ATTENTION Current Mac mini/A+ users: Boot Mavericks from an SD card, load a RAMdisk, dismount your internal SATA drives, and pour a drink for the musicians walking out of your speakers!

[goldsdad]

Each of those zillions of interactions is mediated through the CPU. CPU-RAM interactions are the fastest and most optimized in the entire computer, far more than disk I/O, even with an SSD. So there is always a processing advantage to use of a RAMdisk. Whether there is a sonic advantage is up to you.

 

Audirvana on RAMDisk, when launched, will be loaded from the RAMDisk into RAM, then the program in RAM (not the RAMDisk) will be processed until quit without any further use of the Audirvana package on RAMDisk. I don't understand why you think there is any processing advantage to this Audirvana RAMDisk while it isn't being processed.

[Freann]

Using an app called RAM Disk Creator. Works well.

 

http://bogner.sh/wp-content/uploads/2012/12/RAMDiskCreator.app_.zip

 

[Jud]

You're missing nothing, in my opinion. I'm in complete agreement with you. I think some people have a mistaken idea that the program code on the storage device (real hard drive or, in this case, a virtual drive called a RAMDisk) is somehow directly processed while the program runs. As you realise, the reality is that, when a program is launched, it is loaded into RAM then the instructions in RAM are processed until the program is quit. The program in the RAMDisk seems to make a fine placebo, if nothing else.

 

An experiment you can run at home if you have some facility with Unix-ish OSen: Do a kernel compile with an SSD and a RAMdisk, and see what if any time difference there is.

[Jud]

Audirvana on RAMDisk, when launched, will be loaded from the RAMDisk into RAM, then the program in RAM (not the RAMDisk) will be processed until quit without any further use of the Audirvana package on RAMDisk. I don't understand why you think there is any processing advantage to this Audirvana RAMDisk while it isn't being processed.

 

Thanks for the correction, as always.

 

Nevertheless, I do think I hear a difference; and as noted previously, another memory player developer also advises RAMdisks for both music files and app. No particular reason I can think of for him to say this unless he thinks it sounds better. And unlike me, he actually knows something about computers.

[Freann]

@Jud: way over my head now...

But timing aside – when does an app talk to a disc other than the one controlled by the OS? A+ uses a scratch disc like Photoshop?

[Jud]

music files actually load to RAM as well so other than the SD card, there's no process here that's not already being done........but let your ears be the judge as they almost always tell the whole story.

 

Not only that, but I thought that having the unmounted HDD connected via FW cable while playing back from SD card sounded worse than when the FW cable was disconnected. (That was actually an unintentional blind test. I forgot to unplug the FW cable, then walked to the couch to listen and couldn't understand why SD card playback was sounding more like FW HDD playback. When I got up from the couch to investigate, I saw the FW cable was still plugged in.)

[Jud]

@Jud: way over my head now...

But timing aside – when does an app talk to a disc other than the one controlled by the OS? A+ uses a scratch disc like Photoshop?

 

No idea. Unfortunately, though Audirvana was at one time open source, I know zilch about coding, so I never understood its internal workings.

[Freann]

Ah, the mysteries of software development!

 

Now running OS from SD card and music files from RAM disk. Need to compare against previous setup using a 2008/brick power supply mac mini.

 

Awaiting delivery of a Acoustic Revive Lan filter & a 2.5" FW800 powered 2TD drive.

[goldsdad]

Thanks for the correction, as always.

 

Nevertheless, I do think I hear a difference; and as noted previously, another memory player developer also advises RAMdisks for both music files and app. No particular reason I can think of for him to say this unless he thinks it sounds better. And unlike me, he actually knows something about computers.

 

I accept that he thinks there's an improvement when storing his app on a RAMDisk, but that isn't proof that the sound waves differ from when his app is stored on a real drive.

[Booster MPS]

I accept that he thinks there's an improvement when storing his app on a RAMDisk, but that isn't proof that the sound waves differ from when his app is stored on a real drive.

