Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 17:16:25 GMT
Guys (probably aimed at Sandy and Frans?), **I have vague idea that this is anoter thread somewhere that this question would possibly fit in but I can't find it, so feel free to relocate it. Thinking ahead, a long way ahead perhaps, to when SSDs are much more affordable, I was wondering if installing one in the main PC (not stand alone) with a relatively small capacity, just installing foobar and/or JRiver on it and setting it up to load tracks from a seperate HDD all into memory/RAM for playing would give any benefits in resulting sound quality? - any thoughts? **I'm quite happy receiving different PoVs, all replies will be actively considered TIA, Dave.
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 19:37:27 GMT
I CAN recommend SSD's. But mostly because they are much faster. The main improvement of a SSD is the much faster speed which accelerates every computer greatly.
To be honest, I wouldn't buy a SSD just for getting better audio quality. I have two fast computers here which have both an ASUS Xonar Essence STX audio card. Both have used harddrives initially, but with dropping SSD prices, I've changed for SSD's. If there was an difference, it was certainly small in my case.
The Xonar STX cards have good filtering and shielding of the analog output stage. Electrical noises are filtered already very well. These cards are dead silent and have a great signal-to-noise ratio.
I use both cards with Linux, which has a more advanced audio core (ALSA) than Windows. I wouldn't be suprised to hear some Windows user reporting that they hear an improvement. But that's IMO mainly related to the much faster disk access and higher data transfer rates. This is certainly a benefit for lower audio latencies.
If you compare different operating systems on the very same computer which the exactly equal configuration, you'll notice that Windows does access the harddrive a lot more than any other operating system. IMO, that's another indicator how flawed and crappy Windows is programmed. But since Windows does access the disk very often (even without running anything beside the naked OS), the felt improvement on the speed is even greater that with other operating systems.
Alex is IMO at the moment a bad source for any judgement about SSD's. He doesn't have a SSD and his computer wouldn't be able to make full use of the speed of a SSD. Even a SATA-6G expansion card for the PCI-E bus wouldn't cure that problem entirely. But, of course, even an older computer will see a great speed improvement.
I wouldn't be suprised to read that Alex is reporting an improvement if he'd buy a SSD. I would believe him. But I'm almost sure we wouldn't agree on the reason for this. IMVHO the main improvement would be lower audio latency with the Windows operating system.
It's certainly a plus that a SSD doesn't use a spindle motor like any harddisk does. This means less electrical noise. But I have real doubts that this makes a big difference.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 21:16:07 GMT
Guys (probably aimed at Sandy and Frans?), **I have vague idea that this is anoter thread somewhere that this question would possibly fit in but I can't find it, so feel free to relocate it. Thinking ahead, a long way ahead perhaps, to when SSDs are much more affordable, I was wondering if installing one in the main PC (not stand alone) with a relatively small capacity, just installing foobar and/or JRiver on it and setting it up to load tracks from a seperate HDD all into memory/RAM for playing would give any benefits in resulting sound quality? - any thoughts? **I'm quite happy receiving different PoVs, all replies will be actively considered TIA, Dave. Hi Dave I have already stated elsewhere that not everybody agrees that SSD sounds better than HDD. Some earlier SSDs used internal "supercaps" which were reported to cause more PSU noise, as well as audible noise emanating from the internal inductors due to the current pulses . I would expect that if using something like a SOtM Sata filter for their PSU, or a C-L-C filter in line as I am now doing, that any SQ differences should be negligible. Alex
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 21:33:35 GMT
Alex, what do you think: How good is the filtering/shielding on the Xonar cards? I'm interested to hear your opinion.
I don't think the Xonar cards are perfect and I also don't think my computers are perfect for audiophile use. They are simply NOT designed just only for audio use.
But I can really say that the Xonar cards are DEAD SILENT, even with a lot of electrically noisy components in computers.
