advertisement


MDAC First Listen (part 00100001)

Status
Not open for further replies.
I can confirm that DSD data as DoP works fine with a SB Touch which CPU is much slower than the RPI. I have tested this with Chord Hugo which confirmed it received DSD.
 
Regarding Jplay, if my Windows computer is capable of bit perfect transfer, I don't understand how giving the CPU an easier time by shutting down non-audio related processes could improve upon this performance. As I understood it (quite possibly wrongly!), once you have established bit perfect data transfer to the MDAC, the only other potential computer related problem is the possibility of hardware related noise being be transmitted from the computer to the MDAC via a direct USB connection, and an isolator should solve this. If so, I suspect Jplay is unlikely to improve upon my Windows computer set up as I use it (when it's playing music, it's not doing anything else).

Given that both my RPi and Atom are only running dedicated audio software and nothing else (they are both set up headless and therefore don't even have to drive a display), It seems unlikely that something like Jplay could improve upon their performances. Also, Jplay is a Widows application and the Pi and Atom run Linux...

However, theory must be tested by experiment! Therefore, I will try Jplay on the Windows 8.1 PC and report back once the boss (she who must be obeyed) and I have conducted an ABX test with it against straight JRiver on the same computer. I will try the two computer set up, if I can. I'm not sure if I can do this with a Windows machine and the RPi (different OSs), but we do have more than one Windows machine, so I could use two of these (we have more computers in our household than might be considered healthy!).
 
Have done lots of listening tests myself.. Not ABX, just my own ears.. SB Touch, RPI, CuBox I4, low power laptop and high power Core I5 gaming pc..
The only one I could hear a difference with was the low power laptop.. It injected all sorts of noise down the usb cable when it was charging.. You could hear when it was working hard/spinning drives etc as a sort of white noise morse code that was very obvious. Olimex USB isolator helped but didnt solve this. When it was not charging it was silent.

Have tried out all sorts of playback software, and the conclusion I drew from listening tests there was that the best software to use on a windows pc is the one that has the interface that works best for you.. Couldnt hear any difference but some of the interfaces really did not work for me.
I settled on windows 8.1 with XBMC, but I also use my HTPC for playing films/tv series that are held on the network, as well as Netflix, spotify, tv catchup and gaming... Bit of tweaking of the windows start menu and it works well.. SWMBO can operate it without too much fuss as well so cant be far wrong!
 
Both the Rega Fono A2D and the MDAC are Slave USB devices - they can only be connected to a Master device (USB Host) such as a computer.



The MDAC2 L3 has an internal ADC so will not require an external ADC.



The MDAC2 L3, ADC version = Analogue inputs will require a Phono Pre-amplifier - later I'll release a Phono Gain stage with an EQ bypass option so RIAA EQ can be performed in the Digital domain.

Everything clear ! Thank you John for the explanation. Paradox is, that I'm looking forward for the (2014) MDAC2 for finally listening to my (1972) Thorens TD 160 :) since my two mono ampli have a single line in which is actually dedicated to the MDAC !! :)
 
Re. DSD-capability.

Since the MDAC2 is able to handle DSD, I would be interested in knowing how much cpu/ram power is needed for a computer player to replay DSD.

1) Will the Raspberry Pi be enough?
2) An Atom DN2800mt pc with 8gb ram + Win Server 2012?
3) Or do you need an i3 /i5 /i7 processor with 8gb ram or more?

The pc / rpi will be used as a dedicated audio computer.

Edit: I already own option 2.

Rosewind,

Replaying DSD requires the same CPU overhead as replying a bit-perfect PCM file (with no precessing) - to the CPU they are one and the same.
 
Thanks John. I had seen various claims that a high-spec cpu was needed to play back DSD. Perhaps that was only the case if you wanted everything upsampled to DSD?
Cheers, Peter
 
What I always find strange is that someone would use windows as the operating system on a computer purely dedicated to playing music. If it is a dedicated system, use an operating system that can be optimized for dedicated/embedded operation, not one optimized for general desktop computing.