 

I have tried it and find myself in the camp of those that found a difference being able to unmount my internal drive, unplug my wall wart, and disconnect the external drive from the computer. I would just say, give it a try and report back here on what you find. It's worth the actual experiment to determine for yourself (sorry if I missed in a post where you gave it a try)

[Le Concombre Masqué]

music files actually load to RAM as well so other than the SD card, there's no process here that's not already being done........but let your ears be the judge as they almost always tell the whole story.

As much as i criticize the rd for A+, I advocate the rd for music files.

that the music files end up in ram anyway after processing does not mean that there is no difference wether they originate from ram or ssd. To my ears it does make a difference.

So does playing wav files : same bits than flac or aiff, but in reverse order; and saving, at the crucial stage of processor level, the tiny time of processing them in the right order prior to the main processing makes an audible difference.

[Jud]

I accept that he thinks there's an improvement when storing his app on a RAMDisk, but that isn't proof that the sound waves differ from when his app is stored on a real drive.

 

Absolutely correct. But OTOH, also shows that someone who actually knows how software works considers the idea rational.

[Viola]

Don't forget to check the "prevent app nap" on Audirvana + on your new Ramdisk. :-)

[goldsdad]

I have tried it and find myself in the camp of those that found a difference being able to unmount my internal drive, unplug my wall wart, and disconnect the external drive from the computer. I would just say, give it a try and report back here on what you find. It's worth the actual experiment to determine for yourself (sorry if I missed in a post where you gave it a try)

 

You've missed the point. You cannot unmount, disconnect and unplug your OS and apps drive. My point is that loading Audirvana from RAMDisk to RAM has no advantage over loading Audirvana from OS and apps drive to RAM because, in both cases, it's the code in RAM that is then run, not the code on the drive.

[goldsdad]

Absolutely correct. But OTOH, also shows that someone who actually knows how software works considers the idea rational.

 

Do we know that he does it for a rational reason rather than just preferring the sound regardless of rationality?

[Jud]

Do we know that he does it for a rational reason rather than just preferring the sound regardless of rationality?

 

I don't know of a reason putting the app on a RAMdisk should make any difference (accepting your statement that both are run from RAM anyway), do you? The notion I'm speaking of as rational is that there is a difference. Only then do we reach the question of preference. Whether PeterSt has done any measurements to back up his preference I don't know, though that's something he usually likes to do. My preference is just from listening. However, I try to be careful to go for accuracy over euphony, and sometimes I succeed - please see my comments in the DSD Battle Royale thread.

[Freann]

Last nights setup

2010 Mac Mini unibody/Mavericks on a SD card/Music on a RAM disk

 

Todays setup

2008 Mac Mini/Mavericks on HD/Music on a RAM disk

 

I prefer the 2008 mac. Simply because I hear hardness and grit from the built in power supply in the 2010 Mac Mini. If I modify it with an external analogue power supply the reverse would probably be true. Smaller sound stage and less quick bass with the 2008 setup.

 

Has any one else compared old vs. new Mac Mini?

[Le Concombre Masqué]

music files actually load to RAM as well so other than the SD card, there's no process here that's not already being done........but let your ears be the judge as they almost always tell the whole story.

everybody (here ?) takes for granted as a benefit that PM A+ Amarra (etc?) are RAM players.

 

Once A+ has loaded a file in RAM there is very little subsequent processing activity (check activity monitor)

 

when A+ loads a file in RAM there's a huge processing activity (especially if plugins, up converting included, are involved) : anyone who admits a benefit from RAM read for a small processor activity should expect a bigger benefit from first loading music files in RAM

[goldsdad]

I don't know of a reason putting the app on a RAMdisk should make any difference (accepting your statement that both are run from RAM anyway), do you?

 

I posted because I can't think of a reason other than placebo for the app on a RAMDisk making a difference in what people hear.

