Swapped hdd with ssd....wow...wow....wow...........and wow...again!

[Listener]

if people are using standard PC power supplies ( switchmode) with their computer audio they will always be struggling to get sound quality that can approach a similarly priced audio quality music server. I know we like using our PC's for everything but they have real intrinsic problems with their power supplies when it comes to the best possible audio quality. The following link is just one solution but remember the electrical noise generated by the standard ATX power supply is in all electrical signals output by the computer.

http://www.itemaudio...ear_pc_psu.html

In case anyone missed it, here is my own experience with the same PSU that Item Audio sell in the UK.

http://www.stereo.ne...153#entry695153

With apologies to the OP for an off topic post and for pushing my own thread.

On topic, I have used SSD's for the music PC operating system for two years now and for my main PC for the last year. Fast and [to my old ears] a step up in audio quality. The next Music PC though will have an mSATA as the OS drive.

Steven.

[TP1]

In case anyone missed it, here is my own experience with the same PSU that Item Audio sell in the UK.

http://www.stereo.ne...153#entry695153

With apologies to the OP for an off topic post and for pushing my own thread.

On topic, I have used SSD's for the music PC operating system for two years now and for my main PC for the last year. Fast and [to my old ears] a step up in audio quality. The next Music PC though will have an mSATA as the OS drive.

Steven.

Well timed Steven. I thought the point was well worth being made in the context of understanding the limitations of the average PC when used as an audio source.

Edited November 2, 2012 by Tasso

[Tamir0678]

It's not about voltage accuracy it is about an impure electrical signal that in itself can degrade sound.

Hi,

Thanks for your comment. I'm still learning the different tweaks available, but I guess to do something, or invest money in it, I need to understand how it technically works to improve measurable fidelity. There is a lot of science and technology involved, but it seems snake oil made it's way into the PC realm as well (e.g. silver sata cables for hundreds of dollars...), so now I'm more sceptical about other improvements. From my experience some audiophiles want some equipment just because it's "better" and for the sake of being "better", regardless of objective measurable impact on sq. I'm not saying this is the case here.

Can you please explain how a slightly impure (+/-0.15%) DC signal can degrade sound in digital equipment? Where does the degradation take place (storage/mobo/ram/output device)? How does the impure signal overcome differential signalling and AC coupling to achieve signal degradation?

I'm not saying a digital signal can't degrade, it's just that this is usually associated with long cable runs between units using a different power source (hence no single reference) I really would love to understand the process from someone who's not trying to sell me something.

Edited November 2, 2012 by Tamir0678

[Tamir0678]

I don't really understand your question? Chain is PC -> JK USB / SPDIF MK3 -> DAC -> Amp. The analogue outputs are at the DAC.

I use an adnaco s3 USB hub.

Yes, that's what I wanted to know. (some soundcards have their own DACs). So you're PC only performs digital processing, which is good.

Very interested to hear the results of the blind testing.

[hochopeper]

Hi,

Thanks for your comment. I'm still learning the different tweaks available, but I guess I do something, or invest money in it, I need to understand how it technically works to improve measurable fidelity. There is a lot of science and technology involved, but I it seems snake oil made it's way into the PC realm as well (e.g. silver sata cables for hundreds of dollars...), so now I'm more sceptical about other improvements. From my experience some audiophiles want some equipment just because it's "better" and for the sake of being "better", regardless of objective measurable impact on sq. I'm not saying this is the case here.

Can you please explain how a slightly impure (+/-0.15%) DC signal can degrade sound in digital equipment? Where does the degradation take place (storage/mobo/ram/output device)? How does the impure signal overcome differential signalling and AC coupling to achieve signal degradation?

I'm not saying a digital signal can't degrade, it's just that this is usually associated with long cable runs between units using a different power source (hence no single reference) I really would love to understand the process from someone who's not trying to sell me something.

I am a diyer, nothing for sale from me, I earn enough in my day job. My day job works with 1000's of volts, so the area is a little foreign to me too as I am many many orders of magnitude out of my comfort zone, the issues in electronics design are very different to what I work with normally, though I am learning.

I am not an expert in this field but I'll try to convey the issue in laymans terms.

