Tír Na HiFi

John Dot wrote:8.42 and 8.45 normal - constant clicks (sounds like latency issue).

Increasing driver buffer size increases sound issues.
Lowering it to 1 ms does not remove ticking completely.

So how small is the driver buffersize/latency supposed to be?
BTW buffersize of 2ms is way,way better than 1 ms, so wouldn't to move away from 2 ms.

8.49 sound opens out a bit

this buffer size is 10ms

more improvement with 8.51
still short a little resolution

btw i think new windows is going to sound better for us (seems to be more resolution less noise)

I think the way forward is linux with the device driver integrated with the player so the driver just gets initiated with the ram location and then it just steps through until the end all in kernel mode, ram direct to dma.

sbgk wrote:I think the way forward is linux with the device driver integrated with the player so the driver just gets initiated with the ram location and then it just steps through until the end all in kernel mode, ram direct to dma.

I am wondering if you're capable to do this? I remember you saying sometimes back that your skills with Linux are not as good as with Windows. Correct me if I'm wrong...

sbgk wrote:I think the way forward is linux with the device driver integrated with the player so the driver just gets initiated with the ram location and then it just steps through until the end all in kernel mode, ram direct to dma.

All this might be just so, but I've never heared any Linux player sound as good as MQn on Windows.

sbgk wrote:I think the way forward is linux with the device driver integrated with the player so the driver just gets initiated with the ram location and then it just steps through until the end all in kernel mode, ram direct to dma.

+1

Interesting and I definitely think Linux is theoretically superior to Windows.

Device driver? For Linux, surely none is needed? UAC 2 is natively supported in most Linux flavours and def in OS/X
Also Apple have surely already done what you propose with Direct Drive Mode (and Integer and Exclusive Modes)?

I am open to being told the above is untrue. nige will confirm I believe.

Jonathan

But of course what noone in the Apple World has done I am sure, is apply your determined approach to optimising the render loop as opposed to devoting most of their energies to the snazzy UI and controls.

Now if both approaches could be combined into one 'product', then that's a sure-fire winner.

jrling wrote:

sbgk wrote:I think the way forward is linux with the device driver integrated with the player so the driver just gets initiated with the ram location and then it just steps through until the end all in kernel mode, ram direct to dma.

+1

Interesting and I definitely think Linux is theoretically superior to Windows.

Device driver? For Linux, surely none is needed? UAC 2 is natively supported in most Linux flavours and def in OS/X
Also Apple have surely already done what you propose with Direct Drive Mode (and Integer and Exclusive Modes)?

I am open to being told the above is untrue. nige will confirm I believe.

Jonathan

i got a taste of it this morning
unfortunately i discovered my win 8.1 install was 32bit so mqn wouldn't work with it
i did find my new win server preview (win 10) which is unmodified
sq wise between OS X with audirvana 2.0 and windows media player in win server (10)is similar quality but different OS X is detailed but sounds forced/strained and noisy then tested mqn is just on another planet

then i was wondering which win driver settings i should use to compare to OS X
is there a way to set buffers and latency in OS X?

minamalist linux should be good in theory
just never seen much evidence of great sq
i do remember it is a quiet os but lacks detail and is terribly smoothed

I've never come across a linux developer that wasn't of a bits are bits mindset, so imagine there is plenty scope to tidy things up. With MS charging £1400 a year to licence drivers it makes Linux the only way forward.

Linux seems to have a number of drivers for the different chipsets so it's not true that there is a universal Linux driver.