LogicprObe Posted October 30, 2012 Posted October 30, 2012 Loading the whole track into memory is different to running a buffer. A buffer always has the chance of running out if the computer decides it has to do something more urgent.............like a Win Update or virus scan, etc.
Nada Posted October 30, 2012 Posted October 30, 2012 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. The only thing that matters to me is how the music sounds. If we limit our experimenting to what we think ought to work we are immediately stuck. Usually theory follows observations. So a spinning disc will have a variable current draw that will deregulate the power supply voltage noise level and that is enough to add noise into the audio system. A solid drive will be a benign low noise load. 1
drez Posted October 30, 2012 Posted October 30, 2012 The limiting factor in computer transports setups is not the transport IMO, most people who are finding differences with computer settings and hardware are using high end transports such as Audiophilleo, XMOS etc already with pretty much all of them using galvanic isolation, asynchronous transfer, hardware buffers etc. The latency/timing of the data being fed into the transport still seems to matter. Most of the computer audiophile stuff is fairly easy and cheap now anyway, and makes common sense even if you weren't interested in sound quality, there are some dodgy components out there though, so I think it's important to keep a level head. That is an interesting thought with SSD vs HDD noise and electrical load - there is quite a substantial difference here but unfortunately storing music on SSD is still way too expensive at the moment. Maybe providing external power for the HDD would be easier.
LogicprObe Posted October 30, 2012 Posted October 30, 2012 I can get all my favourite tracks in at under 64GB @ 44.1 wav.
drez Posted October 30, 2012 Posted October 30, 2012 I can get all my favourite tracks in at under 64GB @ 44.1 wav. Hmm I do have a spare SSD from my old install, but I wanted to keep that as a fallback. I guess SSD's will fall in price over time so in a matter of time one could fit a lot of music on solid state storage. I guess you would want to prove the concept before investing in a large SSD anyway, I might give it a try in the future but I am feeling lazy to change my computer setup at the moment.
LogicprObe Posted October 30, 2012 Posted October 30, 2012 I don't have any problem with spinners...........except when they try to be green and go to sleep!
Soundscape Posted October 30, 2012 Posted October 30, 2012 I can get all my favourite tracks in at under 64GB @ 44.1 wav. Do you have any full albums in that collection?
LogicprObe Posted October 30, 2012 Posted October 30, 2012 Do you have any full albums in that collection? Not that many. I just reripped everything I have.............the total so far is 402GB. There's about 150 more discs that failed accuraterip because of scratches which I'm going to attempt to repair with my new CD repair machine! (review to come) So......the total will be about 500GB by the end. That said............I can really cull my favourites down to 15GB to fit on my iphone. It's still a lot of songs and I don't get sick of them.
hochopeper Posted October 30, 2012 Posted October 30, 2012 The limiting factor in computer transports setups is not the transport IMO, most people who are finding differences with computer settings and hardware are using high end transports such as Audiophilleo, XMOS etc already with pretty much all of them using galvanic isolation, asynchronous transfer, hardware buffers etc. The latency/timing of the data being fed into the transport still seems to matter. Most of the computer audiophile stuff is fairly easy and cheap now anyway, and makes common sense even if you weren't interested in sound quality, there are some dodgy components out there though, so I think it's important to keep a level head. That is an interesting thought with SSD vs HDD noise and electrical load - there is quite a substantial difference here but unfortunately storing music on SSD is still way too expensive at the moment. Maybe providing external power for the HDD would be easier. With respect, it is my opinion that most of these computer transports are still flawed in some ways technically, I plan to test these theories. It is my opinion that if the transport is truly immune to jitter, phase-noise and bit-perfect, then the computer settings will have no impact on the sound. Much as Nada suggested above. By that theory, the current transports that show improvements as a result of these tweaks are clearly flawed in some way. So you can either minimise the impact of the computer for the flawed device, or find solutions that are technically better. For most the second option is a far more arduous and expensive option, so it isn't often pursued. For me, it is a hobby to further my understanding of the technical aspects, so I pursue the other approach.
drez Posted October 30, 2012 Posted October 30, 2012 No disrespect intended on my part either, I don't doubt that one can always build a better transport than some of the commercially available USB transports, just I am sceptical that there can ever be a USB transport or digital input design that is truly 100% jitter immune, maybe excluding a few high end DSP's I haven't had the benefit of hearing or affording so far. Again I don't think setting up a computer transport involves anything particularly unusual compared to a typical workstation - nor that you need any boutique audiophile computer parts (in fact I have had bad experiences with these.)
