What High Capacity SSD For Hifi?

[XTRProf]

Say something pointing into the region of HDD?

Thanks again ............

[lark]

Jun 20, 2011 4:41:15 GMT XTRProf said:

Say something pointing into the region of HDD?

Depends on what you want to achieve with moving to SSD from HDD. The only benefits to be had for audio playback are:

1) Acoustically silent
2) Reliability (maybe questionable)
3) Low power
4) Shock resistant

Since all music players read the audio data into memory prior to playback (ram is used as a buffer) there can be no other benefit (music is not 'played' from the drive).

My personal opinion is that moving to SSD is a huge trade off in size or $$$.

SSD's make great sense for data intensive applications and operating systems but provide little to nothing (over the four points above) for music playback.

I'd much rather have a 2GB HDD for < $100 rather than 256MB SSD for > $300 - just seems a waste as cost per MB is not good Bang-For-Buck...

I use SSD's for my OS's for performance improvements but store all my media on HDD for convenience and cost effectiveness - Just think - a HDD has no problems what-so-ever streaming 1080p video - a FAR more intensive process than a piddly 3MB - 100MB audio file which is read into memory prior to playback (most modern machines have in excess of 2GB ram so this will never be an issue)..

Modern PC hardware is soooooo over spec'ed for audio playback (in terms of through-put, we'll leave DAC's out of this discussion) you really don't need to think about these things. An old Pentium 1 90Mhz machine will easily provide the data to a sound card for playback, and the old drives in these babies were oh-so-slow by today's standards)

Your money is better spent on something else.

[Deleted]

The only benefits to be had for audio playback are:1) Acoustically silent2) Reliability (maybe questionable)3) Low power4) Shock resistant

www.computeraudiophile.com/content/Do-solid-state-drives-have-audible-impact-musical-performance "

www.computeraudiophile.com/content/SSD-SATA-interface-recommendation
post - "Hi Clay - WHat I hear is a" -Chris Connaker

[lark]

My comments are based purely on a technical standpoint - data is read from the drive into memory (RAM chips), then processed by the audio software, then sent to the sound card. Unless you are getting buffer under runs, which will be very obvious pauses or pops during playback - of which I have never experienced in the last 10 years of using my PCs as my primary playback device. This was once a problem back in the mid 90's but those days are long gone (thankfully!).

There is no technical reason why a floppy disk, cdrom, hdd, ssd, punch card or paper tape would produce any different results - obviously some of these would require some serious effort and time to input the data into the machine (that is my stupid reference to floppy disks, punch cards and paper tape) But regardless of the storage medium (as long as it's digital) there will be no technical difference to the sound produced by the sound card - ie it will sound exactly the same, always, every time, without fail.

I can find threads on the internet to support just about any theory anyone can think up.... But I much prefer to use common sense and technical merit.

Even big commercial giants like Denon will sell you snake oil and claim it's a whole lot more: $499 for a $5 Cat5 cable - that just makes me laugh! ;D

[Deleted]

quote]There is no technical reason why a floppy disk, cdrom, hdd, ssd, punch card or paper tape would produce any different results [/quote]
lark
I wouldn't be so dismissive of reports from highly qualified members of other forums just because they can't be presently explained to the great unwashed masses, but you can most likely bet your left one that the reasons ARE known in many labs, but not made public due to commercial sensitivities.
That is why the Power supply area of some SSDs is quite different to some other makes, and no, it's not all about trying to avoid paying royalties !
There are many very highly qualified Recording Engineers (Chesky Brothers, award winning Cookie Marenco etc.) , as well as DAC designers (Charles Hansen, Gordon Rankin (of USB DAC software fame) and quite a few software writers who are members of that forum quoted.
I have already reported PSU interaction as have many other people, as part of the reason for some of these differences.
I am more than aware of "expectation bias", and I am damn sure , as are many other people are in this and other forums, that crappy SMPS power with it's often very high ripple, which is of high enough frequency to pass through conventional voltage regulators with only minor attenuation is part of the problem.
That is why I use a +5V JLH Linear PSU with my Corsair Voyager USB sticks.
The Halide Bridge for example is getting excellent reviews. It is a USB–S/PDIF converter that operates in asynchronous mode.
What helps to set it apart is the way it regulates the noisy USB port's +5V down to a squeaky clean +3.3V for it's internal precision oscillator.
Alex

[lark]

Jun 20, 2011 13:18:43 GMT @sandyk said:

