Tír Na HiFi

jkeny wrote: Anyone who has ever used a bench supply ALONE to power their computer will see the current draw spikes that happen even when you just move the mouse on the screen - it's a surprising high current spike. Opening/closing windows (with a small m), everything causes current draws. It would be interesting to have a table of current usage Vs windows actions/activities.

What that shows me is that even the simplest of actions in Windows (moving a mouse) can cause a lot of activity under the hood (as signified by the current spike). It just makes me pause about equating cause with effect in a complex OS although I doubt using the current draw on a bench supply could reveal differences between code changes in MQN

jkeny wrote:

jkeny wrote: Anyone who has ever used a bench supply ALONE to power their computer will see the current draw spikes that happen even when you just move the mouse on the screen - it's a surprising high current spike. Opening/closing windows (with a small m), everything causes current draws. It would be interesting to have a table of current usage Vs windows actions/activities.

What that shows me is that even the simplest of actions in Windows (moving a mouse) can cause a lot of activity under the hood (as signified by the current spike). It just makes me pause about equating cause with effect in a complex OS although I doubt using the current draw on a bench supply could reveal differences between code changes in MQN

I expect what's happening is that PCs are designed to save power and especially minimise heat but you'll have thought of this already. It does show the sophistication of the systems involved and says a little about why this is all so complex.

best to fix the cpu frequency to help stop the fluctuation of current draw
cpus are at idle until theres a load, then spikes usually between 1.6 ghz and 3.4ghz or what ever your cpu does and that can double to triple the cpu current load.

sima66 wrote:Cvrle, be my guest. After glass or two of slivovitza everything will make more sense, everything will be much clearer! ;-)

No problem, it's easy to lure me in with slivovitza, but my ears will get even worse after "contamination", lol!
Let's stop talking about this, as we are already notified that there are more important things to be figured out here.
Sorry guys, I am off this topic, but this time for sure.

sbgk wrote:
neither of the 2.64 versions was optimal, however I can see that intrinsics has potential, shall do an sse4 and sse2 optimal version and then we can all get back to listening to music.

uploaded 2.65 sse4 intel/sse2 intel - the optimal sse4 and sse2 intrinsic versions (according to Intel), haven't listened to them yet.

on the electrical noise topic, aren't eavesdroppers supposed to be able to read what's on pc screens by using the electrical signal generated by the screen, the pc needs to be near a window ie they can read the signal from outside the building.

if the problem is electrical noise, why not just put a load of ferrites on the usb cable ? works on mains cables. AQVOX says usb cables shouldn't have ferrites, but don't understand why not as they are passive, it's all a mystery.

sbgk wrote: uploaded 2.65 sse4 intel/sse2 intel - the optimal sse4 and sse2 intrinsic versions (according to Intel), haven't listened to them yet.

both sound really good, sse4 loses a bit of treble, think I know why, shall upload a v2 version.

the v2 version has a mistook, uploaded v3

The 2.65 sse4 v3 version is absolutely expletive deleted amazing, can't believe how good it is, no need to test any other version, this is the one.

May be able to improve the sse2 version, but I think that's it for sse4.

My God this thread has some growth rate ! I did not read it for 2 - 3 days and I have a back log of 40 pages ... such a shame that my audio-pc is AMD based and because of it I can only read about MQN. Well jmini and hibernate are also good companions and most important I did not hear MQN yet so I dont know what I am loosing. Now I have to keep it this way as long as I possibly can ;)

Trying out 65 sse4 intel and v2. Very good again. No harshness that I can detect just really full bodied rich sound. Need to listen more tomorrow to see if all detail still present but tone is lovely i.e warmth.

sebna wrote:My God this thread has some growth rate ! I did not read it for 2 - 3 days and I have a back log of 40 pages ... such a shame that my audio-pc is AMD based and because of it I can only read about MQN. Well jmini and hibernate are also good companions and most important I did not hear MQN yet so I dont know what I am loosing. Now I have to keep it like this as long as I can ;)

it works on AMD, I can also do an AMD compile, but it should work on AMD 64. Check with cpuz if you have sse2 or sse4.1 and select the appropriate version of MQn.

the 2.65 versions are the best

tony wrote:Trying out 65 sse4 intel and v2. Very good again. No harshness that I can detect just really full bodied rich sound. Need to listen more tomorrow to see if all detail still present but tone is lovely i.e warmth.

you need 2.65 sse4 intel v3, v2 is missing a bit of treble.

had deleted v2, but have brought it back, but think it is missing some treble, well I know it is because there is a command that has that effect and when I remove it the result is v3, I don't think the command is required for music streaming - something to do with out of order execution which modern chips have, but any issues will be picked up as missing bits ie crackle.