LogicprObe Posted October 30, 2012 Posted October 30, 2012 It's always puzzled me, that given the prices of a lot of highend digital gear, that some audiophiles haven't gravitated to Pro ADAC's. You know............the same ones that the music was probably recorded with!?
henry218 Posted October 30, 2012 Posted October 30, 2012 there are few audio interface that can perform very well, and its useful for actuve speakers. motu828mk2 is popular in pro world for high performance with less $$$ top of the game would be Apogee Symphony, having said that.. there are few pro brand entering hi end market such as Lynx, Mytek, Berkeley.
gavtron Posted October 30, 2012 Posted October 30, 2012 It's always puzzled me, that given the prices of a lot of highend digital gear, that some audiophiles haven't gravitated to Pro ADAC's. You know............the same ones that the music was probably recorded with!? Maybe as people figure they are paying extra for the recording features which they wont use? 1
drez Posted October 31, 2012 Posted October 31, 2012 The only thing that matters to me is how the music sounds. If we limit our experimenting to what we think ought to work we are immediately stuck. Usually theory follows observations. So a spinning disc will have a variable current draw that will deregulate the power supply voltage noise level and that is enough to add noise into the audio system. A solid drive will be a benign low noise load. This seems to match what I am hearing - files read from the SSD were better (tighter timing and seperation) than files read from the 5K HDD which in turn were better than the files read from the 7K. There seemed to be a strong correlation between power usage to read the files and the timing accuracy of the audio playback even when using JPLay which buffers the whole playlist to ram before commencing playback. Strangely I noticed something else - there seemed to be more grit to the sound coming from the 5K HDD, comparatively the SSD sounded much cleaner and smoothed. I almost preferred the timber from the 5K HDD - which is more accurate is hard to tell. Is there some magical distinction between overall timing and the timing that produces textural details, or is the texture of 5K HDD playback artificially enhanced? I can't decide for sure.
gavtron Posted October 31, 2012 Posted October 31, 2012 there are so many variables in the connections, platters, electronics, build of the drives between maker, vibrations - depending on how the drives are mounted etc... have you tried formatting different drives or partitions on the same drives with different file systems - like fat, fat32, ntfs, ext etc to see if that has any effect...?
LogicprObe Posted October 31, 2012 Posted October 31, 2012 Maybe as people figure they are paying extra for the recording features which they wont use? You could digitalise your vinyl?
TP1 Posted October 31, 2012 Posted October 31, 2012 It's always puzzled me, that given the prices of a lot of highend digital gear, that some audiophiles haven't gravitated to Pro ADAC's. You know............the same ones that the music was probably recorded with!? In some cases they do. Weiss for example have a consumer DAC version of their pro AD/DA converter which is very well received. I can't speak for all audio tragics but I would think that some would want a purpose built plug and play DAC solution without having to go through menus, cable adapters etc. I've been playing around with an Apogee Duet 2 for A/D purposes but I don't think its particularly convenient to use as a DAC.
LogicprObe Posted October 31, 2012 Posted October 31, 2012 In some cases they do. Weiss for example have a consumer DAC version of their pro AD/DA converter which is very well received. I can't speak for all audio tragics but I would think that some would want a purpose built plug and play DAC solution without having to go through menus, cable adapters etc. I've been playing around with an Apogee Duet 2 for A/D purposes but I don't think its particularly convenient to use as a DAC. Ahhh......yes. A good explanation. Having only ever used ADAC's, it has blinded me to ease of use.