I wouldn't be so dismissive of reports from highly qualified members of other forums just because they can't be presently explained to the great unwashed masses, but you can most likely bet your left one that the reasons ARE known in many labs, but not made public due to commercial sensitivities.
That is why the Power supply area of some SSDs is quite different to some other makes, and no, it's not all about trying to avoid paying royalties !
There are many very highly qualified Recording Engineers (Chesky Brothers, award winning Cookie Marenco etc.) , as well as DAC designers (Charles Hansen, Gordon Rankin (of USB DAC software fame) and quite a few software writers who are members of that forum quoted.
I have already reported PSU interaction as have many other people, as part of the reason for some of these differences.
I am more than aware of "expectation bias", and I am damn sure , as are many other people are in this and other forums, that crappy SMPS power with it's often very high ripple, which is of high enough frequency to pass through conventional voltage regulators with only minor attenuation is part of the problem.
That is why I use a +5V JLH Linear PSU with my Corsair Voyager USB sticks.
The Halide Bridge for example is getting excellent reviews. It is a USB–S/PDIF converter that operates in asynchronous mode.
What helps to set it apart is the way it regulates the noisy USB port's +5V down to a squeaky clean +3.3V for it's internal precision oscillator.
Alex

Alex, we've discussed my thoughts on this to some lengths - I am computer systems engineer as you know, and I am discussing the technical benefits of using a SSD over a HDD. Anyone who understands the logic flow of digital audio knows without a shadow of a doubt that this stage of the process can have no influence on what is delivered to the sound card - from here on it's the DAC and the analogue circuits - I'm not qualified to pass judgement on the analogue circuits, so I won't.

As I've said before, in the digital domain, bits are bits, there simply can not be any other artefacts that can be smuggled along with them - even if we wanted to introduce the idea that quantum entanglement could have some influence on the bits after they reached the DAC this still wouldn't hold true as it is not the same electrons that reach the DAC that were used to read the media. I know I've been through this with you before, but for the benefit of others (since the discussion has now been brought up here).

Data is stored on the storage medium (be it a HDD, SSD, floppy disk, etc) as a series of bytes. Each byte is represented by eight bits such that:

1 = 0000 0001
2 = 0000 0010
3 = 0000 0011
78 = 0100 1110

There is no room for interpretation. 3 is always 0000 0011.

With digital audio, each "sample" is computed as a number, a single number, in the case of 16bit CD audio it is arranged into 16 bit numbers, such:

1 = 0000 0000 0000 0001
2 = 0000 0000 0000 0010
3 = 0000 0000 0000 0011
78 = 0000 0000 0100 1110
56886 = 1101 1110 0011 0110

When the computer reads the audio data these bits are read from the medium and converted back to a number (still represented by bits in memory) - the medium uses checksums to ensure that no corruption has occurred, if corruption occurs your computer throws a hissy fit and fails. But then these bits can go through additional processing to suit the sound card driver and to apply any digital effects required. In the case of compressed audio, these bits are put through an algorithm to produce a totally different set of bits that have no resemblance to
the original bits. This is done using arithmetic such that:

2 + 3 = 5

or

0000 0000 0000 0010 + 0000 0000 0000 0011 = 0000 0000 0000 0101

This all happens by shoving these bits (or numbers if you prefer) in and out of memory (and cpu registers) many times over. Now since 2 numbers are taken and result in a totally different number (a 2, a 3 and results in a 5) there is no transfer of electrons from the 2 and 3 to the 5.... This is how digital data is used - always.....

This works perfectly every time - that's what's do different to analogue, analogue is VERY imperfect, digital is VERY perfect (with regard to reading and converting data)

Now no power supply, line noise, or anything else will effect this in anyway - if it does, it causes catastrophic failure of the processing logic and the computer fails (blue screen anyone?) or a checksum error ...

Now my point is, there is sooooooooo much processing done on the bits after they are read into memory that the medium it is read from is well forgotten before any of this processing happens.

Before any processing occurs, the application will read many bytes into memory - into a buffer. Typically thousands or millions of bytes are read in chunks into memory and then processed (all error free) - in computing terms it sits in memory for a massive amount of time before being used (computers do billions of operations per second, this processing only take a few of those billions) - hence the read speed of your drive is of no consequence here since all modern HDDs are way faster than required for the few piddly bytes required to be read to playback audio..

Now, what happens in the analogue domain, I'll leave for the analogue experts - that's not me.

Basically, I don't care what "(Chesky Brothers, award winning Cookie Marenco etc.)" think or say - unless their thoughts are going to change the way computers work their opinions can not change the reality.

That's be like trying to convince me that petrol doesn't ignite in combustion engine - it's how they work, any conjecture to the contrary is foolish.

I'm not taking other people word for how this works, I use low level debuggers and watch the flow of bits through computer systems (when bug/fault finding) - I see it with my own eyes on a daily basis. I'm not being fooled into believing this stuff - this is stuff I know, just like observing petrol explode in a combustion engine - there is no doubt, no question, just reality.

Sorry for the long rant, but this is not simple stuff to explain but only the boring details can explain it.

