Mistertaz Posted November 22, 2015 Posted November 22, 2015 Hi all! Just bought a Mac Mini and am considering some minor mods to optimise its performance as a music server. Don't want to spend exorbitant amounts of cash on it, but some suggestions of affordable tweaks would be great. The recommendations I've read are to install an SSD, increase the RAM to 8GB and change to a linear power supply. I'll consider the SSD and RAM upgrade, but the linear supply seems a little excessive due to cost. I'm tossing up between the Uptone Audio USB Regen which should filter out any power supply noise anyway. Or the Uptone Audio MMK kit to convert the Mac Mini to 12V DC and using an off board 12V SMPS instead of a linear supply. What are the thoughts on these? Thanks! Taz.
Equaliser Posted November 22, 2015 Posted November 22, 2015 For solely just a headless music server, 8GB is overkill and you will easily get away with just 4GB, but RAM prices haven't really dropped much since the disaster in Taiwan 2 years ago and hasn't really increased so why not you can upgrade to 8GB in case you expand to other things. An SSD improvement may be noticeable due to it using less power, less thermal heat output (no moving parts inside) and faster read/write speed then conventional mechanical hard drives. Both RAM and SSD are seen as upgrades not modifications as they will help enhance the longevity of the Mac mini computation and storage wise (not sound improvements). A linear supply will be the biggest improvement and depends on how well the linear power supply is designed to be.
DoggieHowser Posted November 22, 2015 Posted November 22, 2015 I would disagree on the amount of RAM. 8GB is the bare minimum especially if you use: a. hires b. gapless playback (e.g. full symphonies) The SSD minimises physical and RF noise by not having magnets and moving parts, But the linear PSU really makes a big difference. You can get a kit and ask Clay to make you a reasonably well sized PSU.
Mistertaz Posted November 22, 2015 Author Posted November 22, 2015 It seems the linear supply is pretty unanimously considered a big improvement... Any comments on whether the difference is more audible when the DAC is connected by USB vs Optical? Optical provides isolation from power supply noise but adds jitter right? And which software is considered the best for use as a headless server with an iPhone or iPad as the interface? Taz.
Jventer Posted November 23, 2015 Posted November 23, 2015 Taz I am slowly trying things out in my own systems. I am realising that some things that guys talk about on the forums I can hear and some not. Some of them, even when I can hear it, I don't think it is value for money in my personal circumstances. Almost everything matters in Audio, but it has a lot to do with one's system. Best is to try it out in your system to see what it does and if you like what it does. Regarding software - How much do you wish to spend? I would suggest the mainstream software packages for mac are, Pure Music, Amarra, Audirvana and now JRiver. Which one is best? I am not going to go there. I would download trial versions of each and decide myself. Enjoy! 1
Equaliser Posted November 23, 2015 Posted November 23, 2015 (edited) I would disagree on the amount of RAM. 8GB is the bare minimum especially if you use: a. hires b. gapless playback (e.g. full symphonies) Wrong and it's a common misconception been discussed hundreds of times. I have a 4GB playback system dedicated for only hi-res 24 bit and gapless playback with a good cpu. It's plenty and more so affected by processing power. No music application on its own regardless of what resolution music playback you're playing from does not utilize more then 4GB. The only software application that uses more then 4GB is Photoshop and lightroom rendering applications or running multiple software threads/processes in the background. The SSD minimises RF noise. Got a source to back this up? Edited November 23, 2015 by DefQon
DoggieHowser Posted November 23, 2015 Posted November 23, 2015 I configured an older machine with 4GB of RAM and it was noticeably challenged when I played hires files. Upgrading to 8GB fixed that. To put it in context, a 24 bit 192kHz file is 6 times larger than CD quality. A full length symphony or concert would take up 4.2GB and that's before we factor in OS overheads. And with quad DSD plus DXD files, it just seems a no brainer to go with a minimum of 8GB. And as for moving parts - I went with a SSD, fanless PSU and fanless CPU heatsink PC - noticeably better as a music server than using one with fans/HDDs.
Jventer Posted November 23, 2015 Posted November 23, 2015 My personal opnion is that I put as much as possible ram in every computer I own. After that the biggest bottleneck or issue is the hard disk, so SSD looks like the best option. What not many people are talking about is that really good results can be obtained by using Ram disks and SSD which has very little $ costs.
