether the difference exists.  The most helpful opinions come from those who have good systems and are conservative in  expressing views.

[audiventory]

9 hours ago, Ralf11 said:

if jitter won't matter then can there be any real difference in SQ between the two?

 

It is what I want to say. There is no difference between SSD and mechanical HDD for sound quality.

 

If such difference exists, it is something wrong in the system and must be fixed. But probability of such difference appearing is too low.

 

Need care about DAC's internal clock precision, but no need to care about jitter between computer and DAC, SSD/HDD and motherboard, etc.

 

DAC's buffer is asychronous. Reading clock independent on writing clock. Hence stability of input writing clock is ignored when audio data readed from buffer for digital-analog conversion.

 

Also there many same asynchronous buffers from SSD/HDD to DAC in hardware and software. But these buffers don't impact to result.

[Ralf11]

ok, if that is true then what is the need for a linear power supply for a mac mini

 

e.g. Uptone Audio makes one - so does it improve SQ and if so how?

 

I thought the idea was that noise injection caused jitter(?)

[audiventory]

If DAC synchronized by the internal clock, on jitter at audio output of a DAC device impact only internal clock of the DAC.

 

If there is some way, that some element of a audio system impact to internal clock of the DAC, it may impact to jitter.

 

Reducing noise of audio devices is not theoretical discussion. It is measurements and experiments anyway.

 

As example, you measure noise picture at audio output of DAC. Jitter may be checked as level of noise+harmonics with one element of the system. After it you replace the element and detect difference between the noise pictures.

[davide256]

12 hours ago, audiventory said:

If DAC synchronized by the internal clock, on jitter at audio output of a DAC device impact only internal clock of the DAC.

 

If there is some way, that some element of a audio system impact to internal clock of the DAC, it may impact to jitter.

 

Reducing noise of audio devices is not theoretical discussion. It is measurements and experiments anyway.

 

As example, you measure noise picture at audio output of DAC. Jitter may be checked as level of noise+harmonics with one element of the system. After it you replace the element and detect difference between the noise pictures.

Goodness me... you really do believe this is the best of all possible worlds where hardware is perfect and science has nothing new to learn :<)  Have a happy mid-fi life.