Lark

[toad]

Reaches for popcorn

[Deleted]

Jun 20, 2011 13:58:21 GMT lark said:

Jun 20, 2011 13:18:43 GMT @sandyk said:

I wouldn't be so dismissive of reports from highly qualified members of other forums just because they can't be presently explained to the great unwashed masses, but you can most likely bet your left one that the reasons ARE known in many labs, but not made public due to commercial sensitivities.
That is why the Power supply area of some SSDs is quite different to some other makes, and no, it's not all about trying to avoid paying royalties !
There are many very highly qualified Recording Engineers (Chesky Brothers, award winning Cookie Marenco etc.) , as well as DAC designers (Charles Hansen, Gordon Rankin (of USB DAC software fame) and quite a few software writers who are members of that forum quoted.
I have already reported PSU interaction as have many other people, as part of the reason for some of these differences.
I am more than aware of "expectation bias", and I am damn sure , as are many other people are in this and other forums, that crappy SMPS power with it's often very high ripple, which is of high enough frequency to pass through conventional voltage regulators with only minor attenuation is part of the problem.
That is why I use a +5V JLH Linear PSU with my Corsair Voyager USB sticks.
The Halide Bridge for example is getting excellent reviews. It is a USB–S/PDIF converter that operates in asynchronous mode.
What helps to set it apart is the way it regulates the noisy USB port's +5V down to a squeaky clean +3.3V for it's internal precision oscillator.
Alex

Alex, we've discussed my thoughts on this to some lengths - I am computer systems engineer as you know, and I am discussing the technical benefits of using a SSD over a HDD. Anyone who understands the logic flow of digital audio knows without a shadow of a doubt that this stage of the process can have no influence on what is delivered to the sound card - from here on it's the DAC and the analogue circuits - I'm not qualified to pass judgement on the analogue circuits, so I won't.

As I've said before, in the digital domain, bits are bits, there simply can not be any other artefacts that can be smuggled along with them - even if we wanted to introduce the idea that quantum entanglement could have some influence on the bits after they reached the DAC this still wouldn't hold true as it is not the same electrons that reach the DAC that were used to read the media. I know I've been through this with you before, but for the benefit of others (since the discussion has now been brought up here).

Data is stored on the storage medium (be it a HDD, SSD, floppy disk, etc) as a series of bytes. Each byte is represented by eight bits such that:

1 = 0000 0001
2 = 0000 0010
3 = 0000 0011
78 = 0100 1110

There is no room for interpretation. 3 is always 0000 0011.

With digital audio, each "sample" is computed as a number, a single number, in the case of 16bit CD audio it is arranged into 16 bit numbers, such:

1 = 0000 0000 0000 0001
2 = 0000 0000 0000 0010
3 = 0000 0000 0000 0011
78 = 0000 0000 0100 1110
56886 = 1101 1110 0011 0110

When the computer reads the audio data these bits are read from the medium and converted back to a number (still represented by bits in memory) - the medium uses checksums to ensure that no corruption has occurred, if corruption occurs your computer throws a hissy fit and fails. But then these bits can go through additional processing to suit the sound card driver and to apply any digital effects required. In the case of compressed audio, these bits are put through an algorithm to produce a totally different set of bits that have no resemblance to
the original bits. This is done using arithmetic such that:

2 + 3 = 5

or

0000 0000 0000 0010 + 0000 0000 0000 0011 = 0000 0000 0000 0101

This all happens by shoving these bits (or numbers if you prefer) in and out of memory (and cpu registers) many times over. Now since 2 numbers are taken and result in a totally different number (a 2, a 3 and results in a 5) there is no transfer of electrons from the 2 and 3 to the 5.... This is how digital data is used - always.....

This works perfectly every time - that's what's do different to analogue, analogue is VERY imperfect, digital is VERY perfect (with regard to reading and converting data)

Now no power supply, line noise, or anything else will effect this in anyway - if it does, it causes catastrophic failure of the processing logic and the computer fails (blue screen anyone?) or a checksum error ...

Now my point is, there is sooooooooo much processing done on the bits after they are read into memory that the medium it is read from is well forgotten before any of this processing happens.

Before any processing occurs, the application will read many bytes into memory - into a buffer. Typically thousands or millions of bytes are read in chunks into memory and then processed (all error free) - in computing terms it sits in memory for a massive amount of time before being used (computers do billions of operations per second, this processing only take a few of those billions) - hence the read speed of your drive is of no consequence here since all modern HDDs are way faster than required for the few piddly bytes required to be read to playback audio..

Now, what happens in the analogue domain, I'll leave for the analogue experts - that's not me.

Basically, I don't care what "(Chesky Brothers, award winning Cookie Marenco etc.)" think or say - unless their thoughts are going to change the way computers work their opinions can not change the reality.

That's be like trying to convince me that petrol doesn't ignite in combustion engine - it's how they work, any conjecture to the contrary is foolish.