The problem starts when you consider that the DC signal is not dead flat as you might initially presume. It's actually only flat to a level of tolerance and it actually varies within this 'band'. On top of that conductors and all conductive materials, will act in a way that is similar (basically is) an antenna. An antenna may emit or receive noise from its environment. The components in a PC create quite a bit of environmental noise in the area surrounding them. They also do not produce a very 'clean' DC output. Combined, these issues impact everything in, around or connected to the PC.

I think I said it earlier in this thread, the level of noise allowable on a single ended DAC (assuming 24bit and roughly 120dB SNR) analogue output is in the order of 1 or 2 uV, that is 0.000001V. To achieve 16bit its more like 85uV, still very very low, less than 1mV.

That is the impact on the analogue side of the signal.

The digital signal is susceptible to this noise changing the time at which it triggers the binary 'high' / 'low' state. This means that while all the bits are there and 'perfect' they have still undergone a change compared to the data that is stored on your computer.

Also consider that the concept of 'ground' is also susceptible to the same issues. The ground connection may actually fluctuate in voltage with each switch between 'high'/'low' state of each individual digital signal. This can mean that something else switching will impact the reception of another signal in the time just after a signal is received.

With differential signalling the grounds of the two devices are typically connected and ground noise can be transmitted along this conductor. This may corrupt the reference point used for when a signal is considered 'high'/'low'. If the receiving device is not well designed, or the cable connecting the devices is poorly made, it can also be susceptible to radiated noise.

Having said all of that, there is good quality engineering designs can reduce the impacts of all of these. These good engineering designs are, as you have pointed out, not often the megabuck audiophile approved items that are discussed on hifi forums by enthusiasts.

I am aware that I have left out quite a number of things for the sake of simplicity but this should help put the reasons for this thread in perspective. I recognise that these issues are real and exist and as I have said previously, I plan to address them in a different manner.

A good article on the RF antenna behaviour of all electronics and conductors is here - http://www.hottconsu...s/dipoles-1.pdf

EDIT: I missed your questions about diff signalling etc in your last paragraph, I've edited the above to address that also.

Edited November 2, 2012 by zman

[Tamir0678]

Hi Zman,

Thanks a lot for the thorough reply. I appreciate it.

I am a diyer, nothing for sale from me, I earn enough in my day job.

Never meant to imply differently. You seemed above board to me from the start.

The digital signal is susceptible to this noise changing the time at which it triggers the binary 'high' / 'low' state. This means that while all the bits are there and 'perfect' they have still undergone a change compared to the data that is stored on your computer.

If possible, could you please explain how that works?

Why would it have any impact at the storage device level? If the data makes it to the RAM intact, wouldn't it be retransmitted based on the playback software's sync code (which uses the source's known sample rate, and a different clock)?

Also, why would there be such impact on audio when video is the real bandwidth hog, and appears to make it out at very high fidelity?

There must be equipment to check if what comes out of the PC matches the data on the drive...

[hochopeper]

Hi Zman,

Thanks a lot for the thorough reply. I appreciate it.

Never meant to imply differently. You seemed above board to me from the start.

Cheers, I know you didn't imply anything about me personally. I just like to be clear and thought it was worth stating that categorically I have no involvement in 'the industry' other than to mostly buy pcbs or components that I find technically interesting and buy parts for my own personal use.

If possible, could you please explain how that works?

Why would it have any impact at the storage device level? If the data makes it to the RAM intact, wouldn't it be retransmitted based on the playback software's sync code (which uses the source's known sample rate, and a different clock)?

Also, why would there be such impact on audio when video is the real bandwidth hog, and appears to make it out at very high fidelity?

There must be equipment to check if what comes out of the PC matches the data on the drive...

The theory would be that the additional noise either transmitted over ground connections between boards; OR, the noise picked up by poorly implemented digital interconnect systems would contribute to the noise seen in the signal at the DAC/spdif receiver/whatever.

The video signal is encoded differently and not as susceptible to timing differences, if the whole 'frame' of an image is delivered, it will be displayed when called on by the display's video clock. So in this way, the video signal is generally less susceptible to timing/jitter errors in a digital signal.