There are, of course, a lot of different Xonar cards. I have two Xonar Essence STX cards. These have their analog output section specially shielded. The Xonar Essence ST is the PCI variant (STX is for PCI-E) and should behave almost the same IMO. I have absolutely no clue how the other Xonar cards would perform. Some of the other Xonar cards are quite basic designs and are much cheaper than the ST/STX.
I'm really taken by the Xonar cards. To me, they seem a very easy way to get good (not perfect) audio with a PC. Even crappy audio streams sound usually excellent. I have also used some documentary videos (which have been grabbed from TV). Sound quality of the videos is shite, but the quality seems improved a lot on the Xonar when I compared it with other computers. Everything seems to be automatically upsampled to 24bit/192KHz.
The Xonar Essence STX sounds even better than my Cambridge Audio Sonata CD30 CD-player, which has also an excellent upsampling 24bit/192KHz DAC.
|
|
|
Post by dalethorn on Jul 29, 2012 21:37:08 GMT
Once you run the software from the SSD on a Windows computer, all of the bytes reside in the Windows RAM cache, so it might make more sense to ask about the quality and speed of your RAM chips and the various 'bus' connections that communicate between that RAM and the player and audio drivers.
|
|
oldson
Been here a while!
Older than a well used glove
Posts: 414
|
Post by oldson on Jul 29, 2012 21:40:18 GMT
i just got a new pc with an ssd as main system drive. i do not notice any sound improvement, foobar and my music are stored on it.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 21:56:55 GMT
Christian Although the Asus cards get great reviews , as well as measure well,and sound way better than many other cards via Analogue Out, I find that the SQ doesn't even come remotely close to what I can obtain from their coax SPDIF into my highly modified X-DAC V3. It's in a different ball park entirely. Alex
P.S. The MF X-DAC V3 also upsamples to 24/192.
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 22:13:42 GMT
Once you run the software from the SSD on a Windows computer, all of the bytes reside in the Windows RAM cache, so it might make more sense to ask about the quality and speed of your RAM chips and the various 'bus' connections that communicate between that RAM and the player and audio drivers. I agree with that in general. But it's a requirement that the computer has enough RAM at first. If it is constantly swapping, it will be slowed down which isn't very beneficial. It's hard to qualify the quality of RAM without having ECC memory. Even I have not absolute proof since I don't have ECC RAM. Even with ECC RAM, there are possibilites when a memory error isn't discovered and corrected. RAM with "chipkill" goes further another step, but there are still possibilites for it to fail. I can only recommend Windows users to test their computer also with another operating system (for example Linux). They'll notice that Windows usually has much more disk access than other OS. I've noticed that with various computers at work, which have been reused with Linux as terminals. I know also some system technicians that say the same. It's almost constantly accessing the disk to various extent. I've seen this behaviour with about every release of Windows. Including the server variants. WinXP is by far the worst one, unless we look back even more (thinking of W95/98/Millenium). BTW: Excessive swapping was not a problem with those systems. On the other hand, if just a few bits are falsified, the change that this isn't hearable in most situations is very high. It's a different thing for computer programs. Just one falsified bit can be enough to let your computer crash. I've seen the results of a study from Google. It was about memory errors on servers. Google has one of the biggest (if not the biggest) fleet of servers. They have monitored their servers over a couple of years and they've found out that memory errors occur much more often than was believed before. IMVHO, every computer should have at least ECC RAM. But that's a no-no for the industry since it would increase the overall cost of a computer. Some no-name memory sellers would be out of business almost instantly since their flawed memory would be much more likely to be discovered by the users. I've liked some of the older AMD Athlon CPU's lot. I mean the early designs which had the memory controller integrated on the CPU die. These designs have allowed ECC and non-ECC memory (but not mixed configurations). It was the decision of the customer which memory modules he has preferred to use. There was a choice. Today, you have to buy very costly hardware to get only the possibility of using ECC. I don't worry too much about audio quality when thinking about ECC. ECC would be mainly beneficial for overall system stability since it can detect and correct some memory errors.