I'm not taking other people word for how this works, I use low level debuggers and watch the flow of bits through computer systems (when bug/fault finding) - I see it with my own eyes on a daily basis. I'm not being fooled into believing this stuff - this is stuff I know, just like observing petrol explode in a combustion engine - there is no doubt, no question, just reality.

Sorry for the long rant, but this is not simple stuff to explain but only the boring details can explain it.

Lark

Lark.

As i have said in another thread, myself, and i feel many others, like to learn and gain knowledge from this forum, and in turn, try to add our own forte to the proceedings.

What an interesting and very understandable reply this is for digitally challenged members like me to read, had a rough idea how this system worked, but this is the very best i have read.

Excellent.

Mick.

[PinkFloyd]

Jun 20, 2011 15:01:52 GMT toad said:

Reaches for popcorn

I'm getting a hot dog and a glass of coke

[gommer]

Basically, I don't care what "(Chesky Brothers, award winning Cookie Marenco etc.)" think or say - unless their thoughts are going to change the way computers work their opinions can not change the reality.

Once, many moons ago, I took an exam of algebra at uni. While trying to prove a theorem, i missed a step somewhere in the middle. Of course, the prof. noticed immediately and i tried to cheat my way out by replying: "Why, that step is obvious, isn't it?". Although twenty years have gone by, i still remember exactly what that prof. replied to me. He said: "Dear boy, please look out the window and tell me, isn't it obvious that the earth is flat?"

I think this was one of the best lessons i ever learned: not everything is what it seems, even if it seems obvious.

I'm also a computer systems engineer. I now where you're coming from and i understand what you're saying. But what i'm also saying is: i'm in doubt that bit-perfect files can sound different. Why? I'm also a digital hardware engineer and an analog hardware engineer. I know that there are many possible factors that can play, which are not well understood yet. It all boils down to timing and jitter. How this can influence the output, coming from a digital source, whith -as you explained- all the different digital steps in between, i can't explain. But i can't say without a shadow of a doubt that there's no correlation between source and target in a digital to analog audio system (other than the intended audio contents of course).

I'm not choosing sides here, nor am i trying to refuel the heat. I have no wisdom to add, but I just couldn't resist posting.

Cheers, Marc

[rowuk]

I don't understand the problem. Those that "hear" a difference spend more money on hardware and are happy. Those that don't "hear" a difference don't by the big SSds and are happy. This is the choice of modern times. Something our ancestors NEVER had: 100 highest end headphones, 200 highest end headphone amplifiers, incredible advances in computer technology that reduce hi resolution rips to second.

My only comment to the golden ears: if you hear a difference between media, you will also hear a difference when the same file is copied to a "undeserving" register of the SAME media. Good luck!

[Deleted]

Depends on what you want to achieve with moving to SSD from HDD. The only benefits to be had for audio playback are:1) Acoustically silent

lark
Even that isn't necessarily correct. Although being phased out in newer models, many SSDs still use a "Super cap" in their voltage inverter section as well as an inductor.
There have been many confirmed reports of the buzzing of this inductor with SSD activity being clearly audible in a quiet situation.
There have also been several reports that some SSDs sound worse than a HDD, due to the large brief supply demands, in part due to supercap charging requirements most likely, reflecting through the rest of the system.I could speculate that such power supply interaction may degrade the ultimate stability of th 14.318MHZ Xtal oscillator from which all system timing is derived via PLLs etc.
All the technical explanations in the world will not change the fact that many RG members have heard differences between uploaded .wav files with identical checksums, and although not knowing which file was which, reported very similar differences, as well as consistently picking the more recent file as the better sounding.
Now of course, will come the usual request to prove it with an AB test, so beloved of engineers, and regarded by almost every person who hears differences as a a flawed procedure.
So here we are again, back to square one, with no possible resolution other than people like yourself having the differences demonstrated to them, as well as a demonstration that not all amplifiers meeting appropiate specifications sound the same . That part is clearly inconvenient due to the vast distance between us.
Even then, you would probably dispute the results!
Alex
P.S.
Yet once again, I think we should agree to disagree on this one, and leave it at that?

[XTRProf]

No, no, guys this is a great way to fly! Nobody knows all for sure and it's the combination of all the fields and knowledge that can effectively explain the you think you are "stupid" and you think you are "stupider" phenomenon that we always debate on in a thing such as this. No matter what, I for sure will leave this thread much wiser and not stupider.

Great and kudos to everybody participating. This is what a great forum should be.

[XTRProf]

Jun 20, 2011 8:44:59 GMT lark said:

Since all music players read the audio data into memory prior to playback (ram is used as a buffer) there can be no other benefit (music is not 'played' from the drive).

My personal opinion is that moving to SSD is a huge trade off in size or $$$.