Mr Goldmund Posted November 23, 2015 Posted November 23, 2015 I configured an older machine with 4GB of RAM and it was noticeably challenged when I played hires files. Upgrading to 8GB fixed that. To put it in context, a 24 bit 192kHz file is 6 times larger than CD quality. A full length symphony or concert would take up 4.2GB and that's before we factor in OS overheads. And with quad DSD plus DXD files, it just seems a no brainer to go with a minimum of 8GB. And as for moving parts - I went with a SSD, fanless PSU and fanless CPU heatsink PC - noticeably better as a music server than using one with fans/HDDs. I had the same experience. My system was quite challenged if it's on for more than 2 days at a time, practically everything would lag. Putting in 8 gigs of ram fixed this, now after being on for a week everything still runs smoothly. I found that Jriver uses more memory than foobar. Unsure about the sound quality improvements on my setup but the speed improvement was definitely worth it. 1
Equaliser Posted November 23, 2015 Posted November 23, 2015 My personal opnion is that I put as much as possible ram in every computer I own. After that the biggest bottleneck or issue is the hard disk, so SSD looks like the best option. What not many people are talking about is that really good results can be obtained by using Ram disks and SSD which has very little $ costs. Except if you're talking the same Ram disk as the one I'm thinking of that you can configure in any windows OS, after a restart the whole RAM disk needs to be reconfigured. The only ones I'm aware that give you fast speeds and caching is volatile and the content is deleted once you restart or shutdown.
Equaliser Posted November 24, 2015 Posted November 24, 2015 (edited) I configured an older machine with 4GB of RAM and it was noticeably challenged when I played hires files. Upgrading to 8GB fixed that. To put it in context, a 24 bit 192kHz file is 6 times larger than CD quality. A full length symphony or concert would take up 4.2GB and that's before we factor in OS overheads. And with quad DSD plus DXD files, it just seems a no brainer to go with a minimum of 8GB. And as for moving parts - I went with a SSD, fanless PSU and fanless CPU heatsink PC - noticeably better as a music server than using one with fans/HDDs. You're going to have to be more specific here. If the older machine has a weak processor than putting more RAM may offer a slight improvement to overall responsiveness, why? Because RAM is nothing but a large volatile memory space for caching read/write processes that are temporarily stored when it is dumped after the application has been processed through the stack of the processor. The processor also has in-built RAM (descriptive) but operate at much faster m/s and these are micro registers, L1, L2 or L3 cache of the CPU in that order fastest to slowest (and RAM is considerably slower then the mentioned 3). RAM assists and works in line with the processor as to how much information it should store about current threads and processes. Old processors from few years ago can't compare to even dual core cpu's of today because the architecture is completely different and vastly improved on, a good processor determines how fast the process is handled, how many processes and threads it can handle concurrently and how fast the process is read/write between internal cache and external memory. Regardless of how big a 24Bit Hi Res file is compared to a regular CD quality 44khz 16Bit FLAC rip it all depends on the algorithm/D/A process the Hi res file is ripped/converted to and from and how well it is read and processed by the music player software. If by what you're implying the bigger the file the more application memory space it consumes then this becomes completely false as any old Pentium 2 system will even process a 1GB MP3 file with ease (and I have a Pentium 2/3 system here that can do just that). It all depends on what is done to the file and it's format rather than how big it is. As I said in my previous post, I have a remotely managed, dedicated headless music system that I built a long time ago has a now not so modern Intel i5 quad core cpu but a good cpu nonetheless with 4GB of old Mushkin DDR3 1600mhz ram on a stripped down version of Windows 7 Pro with various music players and it handles any music playback I throw at it without a hitch be it SACD plugin playback in foobar or 800MB Hi Res FLAC rips, no