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 22:14:35 GMT
Alex, which Xonar card(s) do you use?
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 22:19:22 GMT
Just for the audio quality, I'd agree that the best DAC's are probably in an external case.
PC's are quite noisy.
What I really do like the the Xonar Essence STX is the built-in headphone amp. It's sounds very good and the signal path is short.
The most advanced Xonar design is also an external DAC with own toroid. But it cost a lot more than the PCI(-E) expansion cards, which isn't really a surprise.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 22:29:24 GMT
Alex, which Xonar card(s) do you use? D2X. I have also heard a cheaper Asus Xonar card that I gave to my son after I discovered the D2X problems were due to an "iffy" molex 4 pin connector. (intermittent relay dropouts) I have also heard a friend's Essence STX. It's specs are slightly better , but the audible difference wouldn't be obvious to most users.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 22:50:04 GMT
For those with deep pockets the ideal solution will be a USB DAC connected with a USB optic fiber extender which will provide full galvanic isolation from the PC. It is a 2 box solution, one box is connected to the PC with a standard USB cable, it convertes the electrical USB data to optical and completely removes the electric part, sends it over glass optic fiber to the second box (with its own upgradable PS) wich converts it back to USB data plus power and there you would plug the USB DAC. They go for around 900USD for a 4 port solution (example: www.icron.com/products/usb_new/usb20-ranger-2224-fiber-extender.php) plus the cost of the optic fiber patch cord.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:09:14 GMT
Javier It still won't overcome the problems that USB has due to the noisy SMPS +5V Vbus wire interacting with the D- and D+ leads before it is processed in the 1st box.This would still affect the differential receiver wherever it is located.However, if something like a SOtM PCI USB card was used to replace the noisy +5V SMPS with clean regulated +5V derived from the PC's +12V supply, then this should further improvre matters. Regards Alex P.S. Reply 58 from an industry insider, and experienced in listening evaluation of products before their release, may be of interest. www.computeraudiophile.com/f10-music-servers/optical-out-good-12759/index3.html#post170374
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:19:31 GMT
I guess for someone really OCD about power noise on a spending spree another $400 wouldn't be too much of a problem. $1,200 and power problem almost solved because there is still the receiver unit's PS to tweak, one wouldn't want to feed it with a cheap walwart SMPS, so another $300-400 would need to be added for a fully regulated and filtered linear supply with a HQ traffo in a nice case. What do you reckon? BTW, something I never have fully undestood is why power noise seems to only affect USB and not PCI or PCIe.
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 23:20:41 GMT
I agree that a full galvanic isolation with external DAC would give the best results.
But the (USB-)specific solution described in your post give only the best possible USB solution. And USB isn't exactly the best solution. But it would be quite versatile since a lot of DAC's use USB.
The fiber optic variant of Thunderbolt could be interesting from that viewpoint. But so far, I've seen no fiber variant. But at least, it was specified for Thunderbolt.
Optical SPDIF has only a very low quality of the cables and the cable length has rather tight limits. No comparison to fiber optics for 1000BaseSX Ethernet. But the protocols for Ethernet just aren't suited for low-latency audio, althought the media itself (850nm multi-mode fiber optic cables) would be great (and very costly).
I've had another idea this evening: A steal case which is divided in two sections (separated by steel walls). One for the audiophile computer and one for the DAC. Both sections fed by their own toroid power source (internal or external). A kind of fully galvanic isolated optical connection as link between them. Probably the easiest way here would be still an optical SPDIF connection. On the computer side, it would be only possible to use a mainboard/CPU which has not big power demands and can be fed with (for example) a 12V input. At least, this solution would give "one" device and cable mess could be somewhat reduced. Not really a new idea (I know) but it should be easy to use in pratice. Personally, I'd prefer if the PSU's would be inside both sections, although external PSU's could be even more beneficial.