I'm not the computer expert. You are. But how about those jitters or whatever errors that had already occur at the HDD. Wouldn't the ram be holding a "distorted" version of the data after being transferred from the HDD? Really, I'm not sure.

Jun 20, 2011 8:44:59 GMT lark said:

Your money is better spent on something else.

That's just being sensible based on what you know in your field. Ok, based on your field of knowledge, what do you think an ideal value for money pc audio setup should be like? So shoot your list of things you will have in your pc for audio. Btw, in another thread I had mentioned the AC Ryan as probably a suitable candidate as it's cheap and have all things that we would like to decode. I'm holding back as it's not yet 3D and hdmi 1.4 compliant yet. Yeah, this damn thing here .............







So will this be good enough to cover for pc audio? Yeah, hirez enough to tickle our hearing senses. I wish it can as that will save me a big chunk of time to setup a pc for audio as well $$$$.

[XTRProf]

Jun 20, 2011 12:16:51 GMT lark said:

My comments are based purely on a technical standpoint - data is read from the drive into memory (RAM chips), then processed by the audio software, then sent to the sound card. Unless you are getting buffer under runs, which will be very obvious pauses or pops during playback - of which I have never experienced in the last 10 years of using my PCs as my primary playback device. This was once a problem back in the mid 90's but those days are long gone (thankfully!).

Yup, fair enough based on what you are working on.

Jun 20, 2011 12:16:51 GMT lark said:

There is no technical reason why a floppy disk, cdrom, hdd, ssd, punch card or paper tape would produce any different results - obviously some of these would require some serious effort and time to input the data into the machine (that is my stupid reference to floppy disks, punch cards and paper tape) But regardless of the storage medium (as long as it's digital) there will be no technical difference to the sound produced by the sound card - ie it will sound exactly the same, always, every time, without fail

can find threads on the internet to support just about any theory anyone can think up.... But I much prefer to use common sense and technical merit.

Even big commercial giants like Denon will sell you snake oil and claim it's a whole lot more: $499 for a $5 Cat5 cable - that just makes me laugh!

Really, I wouldn't laugh so fast through a highend hirez system although the technical part of me will say to laugh and drop dead.

Even the bloody SPDIF interlink which is supposed to carry 1s and 0s for the dac to interpret will make the digital system sounds markedly different when we change from one material to another. Of course, it's in the degree and not cannot hear.

[Deleted]

I'm not the computer expert. You are. But how about those jitters or whatever errors that had already occur at the HDD. Wouldn't the ram be holding a "distorted" version of the data after being transferred from the HDD? Really, I'm not sure.

The answer is NO, that is the beauty of digital.
a 1 is a 1, a 0 is a 0 there are no in-betweens stored.
when there has been a reading error a 0 would become a 1 and a 1 a 0.

The copy would NOT be 'bit perfect' anymore and would be detected as a corrupt file as checksums would not 'fit' anymore.
depending on the used checksum/error correction/redundancy the wrongly interpreted bit can be recovered or not.

Take a small excerpt from a file containing music (a sample of a few seconds) and load it into a program that can edit the binairy numbers.
just edit one (or more) 1's or 0's to the opposite value and save. See/hear what what happens when you try to play that back (won't tell because you will have to find out).

There will be those that think a bit can be distorted or have say 0.9 value or another value when stored.
Ofcourse this is vVERY true for actual storage and data transport and is THE arguement that is always used to 'prove' digital is similar to analog and would explain subjective perceived differences.

a charge in a capacitor (or other form of SS storage) doesn't need to have an exact voltage/charge/value.
above a certain voltage/value it will be interpreted as a 1, below that same point (atually slightly above it, as a hysteresis is present) will be seen as a 0.
The same goes for optical bit retrieval and magnetic storage b.t.w. regarding the stored LEVELS.

Once 'wrongly' interpreted, this will be processed that way and depending on the error detection method be handled by software (marked as corrupted or repaired)

Jitter is present in transport of digital data streams and NOT in storage (it is 'visible' in magnetic and optical storage due to rotational errors/speed errors/rise/fall time errors, but ultimately, stored data on these media is NOT retreived at the rising or falling edge but in the exact centre of the stored bit (simply by doubling the data clock) and looking at a 'range' of values which will be determined to have been a 1 or 0.

The same principle applies to transport of data, be it serial or parallel, and be it electrical or optical regardless of format.
a Format only determines how bits and bytes are 'packed' and handled.

Jitter (and thus clock frequency accuracy/constantness) IS of MAJOR importance in 2 stages of the analog-digital-analog conversion process and is measurable.
This is where analog is converted into numbers and where the numbers are converted into an analog signal again.
WHEN jitter/inaccuracy of the clock is present, the analog signal is simply NOT sampled at the 'expected' time and slight errors in interpreted LEVELS can occur. In the anolog to digital domain this cannot be repaired afterwards anymore and will be present from then on.
The same thing can happen/happens when digital is converted to analog again but this time it is not a level but TIME that is affected.
a number which represents an analog voltage step is not clocked out at the expected time but a bit too soon or too late or anything inbetween.