drez Posted October 31, 2012 Posted October 31, 2012 (edited) there are so many variables in the connections, platters, electronics, build of the drives between maker, vibrations - depending on how the drives are mounted etc... have you tried formatting different drives or partitions on the same drives with different file systems - like fat, fat32, ntfs, ext etc to see if that has any effect...? Nope I havent got into that stuff yet Just comparing drives with NTFS probably using MBR for the 5k HDD and GPT for the SSD (not sure as I partitioned those drives a long time ago). Vibration management might be an interesting thing to try out - I can't really do much in terms of mounting in my current setup except for vertical or horizontal mounting and mass loading the drives. ideally if SSD's can give optimal performance then this would make life much easier as they are not actually creating vibrations themselves. There is also of course the option of externally powering HDD's but I'm not sure how practical that is at the moment until I get some more GPO's installed, and even then I would probably have to leave the drives on 24/7 which might waste a bit of electricity over time. IMO if I can get a intel 520 SSD with large enough capacity and hook it up to SATA 1 controller I can't see how this could be improved upon fro a technical perspective. Edited October 31, 2012 by drez
Tamir0678 Posted November 1, 2012 Posted November 1, 2012 (edited) Hi everyone, I'm new to this forum, so I don't know what equipment each of you use. There have been a lot of interested ideas mentioned in this thread and I thought I'll add my 2 cents. I'm running an HTPC (PCM streamed over HDMI @ 24bit/192khz) to an Onkyo TX-NR3009 receiver to an Elektra Theatron power amp to Quad 22L/12L/L-Cenre speakers and a Def Tec Supercube subwoofer. Although not audiophile by any standards, this system sounds very good. I've also been building (and used to sell) PCs for the past 17 years. I have an engineering background, so I'm very sceptical towards "high end" tweaks that are not supported by science. I have been an avid music lover for the past 20 years and movies in surround for the past 10. My view on using HTPC as an audio source: 1) The hardware should be adequate to run whatever you throw at it without latency ever being an issue. If it is not, this is probably one of the cheapest improvements to make to your system. Therefore, files should be encoded in lossless formats to save space. 2) The PSU is probably the most important and most underrated piece of hardware in an HTPC. Use a good one. 3) It's better to use good external DACs and let the PC bit-stream in PCM. I prefer HDMI because it serves both audio and video. 4) Cables for digital data make no difference for short runs. It is very difficult to turn 0 into 1 and vice versa with external interference, and if there is such interference, the entire signal would get trashed - it won't be a minor sq issue. 5) Stable voltage does matter and can impact fidelity, but does not create duplicable errors - which can make subjective comparisons sometimes unreliable. 6) If using more than 1 drive, the OS and media playing software should be on the better (faster, more stable) drive (preferably SSD). If you really want error free reading from your drive(s), set them up in raid. 7) Use Win7 over Vista, as it consumes far less resources (memory, power). Use the 64 bit version if you have more than 4gb of ram (the 32 bit version cannot efficiently use more). 8) Processes not used should be disabled (I use the built in msconfig). 9) Some thoughts on latency, read times and transfer rate: - SATA 3.0 has a transfer rate of 6 Gbit/sec, USB 3.0 has a transfer rate of 3.2Gbit/sec. - For comparison, the data bandwidth required to transfer 6 channels @ 24bit/192khz = 6x24x192000= 27648000 or 27.6Mbit/sec, so about 1/100 to 1/200 of the available bandwidth - A slow 5K HD (e.g. WD Green) has a read time of 800Mbit/sec. A fast SSD (e.g. Intel 520) has a read time of 4.4Gbit/sec. (which BTW means that If you are using SSD over USB you are reducing the transfer rate by 27%). - For comparison, an almost optimal high definition video signal (1080p @ 32bit/60hz, 32bit because very little graphics cards and content support a higher colour depth) requires 1920x1080x32x60= 3981312000 or almost 4Gbit/sec - 144x the bandwidth requirement for lossless multi-channel audio. So, again, full lossless audio isn't even scratching the bandwidth available to it in a decent PC. Here are some more thoughts: Just swap my hdd in my laptop with an 120gb intel 520 hdd. I am absolutely and totally blown away by the improvement across the board. Immediacy, impact, detail and darker background...more stable soundstage.....but the increased immediacy and impact......wow.... When you think about it, the hdd can introduce interference from the spinning motion and increased latency from being slower to read off. That is a great SSD. IMHO it's the only one to get. Best Value/Performance/Reliability (This is what I use in my HTPC as well, chosen after much research). I'm not sure how the spinning motion would create interference with a digital signal? Is there a lot of processing done on an internal sound card in your configuration? I haven't had any sound/playback issues resulting from latency in any