EDIT: yes, there is equipment, but measurement of the time domain performance of the digital signal is not able to check for shifts in the rising/falling edge by amounts of time like 1 picosecond or 1 nanosecond. It's not clearly understood, its not an easy problem to solve and it is often oversimplified.

Another thing that is often oversimplified and miss-understood is that the digital signal while it is being transmitted, is really an analogue RF signal. It is actually BETTER to let it have a slower rising time so that it looks more trapezoidal on the oscilloscope. As long as the rising and falling edges are consistent this will produce a cleaner transfer of data over the cable between devices.

Edited November 2, 2012 by zman

[TP1]

Hi Zman,

Thanks a lot for the thorough reply. I appreciate it.

Never meant to imply differently. You seemed above board to me from the start.

If possible, could you please explain how that works?

Why would it have any impact at the storage device level? If the data makes it to the RAM intact, wouldn't it be retransmitted based on the playback software's sync code (which uses the source's known sample rate, and a different clock)?

Also, why would there be such impact on audio when video is the real bandwidth hog, and appears to make it out at very high fidelity?

There must be equipment to check if what comes out of the PC matches the data on the drive...

I think its important to understand that while the digital information will determine what music will be recreated, other factors can still affect the process. Firstly with switch mode power supplies, a 'dirty' electrical signal carrying digital information ends up in your DAC. While the digital information is extracted accurately , the resultant electrical analog signal could still carry some of the electrical artefacts that were introduced at the input. Although the information is digital, the means of communicating the information is electrical. In addition, RF interference generated by normal switch mode power supplies gets fed back into the same electrical circuit that the rest of the equipment is also plugged into. Now it would be possible to have high quality SMPS that do not have these same effects but they don't seem to be currently available for computers.

For my part I use large balanced transformers between the AC mains and my hifi system ( 6 KVA capacity) to eliminate all sources of electrical impurities from the power supply, and the difference is not subtle.

Edited November 2, 2012 by Tasso

[drez]

There are also plenty of switching VR's on the motherboard itself, I am interested in the digital VRM's used in Asus' new motherboards as these are supposedly better. Generally the best switching computer PSU's available have around 5mV ripple, while that linear supply has 1mV, but I have also heard that switching PSU's can be better for high speed electronics. Some computer PSU's are coming out with digital VRM's and they measure very well, equal to the best switching supplies available: http://www.jonnyguru.com/modules.php?name=NDReviews&op=Story&reid=300.

[Tamir0678]

There are also plenty of switching VR's on the motherboard itself, I am interested in the digital VRM's used in Asus' new motherboards as these are supposedly better. Generally the best switching computer PSU's available have around 5mV ripple, while that linear supply has 1mV, but I have also heard that switching PSU's can be better for high speed electronics. Some computer PSU's are coming out with digital VRM's and they measure very well, equal to the best switching supplies available: http://www.jonnyguru...=Story&reid=300.

Yes, those are some crazy performance numbers on ripple and VR. Which means good things for the future of internal PSUs.

Now I'll just wait for them to put it in their 650-750w line at a more reasonable price... I don't really need this much power.

Edited November 4, 2012 by Tamir0678

[TP1]

Well I've learned something on this thread in terms of the relationship between sound quality and speed of data access. As an experiment I played music from a powered external drive ( I used a linear power supply......) vs the USB powered drive and the powered drive gave noticeably better results. My next move is to try an external SSD drive

[LogicprObe]

if people are using standard PC power supplies ( switchmode) with their computer audio they will always be struggling to get sound quality that can approach a similarly priced audio quality music server. I know we like using our PC's for everything but they have real intrinsic problems with their power supplies when it comes to the best possible audio quality. The following link is just one solution but remember the electrical noise generated by the standard ATX power supply is in all electrical signals output by the computer.

http://www.itemaudio...ear_pc_psu.html

PSU's of today, especially the higher end ones are nothing like the old ones or the current yum cha models.

Then again, just look at the filtering mounted on the current motherboards

This is another reason why you'll get a better sound from a desktop PC rather than a laptop.

[drez]

Interesting Haswell (Intel's next CPU range) will use on-chip VR, I wonder whether this will be a good or bad thing.