[Superdad]

Last nights setup

2010 Mac Mini unibody/Mavericks on a SD card/Music on a RAM disk

 

Todays setup

2008 Mac Mini/Mavericks on HD/Music on a RAM disk

 

I prefer the 2008 mac. Simply because I hear hardness and grit from the built in power supply in the 2010 Mac Mini. If I modify it with an external analogue power supply the reverse would probably be true. Smaller sound stage and less quick bass with the 2008 setup.

 

Has any one else compared old vs. new Mac Mini?

 

 

I agree that the power supply is the issue. Last year, when I ran a controlled 3-way shootout between a late-2009, 2010, and 2011 (all with same clean and trim ML, 8GB RAM, A+ 1.4.x), my 2009 had an $80 13.8V/10A regulated linear PS, while the others had their internal switchers. Bass detail, tightness, authority, etc. were all better with the 2009 on external supply. But the 2009 still lost to the 2010 in a bunch of other areas (and the 2011 fell short too, though I may have used the wrong USB port), so I went with the 2010 with the intention of going linear (next week!).

 

Of course at the time we did not have Mavericks or the gloriousness of A+ 1.5.10, I was not using a RAM disk, and tunes were coming in from an external FW. Now add in SD card booting (something the non-unibody minis lack), Thunderbolt or SATA options, and the external PS, and there is not much looking back.

 

By the way, what mini model are you referring to as a 2008? Intros were 2007 and 2009. I am guessing you have one of the 2007 Core 2 Duo units. Don't those max out at 4GB RAM total?

 

Ciao,

Alex C.

[Superdad]

Something I do periodically on this journey--especially when a bunch of improvements come in succession over just a couple of months--is to jump backwards a few steps to remind myself of where I was and to assess the cumulative magnitude of the changes.

 

So last night I rebooted my mini into Mavericks from the internal hard drive (a clone b/u of my trimmed and tweaked SD card boot), launched A+ (from the HD) in iTunes integrated mode, and played tunes straight from my best sounding external FW400 drive. No RAM disk, no SD card, no Playlist mode.

 

The difference in presentation was so large it made me laugh. I even dragged my wife in, and her comment was that it sounded like a different recording of the same piece.

Then for fun while I had her ear, I quit Audirvana and played straight from iTunes. She made a really ugly face and walked out of the room.

 

These are the times when I try not to think about people like Archimago or others who test playback s/w with null tests and/or can't measure differences among any of this, and then declare that there is nothing to be heard. Sorry, I guess I am violating my own thread intention to avoid polarizing discussion about what can and can't be heard/measured.

 

I am sure my room/system spoils not just myself, but my wife as well. Yet when I jump back just those few steps (no hardware changed!) and A/B against where we are today, it just bowls me over.

[Freann]

@Superdad: interesting shootout. Mine is a server model with a 2.53 Core 2 Duo with 8GB ram.

Which linear PS are you getting for your 2010?

[Jud]

@Superdad: interesting shootout. Mine is a server model with a 2.53 Core 2 Duo with 8GB ram.

Which linear PS are you getting for your 2010?

 

He and an engineer friend are making one.

[Jud]

I posted because I can't think of a reason other than placebo for the app on a RAMDisk making a difference in what people hear.

 

Right, I think we've established neither one of us can think of a reason why this should make a difference. But I think I hear one. Of course placebo is always a possible explanation. But PeterSt certainly knows more about this stuff than I do, and he apparently does not discount the possibility it is real. (As I noted, I don't know whether he's measured a difference, though he usually likes to carry out measurements.) So for the moment at least, I will not discount the possibility it is real either.

[elcorso]

Right, I think we've established neither one of us can think of a reason why this should make a difference. But I think I hear one. Of course placebo is always a possible explanation. But PeterSt certainly knows more about this stuff than I do, and he apparently does not discount the possibility it is real. (As I noted, I don't know whether he's measured a difference, though he usually likes to carry out measurements.) So for the moment at least, I will not discount the possibility it is real either.

 

Good answer Jud!

 

If I followed this politics with my wife each time whenever she contradicted me (ALWAYS), maybe we would be living in the same house...

 

Roch