Note that I clearly state only at the conversion stages this jitter has influence. NOT in storage nor transport.
There are those that are of the opinion that jitter in storage and transport also have an audible effect, yet this cannot be proven with any kind of electrical test. This boils down to the religious and or .. I can clearly hear it part. debating this is pointless as no evidence can nor will be found that can prove this in a court of law.

Now as Alex correctly states, power supply voltages in PC power supplies can vary and cause clock frequencies to deviate marginally.
depending on the used circuitery and immunity of these circuits to varying supply voltages this will give small errors that differ in kind at the moment of conversion ONLY.

Ofcourse everyone is free to suspect something more is going on that has the exact similar results as what can be found in analog.
This, however, is might be caused by not (fully) understandig the mechanics/workings of all processes involved and knowing what does what and when.

I can't comment nor recommend computer stuff, only have knowledge of data handling/storage/processing at electrical levels (not software) and the interface between digital and analog and analog circuitery.

These are my everyday working fields (digital/analog interface area and the analog domain and electrical data transfer) and my competences in this.
ofcourse this may be questioned as I do NOT know all there is to know about it.

How is that for a disclaimer eh ?

[XTRProf]

Jun 20, 2011 15:37:52 GMT PinkFloyd said:

Jun 20, 2011 15:01:52 GMT toad said:

Reaches for popcorn

I'm getting a hot dog and a glass of coke

Hey, this is not quite good enough! Where is the pom pom girls? Ok, I'm talking more like an American instead of British! Sorry.

[XTRProf]

Jun 20, 2011 20:30:38 GMT rowuk said:

I don't understand the problem. Those that "hear" a difference spend more money on hardware and are happy. Those that don't "hear" a difference don't by the big SSds and are happy. This is the choice of modern times. Something our ancestors NEVER had: 100 highest end headphones, 200 highest end headphone amplifiers, incredible advances in computer technology that reduce hi resolution rips to second.

My only comment to the golden ears: if you hear a difference between media, you will also hear a difference when the same file is copied to a "undeserving" register of the SAME media. Good luck!

Yeah, that's the problem with the so call "progress". Our ancestors don't care for such things. I remember telling my 80+ old man that I need to change this and that. The retort is usually, why do you need to change it? I'm very confident he is still in, and still is, ........

[XTRProf]

Jun 21, 2011 8:52:00 GMT @solderdude said:

The answer is NO, that is the beauty of digital.
a 1 is a 1, a 0 is a 0 there are no in-betweens stored.
when there has been a reading error a 0 would become a 1 and a 1 a 0.

The copy would NOT be 'bit perfect' anymore and would be detected as a corrupt file as checksums would not 'fit' anymore.
depending on the used checksum/error correction/redundancy the wrongly interpreted bit can be recovered or not.

Take a small excerpt from a file containing music (a sample of a few seconds) and load it into a program that can edit the binairy numbers.
just edit one (or more) 1's or 0's to the opposite value and save. See/hear what what happens when you try to play that back (won't tell because you will have to find out).

There will be those that think a bit can be distorted or have say 0.9 value or another value when stored.
of course this is very true for actual storage and data transport and is THE arguement that is always used to 'prove' digital is similar to analog and would explain subjective perceive differences.

a charge in a capacitor doesn't need to have an exact voltage.
above a certain voltage it will be interpreted as a 1, below that same point (atually slightly above it, as a hysteresis is present) will be seen as a 0.
The same goes for optical bit retrieval and magnetic storage.

Once 'wrongly' interpreted, this will be processed that way and depending on the error detection method be handled by software (marked as corrupted or repaired)

Jitter is only present in transport of digital streams and NOT in storage (it is present in magnetic and optical storage due to rotational errors/speed errors/rise/fall time errors but ultimately stored data on these media is NOT retreived at the rising or falling edge but in the exact centre of the stored bit (simply by doubling the data clock).

The same principle applies with transport of data be it serial or parallel and be it electrical or optical.

Jitter (and thus clock frequency accuracy/constantness) IS of MAJOR importance in 2 stages of the analog-digital-analog process and is measurable.
Thise is where analog is converted into numbers and where the numbers are converted into an analog signal again.
WHE jitter/inaccuracy of the clock is present the analog signal is simply NOT sampled at the 'expected' time and slight errors in interpreted levels can occur. In the anolog to digital domain this cannot be repaired afterwards anymore and will be present from then on.
The same thing can happen/happens when digital is converted to analog again but this time it is not a level but TIME that is affected.
a number which represents an analog voltage step is not clocked out at the expected time but a bit too soon or too late or anything inbetween.