[LogicprObe]

I haven't read about it............but it'd be good for small applications.

Can't see the point for desktops and laptops though.

[THOMO]

Personally, I find it hard to understand how a SSD could make much difference over conventional HD.

Most music playback software uses some form of memory buffering rather than reading directly from hard disk. That is, data is read from disk into computer memory where it is held until the software needs it.

This means that you are already getting the benefit of the software reading from fast silicon. It would be a very slow, or very busy, computer that couldn't keep the memory buffer filled...even with hires music files.

SSD or conventional HD, the data in the music files will be read from there in to the same memory buffer.

Yes but if you make the change it HAS to sound better.

So people will think it does.

Unless you have two components that are exactly the same and you only change one and keep the other for comparison purposes then you will fall for this sort of stuff every time.

The world of computer audio seems to have overtaken cable and capacitor changing on the well trodden path of audiophile self delusion.

Edited November 5, 2012 by THOMO

[drez]

obvious troll is obvious.

[drez]

http://www.anandtech.com/show/6432/the-intel-ssd-dc-s3700-intels-3rd-generation-controller-analyzed

want

[LogicprObe]

http://www.anandtech...roller-analyzed

want

They'll only keep getting better and cheaper.

[arkticfox]

Hi all,

This is a little off topic so I apologize. I don't really see how this would make a difference but does the type of HD make a diff when streaming music wirelessly? For instance in my situation from a PC via a sonos? I have all my music on a spare 1TB WD green HD I pulled out of a WD MyBook which had karked it. Have plenty of bays in the PC for HDs (probably should set up a raid but don;t really know how to do it) so would putting all my music on to an SSD make any diff in this regard?

Cheers

sam

[LogicprObe]

The RAID is pretty easy to set up.

We can help you if you want.

SSD's as far as we know so far, are not any more reliable than spinners.

Just faster.

[drez]

Hi all,

This is a little off topic so I apologize. I don't really see how this would make a difference but does the type of HD make a diff when streaming music wirelessly? For instance in my situation from a PC via a sonos? I have all my music on a spare 1TB WD green HD I pulled out of a WD MyBook which had karked it. Have plenty of bays in the PC for HDs (probably should set up a raid but don;t really know how to do it) so would putting all my music on to an SSD make any diff in this regard?

Cheers

sam

IMO difference between HDD and SSD for playback is marginal - you would have to really listen hard to notice. I just happen to have an old SSD lying around which I can use to put a few albums on, but otherwise I would say investing in a large SSD is probably not the best way to spend money at this point in time. Maybe in a year or two things might be more affordable, and SSD performance might be even better.

[stuarth]

My CAPS v2 server started out with a 60GB SSD for the Windows 7 O/S plus JRMC etc, with FLAC files on an external 2TB eSATA drive. That drive was a bit noisy and eventually I moved both O/S and music files onto a 1TB WD Scorpio Blue 2.5" spinning drive (ditching the SSD to use elsewhere). Apart from slower boot-up, I noticed no decline in SQ with this arrangement, and the smaller drive is much quieter, but the slow boot-up was annoying as I power down the server when not listening. Also I'd had problems with JRMC getting corrupted so I could not always connect remotely using JRemote app on my iPod touch. So, a rebuild seemed in order.

Last night I slotted in an OCZ Agility 60GB SSD and re-installed Win7Pro on that then disabled all unwanted services. The spinning 1TB disc still lives in the case to store my FLAC/MP3 library. I've yet to do critical listening but I think this is the best arrangement for me so far.

[Tamir0678]

I just wonder with people considering exotic items and out of case power supplies, how many of you are using ECC memory?

[stuarth]

how many of you are using ECC memory?

I had to look up ECC memory - http://en.wikipedia.org/wiki/ECC_memory

I doubt that the Jetway mainboard in the CAPSv2 would support ECC, in fact it uses last-generation DDR2 RAM. JRMC and most other players now have a 'memory play' option so fast and accurate writing and reading to/from RAM _would_ seem to be important.

[drez]

I think ecc memory is for servers right? So that would mean Xeon systems? Has anyone tested ecc memory for this kind of application. For ram so far I have heard that low latency is good but I haven't really looked into ecc.