hitch whatsoever. I found that Jriver uses more memory than foobar. As a programmer/software coder I find Jriver to be a relatively bad coded application with lots of internal overhead and uses more memory then it should for music playback. Although the quality of playback is quite good compared to foobar, foobar is a very nice minimalistic simple music player that I recommend and should be used a baseline for people listening on digital playback systems before spending money on other players such as Amarra, Jriver, Audivirna etc. Now the final question, is it worth upgrading to 8GB for a dedicated music player (nothing more)? Yes and no considering prices aren't as cheap as before but have stabilised at a certain point. Is it recommended to upgrade to 8GB or more for dedicated music playback? No, you can easily get away with 4GB if you're on a relatively modern machine, need proof and factual evidence? I have spare parts of all circa's over the past 2 decades so I can easily build one and show the usage and statistics. Edited November 24, 2015 by DefQon
Jventer Posted November 24, 2015 Posted November 24, 2015 Regarding Ramdisks: http://www.computeraudiophile.com/f27-uptone-audio-sponsored/attention-current-mac-mini-users-boot-mavericks-sd-card-load-ramdisk-dismount-your-internal-sata-drives-and-pour-drink-musicians-walking-out-your-speakers-18159/
Equaliser Posted November 24, 2015 Posted November 24, 2015 (edited) So he is doing something that has been around since the dawn of windows, ok here is what I had in mind (people usually configure it on a piece of volatile memory) https://www.softperfect.com/products/ramdisk/ I find a large percentage of the posts I read from time to time on the CAP's forum skeptical when it comes to computer audio playback because I never achieve the same result with the same thing (and audiophool non-sensary + computer audio = a mess to read). Such as night and day improvement on a system using a high quality power supply by swapping out a SATA hard drive to an SSD, there was one post online on reddit while back somebody took the liberty of emailing a hard drive manufacturer about sound improvements from HDD to SSD and they received a response from the companies technician laughing it off. Think the other funny one I read was somebody put there hard drive into a freezer, took it out and reported better music playback coming alive. Really eyebrow raising stuff. Edited November 24, 2015 by DefQon
DoggieHowser Posted November 24, 2015 Posted November 24, 2015 The RAM upgrade was on the same machine - as was the case with ethan. I replaced my Mac Mini 2010's HDD with an SSD and there was an improvement too. Thats why I opted for the factory SSD on the 2012 Mini I bought
Equaliser Posted November 24, 2015 Posted November 24, 2015 I replaced my Mac Mini 2010's HDD with an SSD and there was an improvement too. Thats why I opted for the factory SSD on the 2012 Mini I bought I've been in the computing field for a very long time, it and networking is my primary profession, audio is a just a side hobby from yesteryear, when you said you saw an improvement...what sort of improvement are we speaking here? I have tried over 30 brands and types of hard drives and SSD's for this sort and been told by some they hear an improvement but I have not been able to replicate the same result. I've definitely seen improvements in read/write speed for sure, lower thermal output/lower power usage, this leads to a quicker response with the OS boot time and applications I click on but I have not been able to hear an sound improvement. Now the old saying comes in, sometimes you need somebody to describe to you the difference for you to notice it. What sound improvements did you hear? Bits are bits and I find it hard to hearing a change. 1
DoggieHowser Posted November 24, 2015 Posted November 24, 2015 Just curious do you find all music playing software to sound the same if they are bit perfect?
Equaliser Posted November 24, 2015 Posted November 24, 2015 (edited) No I don`t and I`m talking about bits and bytes on hard drives not the source file itself. Edited November 24, 2015 by DefQon
DoggieHowser Posted November 24, 2015 Posted November 24, 2015 No I don`t and I`m talking about bits and bytes on hard drives not the source file itself. So you don't hear a difference between different versions of say Audirvana Plus or PureMusic - even when they are all outputting bit perfect signals?