Now as Alex correctly states, power supply voltages in PC power supplies can vary and cause clock frequencies to deviate marginally.
depending on the used circuitery and immunity of these circuits to varying supply voltages this will give small errors that differ in kind at the moment of conversion ONLY.

Ofcourse everyone is free to suspect something more is going on that has the exact similar results as what can be found in analog.
This, however, is might be caused by not (fully) understandig the mechanics/workings of all processes involved and knowing what does what and when.

I can't comment nor recommend computer stuff, only have knowledge of data handling/storage/processing at electrical levels (not software) and the interface between digital and analog and analog circuitery.

These are my everyday working fields (digital/analog interface area and the analog domain and electrical data transfer) and my competences in this.
ofcourse this may be questioned as I do NOT know all there is to know about it.
How is that for a disclaimer eh ?

That's precisely the problem. So who wants to be GOD and sacrifice himself to come back as an angel from GOD with a message to tell us why the objective and subjective cannot agree when present technical knowledge points to error free and so no difference? Still I wouldn't take a place in technical unless my real stuff ears tell me so for sure.

So who has a Doctor of Science in GAIA? That's the nearest to being GOD.

[lark]

Jun 20, 2011 18:45:31 GMT gommer said:

I know that there are many possible factors that can play, which are not well understood yet. It all boils down to timing and jitter. How this can influence the output, coming from a digital source, whith -as you explained- all the different digital steps in between, i can't explain.

Hi Marc,

I sort of agree here - yes there are timing issues that can cause jitter - but these do not stem from the storage medium - remember it's all read into memory buffers prior to processing.

I'm of two minds how much jitter can be perceived, but I won't go into that, instead I'll acknowledge that jitter is an observable artefact of digital audio and hence has the possibility to influence the sound (yeah, I'm fence sitting until I know better).

Jitter is typically introduced due to the transport between processor and DAC (USB, SPDIF, etc) and in the DAC it's self due to its clock implementation and limitations. Again, no influence by HDD, SSD, Floppy disks, etc..

Also consider, jitter in CD players is typically measured in milliseconds and has all sorts of hurdles to overcome due to the reliance on the reading the spinning disk being able to deliver the data in time - this can be influenced by the master quality used to produce the CD. But when ripping a CD, any possible jitter artefacts are totally removed since it is not time critical - playback is time critical because if the data is not received in time then this may be observable. CD ripping is not time critical because if the data is received late, no biggie as it will still be written 100% correctly and will not influence later playback (when it gets time critical again).

So, to recap, jitter in CD players is typically measured in milliseconds. Now if we start looking at USB (let's ignore internal sound cards for now, as their interface is very different and most here use external DACs). Ok, typically over USB jitter occurs a lot more frequently - a lot, but it is measured in nanoseconds - at least an order of magnitude less influence on the sound. If there is debate whether CD jitter has much perceivable influence on the sound, then there is even more debate if it can be perceived when we're talking nanoseconds.. - Remember I'm not making any claims over the ability to perceive jitter, I'm just explaining where jitter exists and what effect it can have.

This same argument can be made of most SPDIF implementations - again typically measured in milliseconds..

The CD and SPDIF jitter issues may not be so profound or even eliminated on some esoteric super expensive implementations, but not the gear we're using...

Conclusion - HDD, SSD, Floppies etc do not influence jitter on PC audio playback because they are too early in the processing chain and are not really in the time critical domain because of buffering..

[Deleted]

That's precisely the problem. So who wants to be GOD and sacrifice himself to come back as an angel from GOD with a message to tell us why the objective and subjective cannot agree when present technical knowledge points to error free and so no difference? Still I wouldn't take a place in technical unless my real stuff ears tell me so for sure.So who has a Doctor of Science in GAIA? That's the nearest to being GOD.

Now if THIS doesn't qualify as religion what does ?

[lark]

Jun 20, 2011 22:30:41 GMT @sandyk said:

Depends on what you want to achieve with moving to SSD from HDD. The only benefits to be had for audio playback are:1) Acoustically silent

lark
Even that isn't necessarily correct. Although being phased out in newer models, many SSDs still use a "Super cap" in their voltage inverter section as well as an inductor.

Hi Alex, but that is necessarily correct - I specifically said Acoustically silent what you are describing is not 'audible' you talking about it's effect on the analogue circuitry later in the processing chain where it might then produce noise (where it gets spat out the DAC - in the analogue domain).. I'll leave the analogue experts to discuss what happens there.. But I really don't believe that this will have much influence, but that's not my domain so I won't press my opinion.

If you put your ear to an SSD it will be silent, there are no moving parts like a HDD which is acoustically noisy - this is the silence I was referring to.

[Deleted]

There have been many confirmed reports of the buzzing of this inductor with SSD activity being clearly audible in a quiet situation.