Personally, I'd go probably for a low-power mainboard/CPU which is fed by a toroid PSU. I could live with a simple Xonar expansion card as DAC. Probably not the best solution, but cost effective and most problems with the usual PC SMPS PSU's could be avoided by using a toroid as power source. Internal peripherals like SSD, HD, optical drives and some Xonar cards which require an additional 4-pin Molex or SATA power input could be a problem though. The voltage transformer of mainboards with 12V DC input usually do reside on the PCB of the mainboard. Those aren't really of a high quality respectively they are not very efficient. Usually, there's also no possibility to connect 4-pin Molex or SATA power connectors directly to the mainboard. These connectors need 12V and 5V DC.
Perhaps "normal" mainboards with one of the tiny converter boards for 12V DC input would be more pratical. But these converter boards are usually a plain rip-off. It's easy to pay as much for the converter board as for a fully-fledged high-quality ATX PSU.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:23:52 GMT
And USB isn't exactly the best solution What do you think is wrong with async 2 USB?
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 23:28:56 GMT
I find this SOtM PCI USB card very questionable. The price is a plain impertinence, even if it would give benefits. If we look on the PCB of that SOtM PCI USB card, I see nothing special. A standard NEC USB chip and a few more caps than usual. Cheap as chips to mass-produce.
And no one on RG has tried it. I won't. It's a rip-off, that is.
I find it somewhat questionable to praise that card only based on a C.A. article in which those idiots have done almost everything wrong. That cheap external SMPS is a pure joke for such a costly rig. Most money was invested for the very costly OrigenAE case, the SOtM PCI USB card and the SOtM SATA filter.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:33:07 GMT
It probably does, but there isn't normally a great deal we can do about it unless we use Linear PSUs and suitable Motherboards. Mr.USB ( Gordon Rankin) also advocates using wherever possible, different interfaces to avoid interaction. For example, only one USB device at a time. I would like to go back to a previous keyboard but they all seem to be USB these days, or wireless via USB. Alex
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 23:37:29 GMT
I have no problem with it. But it's certainly not the best possible bus for audio. FireWire is more advanced for such a purpose. Even FireWire 400. The availability of USB DAC's is much better though. It's used much more often and therefore more versatile.
I also tend to prefer a real optical interface instead of putting some copper-to-optical converters in between. But SPDIF has certainly also it's limitations.
It's probably save to say that all these interfaces haven't been designed with audio purists in mind. All of them will do for most people though.
There is certainly a point where I think I have to stop tossing out money for only small (if any) improvements. Personally, I can live well with the output of the Xonars cards. Perhaps not the best, but clearly above the average audio solution.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:41:24 GMT
Mr.USB ( Gordon Rankin) also advocates using wherever possible, different interfaces to avoid interaction. For example, only one USB device at a time. That woulnd't be much of a problem with current motherboards sporting USB 3.0 ports, they use a completely different chip for these so you can plug the DAC in it and use the Intel chipset one for the rest of your USB devices. Both my motherboards (Asus P8H67-M Pro and P7P55D-E Pro) have independent USB 3.0 ICs.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:49:09 GMT
Firstly, almost everything done wrong is your personal opinion. As for being a rip off price wise, perhaps if it was mass produced it wouldn't be ? It takes the PC's +12V supply and regulates it down to a very clean +5V. The card also permits the +5V supply over the USB cable to be switched off where the external device can take advantage of this.Quite a few C.A. members have also used the SOtM products in other areas and posted their findings. For the vast majority of people,using these cards would be an easy, but not cheap, alternative to external DIY Linear PSUs. C.A. has >130,000 members. Quite a few of these members are involved in the industry and many are designers/manufacturers of highly regarded commercial products. Demian Martin , as an example is very highly regarded in this area, as is Gordon Rankin and many others.
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 23:49:14 GMT
That's a point where I do absolutely agree with you. Instead of putting a lot of filters and snake oil between the components, I'd have a look on the power source itself.
Such a solution is only possible for a limited choice of mainboards, but it's possible. But so far, I've seen no one that really did it. The guy from the C.A. "audiophile PC" article certainly has missed the most important point.