Jventer Posted November 24, 2015 Posted November 24, 2015 (edited) I've been in the computing field for a very long time, it and networking is my primary profession, audio is a just a side hobby from yesteryear, when you said you saw an improvement...what sort of improvement are we speaking here? I have tried over 30 brands and types of hard drives and SSD's for this sort and been told by some they hear an improvement but I have not been able to replicate the same result. I've definitely seen improvements in read/write speed for sure, lower thermal output/lower power usage, this leads to a quicker response with the OS boot time and applications I click on but I have not been able to hear an sound improvement. Now the old saying comes in, sometimes you need somebody to describe to you the difference for you to notice it. What sound improvements did you hear? Bits are bits and I find it hard to hearing a change. All over the world people are experimenting with computers and posting things on forums. It is impossible to test what they are all saying, so you look for runs on the board. I can site a few examples: foobar, jplay, jriver, XXHighEnd, SOTM. As of today jplay has a loyal following whilst the jriver and other guys say it is a hoax. Some people swear by the latest and the best computers, others reckon you have to have a different motherboard, a slower type .Then you have Antipodes, Aurender etc. In my own limited tests I can sometimes hear differences and sometimes not. So we have a dilemma. Why dont we hear what others purportedly can? In the first instance, I trust the hearing of my wife. After 25 years of marriage and 4 years of her listening to hifi equipment I have not been able to prove her hearing wrong - point to note here, I am only referring to her ears and hearing ability . To date I have not been able to fool her with cables, interconnects even the Lenehan Toppers. Getting back to topic, the Ramdisk topic I referred to is the result of the experiments of a member on that forum named: Superdad. One could ignore him, but then he is the guy behind the Usb Regen as well as the Uptone powersupply. The mentioned products is getting some very good feedback and people are paying good money for it. Even on this forum @@Tasso who has IMO an impeccable standing is saying how good the Uptone is.. In other words, when a guy like Superdad, whose producs are selling worldwide, one of which I own and actually can hear the difference, say that ramdisks work - who am I to say it does not? If I cant hear something, good luck to those that can, and I want them to enjoy it as much as they can. Edited November 24, 2015 by Jventer 1
Jventer Posted November 24, 2015 Posted November 24, 2015 (edited) Post deleted Edited November 24, 2015 by Jventer
Equaliser Posted November 25, 2015 Posted November 25, 2015 So you don't hear a difference between different versions of say Audirvana Plus or PureMusic - even when they are all outputting bit perfect signals? I haven't used PureMusic so I can't comment on it. However I have been at times been able to detect very nuance differences of sound between say foobar and vlc player playing the same source file.
DoggieHowser Posted November 25, 2015 Posted November 25, 2015 I haven't used PureMusic so I can't comment on it. However I have been at times been able to detect very nuance differences of sound between say foobar and vlc player playing the same source file. Were the players configured in bit perfect mode? A number of DACs have a bitperfect test track so you can check if the app resamples. If bits are bits, then a bit perfect software should have no impact on the sound. The fact that it does suggests there is a lot more to this matter than just an averaged "CPU utilisation" figure. I remember one of the rewrites for ChannelD's PureMusic resulted in big gains in SQ as did AudirvanaPlus. In ChannelD's instance, the developer said he made changes to ensure there was no sudden spikes in loads. Most measurements record steady state or average end results but not the slight peaks and jumps. I remember the developer of Sabre32 made a similar observation - and it was in minimising the big dips/changes before steady state was achieved that they got their best "listening" results.
Andrew S. Posted November 25, 2015 Posted November 25, 2015 I take it everyone is talking about the transfer of data via USB here.
TP1 Posted November 25, 2015 Posted November 25, 2015 I've been in the computing field for a very long time, it and networking is my primary profession, audio is a just a side hobby from yesteryear, when you said you saw an improvement...what sort of improvement are we speaking here? I have tried over 30 brands and types of hard drives and SSD's for this sort and been told by some they hear an improvement but I have not been able to replicate the same result. I've definitely seen improvements in read/write speed for sure, lower thermal output/lower power usage, this leads to a quicker response with the OS boot time and applications I click on but I have not been able to hear an sound improvement. Now the old saying comes in, sometimes you need somebody to describe to you the difference for you to notice it. What sound improvements did you hear? Bits are bits and I find it hard to hearing a change. One needs to be careful about how they describe sonic improvements but I'll give it a go. For SSD I think there is better clarity and timing and in some case better tonality. Whether this or any other SQ improvement is worth it, depends upon the individual's perspective. As for the reasons that it could sound better, one theory is that there is less EMI ( no spinning magnets, less current consumed), and faster data access speed. I dont think the faster transfer speeds make a difference ( with the gear I am using) because when using SSD in a RAID 0 array I couldn't detect an improvement in sound ( there was a difference but not improvement IMO. ) .
davewantsmoore Posted November 25, 2015 Posted November 25, 2015 Why dont we hear what others purportedly can? Interconnection of fast signals the way it is done in a typical multibox digital audio system ...... is almost always a very complex mess (much more than meets the eye), and there is often too much generalisation flung around.
Recommended Posts