And no, this is not about being pro-windows or anti-windows - I can totally see using windows if you also use your computer for other stuff besides playing music. But taking a windows PC, and shutting down "all but essential activities" renders it useless for general use, so why not go for a dedicated solution in the first place?
I'm definitely coming round to this point of view. I recently decided to dip a toe into the world of linux, partly in preparation for the arrival of my cubox-i, and partly in preapration for gettign the mdac 2 to work with its embedded board running squeezelite and maybe even lms.

It was a wee bit of a fiddle, but not too bad, and I am amazed at how my 6 year old machine which creaked and spluttered running vista now skips along merrily running debian wheezy.

Well obviously you are not able to use the computer whilst music is playing, that is self evident. And yes if you have the luxury of having a dedicated computer for music . . . But many people don't, hence JPLAY's success.
I get that, but I still think that ultimately having a machine running lms as a server is much better. It can run 24/7 serving music all over the house and can be hidden away in a cupboard.
And what's more i reckon that most people could lay their hands on a machine capable of running it for nothing, or next to nothing.

People that don't have a second computer could get a RPI.
That too. Or any number of similar machines. Or a NAS. My QNAP can run LMS pretty well too, although it can't do DRC.
 
And what's more i reckon that most people could lay their hands on a machine capable of running it for nothing, or next to nothing.

Indeed - for a long time my server was a 10-year-old "past it's useful life" ex-desktop machine stuffed away in the boiler room.
 
Indeed - for a long time my server was a 10-year-old "past it's useful life" ex-desktop machine stuffed away in the boiler room.

I'm digressing here but I was amazed by the fact that I could get the mail program to sinc with my office exchange server with one simple plugin. And I rather like ice weasel too.
That said I'm a bit apprehensive about getting Arch running, which I understand I may need if I'm going to get a cubox-I running Xbmc and Squeezelite.
 
JohnW,
just a quick question - will it be possible (with MDAC2) to use the DAC and ADC independently at the same time via USB? I wouldn't mind limited functionality (ie. PCM only), it's just that it would allow me to use the MDAC for both playback of some base "guide" track (to headphones) and recording of a new track (via a microphone amp -> ADC), ideally with quite low latency (though not required).

I presume the ADC is going to appear like a PCM recording device in the OS, so it could work - the OS would simply open the device (request recording), which would activate the ADC, much like opening the device for playback activates the DAC now.

If it's not possible, I can live with that / use another DAC for playback.

Thanks,
Jiri
 
Jiri,

There is no technical reason the DAC / ADC section could not be used independently, it would be more of a software issue. I guess it would be best that the ADC / DAC sample rates would be the same or multiples of the FS rate (if not the same) such as 44.1 / 88.2 or 48 / 192 etc.
 
Thanks John. I had seen various claims that a high-spec cpu was needed to play back DSD. Perhaps that was only the case if you wanted everything upsampled to DSD?
Cheers, Peter

"Upsampling" to DSD could be processor intensive, but we just require "bit Perfect data" sans any processing.... just "dumb" data streaming...
 
Jiri,

There is no technical reason the DAC / ADC section could not be used independently, it would be more of a software issue. I guess it would be best that the ADC / DAC sample rates would be the same or multiples of the FS rate (if not the same) such as 44.1 / 88.2 or 48 / 192 etc.
Okay, thanks. I presume that by "software issue" you mean "firmware / clock" as the OS sees those two subdevices as independent (with independent sample rates) and has no problems using one with different sample rate than the other (tested with onboard Realtek soundcard, under both linux and windows).
 
Thanks John. I had seen various claims that a high-spec cpu was needed to play back DSD. Perhaps that was only the case if you wanted everything upsampled to DSD?
Cheers, Peter
By the way, DoP is just a packaging format, it contains the original DSD data, so even converting DoP to DSD on-the-fly has little overhead. Converting normal PCM to DSD is something completely different (and theoretically not worth doing).
 
Hi John, all

known issue to keep an eye on:
taking measurements for a wood cabinet, I just noticed 2 electrolytics 470uf 6,3V on both side next to the middle 7805ct regulator are swollen and some yellowish content have been "sweat out" on top of them, I think I better buy a handfull quality replacement for them all, can you recommend a brand/type or will standard Panasonic M series do?
I have photos but cannot attach to this list
mine are the Premium TOY version
 
Status
Not open for further replies.


advertisement


Back
Top