PC I've used in the last 5 years. Hey do you think a fanless desk top performance will yield even better performance. Surely a spinning fan in a laptop will introduce its own magnetic and vibrational interference....compared to a completely silent desktop with huge heat sinks? Is there such a solution as a fanless desktop PC? Yes you can build a fanless HTPC, but I don't see how any fan would create enough magnetic interference to impact a digital signal. Vibration is only relevant to the ambient noise the PC creates - and that's the true benefit of a fanless PC - minimum ambient noise (and with a good case, power supply and SSDs no audible noise at all). Ahhh...I just Remembered the reason I chose not to go down the desktop route....it's the need to plug power to the wall. To me, stable power supply will impact the sound more than fans running on a laptop. Unless you could power a desktop with a battery. One of the strength is the ability to go off th grid and I can hear the difference when the power supply adaptor is switched off. Why not just get a good PSU? A battery won't provide perfect DC either. I suppose in the case of using a pc as a transport, the power supply is not as critical as say in a dac or an amp, latency is the bigger issue??? I think the contrary is true. The PSU impacts everything (including latency, which is clock dependant, which in turn is voltage dependant) The OP referred to a certain group of guys and jplay and I have now read all 24 pages there relating to this issue. My head is spinning! But in summary I interpret the recommendations as follows: Build a PC for music only Use windows 7 or 8 Operating system on a USB stick or on an SSD drive. It appears that the best results are obtained by having this SSD drive externally and connecting via usb to the pc with a USB/optical cable or via an "Adnaco USB hub" the external drive and or USB hub shuold have a good quality power supply The music can be on a normal hard drive or on another SSD or on a NAS Even better results can be obtained by using software called Total Commander. It appears that you lose a lot of functionality, but have better music. (There are variations to this where they use special SATA cables to connect drives to the motherboards etc.) According to them the operating system (and not necessarily the music) should ideally be on media externally and isolated from the computer. This makes sense because I have a 1 TB of music which I cant afford to put on SSD drives. I do find this topic extremely interesting. More views? I can't see the benefit of having the OS externally when using SSD. Theoretically, an internal HDD could be effected by vibrations, but my intuition tells me this probably isn't the case. Separate high quality power supply to the OS drive could have merit though. Edited November 1, 2012 by Tamir0678 3
Tamir0678 Posted November 1, 2012 Posted November 1, 2012 Continued: I believe connecting a ssd via USB sort of defeats the purpose of having an ssd in the first place. My understanding USB connection is no where as fast as say a sata connection. I was just really quite shocked by how much improvement I got....now I am greedy. The beauty of pc audio is everything is scalable and modifications are very modestly priced compared to hifi component upgrades. It's a market size thing at the end of the day. Agree. And I love how quickly programs load too... Prices of SSD are going to have drop more before I purchase 1TB+ for my music files! Has anyone compared whether SSD makes a difference when the files are read into memory and then fed to the DAC from memory? Greg I think the benefits from SSD for SQ are: 1) Faster, so has more redundant speed to avoid bottlenecks 2) Lower power consumption, so makes life for the PSU easier, which could reduce voltage related errors 3) More stable current draw, which again could reduce voltage related errors 4) Quiet operation, which reduces background noise So any errors introduced by the drive will remain whether pre-buffering to the RAM or not, unless the software also performs parity checks or has another ECC code. It's all about taking turns at the CPU. That is why loading up the USB is folly. For audio, I would recommend one device per chip. Good point! Hi Greg, In my experience the storage medium plays a part on the sq of memory based playback. I am unsure why this is the case - however I can hear subtle differences between ssd's, memory sticks, hdd's, loading from NAS, loading from USB hdds. The cause of this is beyond me as it is well outside any understanding I have of how computers work, but I have to trust what my ears tell me. This is very interesting. Did you notice exactly the same differences in detail and dynamics every time? Can you describe those differences? Differences mean errors are introduced, but with a digital signal , errors should be nearly impossible to replicate with normal use. I wonder if it's error averaging that's creating different sound characteristics? It would be extremely interesting to run this as a blind test. Hi JVenter, Mostly the JPlay consensus seems to be that the least amount of activity happening in or around the PC while either loading the file to ram or when a file is playing back is beneficial to sq. This is why people suggest tweaks such as disabling the video card in device