I'd like to build such a PC somewhen. Just for seeing how big the difference could be. It's wouldn't toss out money for the SOtM rip-off components though. Unfortunately, I have still no job and this project has a very low priority at the moment.
I'll probably buy a mainboard, CPU and RAM for a small system this autumn. Most likely one of the upcoming IvyBridge i3 or AMD Trinity desktop solutions. Main goal here is to go cheap and it will be a general purpose computer. I'll do it only since I have most other needed components already here. Otherwise, I'd buy nothing at the moment. It gives me the chance to "get rid" of a bunch of components which are unused at the moment (so they see some use) and I have the time to complete it easily.
To get back on track: I guess Dave's questions is answered. No problem to go for a SSD, but benefits for audio quality should be minimal (if any). The improved speed will make the investment worthwhile though.
@alex, Javier: What do you think guys? Shall we start a project for a nice little audio PC somewhen? It would be fun to make one. But we all have not much money to toss out at the moment. I guess, it will have to wait.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:54:11 GMT
Firewire, for all intended purpose, is as dead as the last tyranosaurus in the PC world. Some Apple gear owners still use it but I don't think it will last much longer even in the Mac world with Thunderbolt pushing it away, Mr. Cook wanting badly their cash.
From a technical POV, I2S from Async USB is a much, much better solution than any SPDIF type. No muxing required and the whole chain can be enslaved to a single clock very close to the DAC IC in synchronous mode (2 clocks actually, one for each fundamental freq family). Of course, good design and implementation is critical for best performance in either technology.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 29, 2012 23:56:05 GMT
Mr.USB ( Gordon Rankin) also advocates using wherever possible, different interfaces to avoid interaction. For example, only one USB device at a time. That woulnd't be much of a problem with current motherboards sporting USB 3.0 ports, they use a completely different chip for these so you can plug the DAC in it and use the Intel chipset one for the rest of your USB devices. Both my motherboards (Asus P8H67-M Pro and P7P55D-E Pro) have independent USB 3.0 ICs. Javier Agreed. However many people still use PCs from a few years before USB 3.0 became common place.I do however have a spare PCI-E slot that could be used for a USB 3.0 card. Presently, I don't have any USB 3.0 devices, although I am contemplating purchasing a USB 3.0 Corsair Voyager ( $35 for 32GB) which is backwards compatible with USB 2.0 , but not as fast with USB 2.0. Regards Alex
|
|
elysion
Been here a while!
Team Anti M$ AND Facebook.
contra torrentem
Posts: 2,375
|
Post by elysion on Jul 29, 2012 23:58:18 GMT
Firstly, almost everything done wrong is your personal opinion. As for being a rip off price wise, perhaps if it was mass produced it wouldn't be ? It takes the PC's +12V supply and regulates it down to a very clean +5V. The card also permits the +5V supply over the USB cable to be switched off where the external device can take advantage of this.Quite a few C.A. members have also used the SOtM products in other areas and posted their findings. For the vast majority of people,using these cards would be an easy, but not cheap, alternative to external DIY Linear PSUs. C.A. has >130,000 members. Quite a few of these members are involved in the industry and many are designers/manufacturers of highly regarded commercial products. Demian Martin , as an example is very highly regarded in this area, as is Gordon Rankin and many others. If the number of registered members would be a real indicator of knowledge, you should have a look at Facebook (plus 900 million users). If every tenth of their member would buy a SOtM card, it would be easily to go for a mass product. The card also doesn't look like a small batch. You haven't even tried the card. And your are judging it. According to your description, it's used to overcome some USB power issues which are related to noise PSU. But instead of using a cleaner (linear or toroid) power source, they use a questionable overprice USB PCI card which costs more than a handmade Little Pinkie from Mike. Before I'd toss out money for the SOtM stuff, I'd order an appropriate LittlePinkie for that computer from Mike. Instead of treating only the symptoms, it would cure the origin of the problem.
|
|