There have been many confirmed reports of the buzzing of this inductor with SSD activity being clearly audible in a quiet situation.

lark
What I am saying here is that there are quite a few reports of audible activity noise emanating from SSDs due to the inductor in it's internal PSU section. It includes a voltage inverter(voltage step up) section .
Some believe it is due to high currents associated with the Supercap.
It is not unknown for inductors to buzz.
Alex

[lark]

Jun 21, 2011 8:25:31 GMT XTRProf said:

I'm not the computer expert. You are. But how about those jitters or whatever errors that had already occur at the HDD. Wouldn't the ram be holding a "distorted" version of the data after being transferred from the HDD? Really, I'm not sure.

Good point, but no - so jitter only effects the time critical domain - that is at the point where the data must be received on time - the only distortion here is time distortion (it came late) - but since it's read into the buffer first it comes down to the ability to read the buffer - here RAM, CPU bus speed, USB clocks, etc may have influence, but not the storage medium. And remember that RAM, CPU, CPU bus speeds are several orders of magnitude faster than the storage medium (billions or more operations a seconds and billions of bits per second throughput) in modern hardware it just isn't an issue.

Jun 20, 2011 8:44:59 GMT lark said:

Your money is better spent on something else.

That's just being sensible based on what you know in your field. Ok, based on your field of knowledge, what do you think an ideal value for money pc audio setup should be like? So shoot your list of things you will have in your pc for audio. Btw, in another thread I had mentioned the AC Ryan as probably a suitable candidate as it's cheap and have all things that we would like to decode. I'm holding back as it's not yet 3D and hdmi 1.4 compliant yet. Yeah, this damn thing here .............

Well, I'm not really qualified to say what the ideal setup is - this is a very subjective thing. My priorities are more about the user interface - it must store all my music, be easily navigated, accessible from my network (so I can control it from anywhere in my house) and sound sound great. There are soooo many choices that will fit this criteria, anything I say would just be a random pick from a bunch of good stuff - if you know what I mean. Your priorities will most likely be different from mine.

Perhaps start a thread stating asking about the most ideal setup for your needs - state your priorities and see what we come up with.

So will this be good enough to cover for pc audio? Yeah, hirez enough to tickle our hearing senses. I wish it can as that will save me a big chunk of time to setup a pc for audio as well $$$$.

I'm not familiar with this device - looks nice though and is similar to what I use for my HiFi setup.

I think a device like this could be good - PC's sorta suck for a dedicated audio setup - all the fans, heat, big box - all that power just isn't needed for audio playback - look what a little ol' iPod can do with such little hardware... (I'm not promoting iPod here, just pointing out that big box hardware is not needed).

At home my audio setup is probably considered low-fi by standards here. I use a Popcorn hour running MPD - MPD is the music playing software, that is controllable over my network so I can control it from any PC, my mobile phone, my remote control etc - this is high priority for me. The Popcorn Hour doesn't do this out of the box, MPD must be manually installed and configured. The the audio is passed over HDMI as raw digital to my Denon 3808ci amp which has it's own DAC - such my analogue path is extremely short (all inside the amp). This setup is configured for convenience and sounds perfectly good to me.. No doubt I could get better equipment, but this is all I need. Note, my amp has a dedicated 'Pure Stereo' function to provide unadulterated playback avoiding all the theatre processing which are common in these types of amps. The Popcorn Hour is probably a bad bit of kit here, but it only used to read the data from the drive and pass it (digitally) to my amp for decoding... For this it performs flawlessly. But this does not answer your question, sorry

[lark]

Jun 21, 2011 8:44:20 GMT XTRProf said:

Really, I wouldn't laugh so fast through a highend hirez system although the technical part of me will say to laugh and drop dead.

Even the bloody SPDIF interlink which is supposed to carry 1s and 0s for the dac to interpret will make the digital system sounds markedly different when we change from one material to another. Of course, it's in the degree and not cannot hear.

Naa, sorry can't agree here - your TV antenna cable, unless it is faulty will work the same as any SPDIF cable you'll buy - the difference is only in the physical characteristics of the cable - less prone to tangling, better connectors that wont corrode, etc, but the bits that flow through them will not be 'better' or 'worse' they will be the same bits and can not effect the sound. Caveat - it can not be damaged, should be the correct ohm-age (75-ohm) etc, but if it works, it works. That's the beauty of digital and digital interconnects.. Same goes for network cable, cat5, HDMI, digital optical, SPDIF, USB, etc. This is not by chance, it was by design.

The idea of interconnects making a difference is a bad translation from the analogue domain that just doesn't apply to the digital domain. (I think some will even challenge its applicability in the analogue domain, but I see more chance of it making a difference there).

I have some high quality SPDIF cables, but that's not for sound benefits, its for its nice look, quality connectors (for longevity) etc..