manager (obviously best used in a headless media server!) or removing the windows explorer shell. Also perhaps why they recommend loading the files from a NAS as there is less chance of interference (RFI and such) being generated in the PC. I have just gathered this from reading their posts - but have no real evidence as to why this is - but I have lots of time to experiment and am slowly finding what works best for me. Some of the changes are very subtle, some of them, like shutting down extra processes don't seem to do much at all I have noticed on my setup that when I used a usb wifi adapter that there was a tiny bit of noise in the mid range that was not there when I unplugged the adapter - or put it into the adnaco hub. So I think certain usb devices can insert subtle noise into the system, but I'm not an expert on this and don't understand all the theory I have read on the subject. I think logic probe may be referring to the fact that motherboards have USB hubs built in and even though your computer may have 4 ports on the back they are just connected to the one usb controller. (please correct if I am wrong, lp Less processes should be beneficial to overall performance as it equals less competition for resources (CPU, RAM, Power), and less power consumption spikes. Does your PC connect to your system with analogue outputs? Most high quality boards would have 2-4 USB controllers. Each would commonly be a hub for 2-4 ports. 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. True. However there are some differences which I've mentioned above. Nope I havent got into that stuff yet Just comparing drives with NTFS probably using MBR for the 5k HDD and GPT for the SSD (not sure as I partitioned those drives a long time ago). Vibration management might be an interesting thing to try out - I can't really do much in terms of mounting in my current setup except for vertical or horizontal mounting and mass loading the drives. ideally if SSD's can give optimal performance then this would make life much easier as they are not actually creating vibrations themselves. There is also of course the option of externally powering HDD's but I'm not sure how practical that is at the moment until I get some more GPO's installed, and even then I would probably have to leave the drives on 24/7 which might waste a bit of electricity over time. IMO if I can get a intel 520 SSD with large enough capacity and hook it up to SATA 1 controller I can't see how this could be improved upon fro a technical perspective. GPT should be used if your drive is over 2.2TB. That's what it was created for and should not provide any other benefits pver MBR. MBR has the benefit of being supported by more software I don't think vibrations make a difference, but would love to see hard data that shows otherwise. You could use a master/slave power outlet to control your SSD external power supply, or Instead maybe just get the best power supply you can. 2
TP1 Posted November 1, 2012 Posted November 1, 2012 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.co.uk/linear_pc_psu.html
Tamir0678 Posted November 1, 2012 Posted November 1, 2012 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 That's not completely accurate. When the output is through HDMI on the graphics card, the PSU voltage is DC-DC converted by the card using an onboard VRM (Voltage regulation module), which reduces voltage fluctuations. Good internal PSUs can produce less than 0.15% voltage fluctuation on a 12v rail, which is ~18mv. If this produces more errors than a 1mv fluctuation on a digital 400-600mv signal (sata), I don't know, but I would love to see test result measurements for this.
gavtron Posted November 1, 2012 Posted November 1, 2012 This is very interesting. Did you notice exactly the same differences in detail and dynamics every time? Can you describe those differences? Differences mean errors are introduced, but with a digital signal , errors should be nearly impossible to replicate with normal use. I wonder if it's error averaging that's creating different sound characteristics? It would be extremely interesting to run this as a blind test. The differences are quite subtle - I will try and find a friend who is happy to run a blind test and report back my results! Maybe best described as a lack of depth in the sound, with some harshness in the mid range. But the reason why I hear this is well beyond my comprehension of how PC's work. So I just have to trust my ears Does your PC connect to your system with analogue outputs? Most high quality boards would have 2-4 USB controllers. Each would commonly be a hub for 2-4 ports. 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.
TP1 Posted November 1, 2012 Posted November 1, 2012 That's not completely accurate. When the output is through HDMI on the graphics card, the PSU voltage is DC-DC converted by the card using an onboard VRM (Voltage regulation module), which reduces voltage fluctuations. Good internal PSUs can produce less than 0.15% voltage fluctuation on a 12v rail, which is ~18mv. If this produces more errors than a 1mv fluctuation on a digital 400-600mv signal (sata), I don't know, but I would love to see test result measurements for this. It's not about voltage accuracy it is about an impure electrical signal that in itself can degrade sound.
Recommended Posts