Tír Na HiFi

sbgk wrote: the files are bit identical, what mechanism is at play ? you could copy the contents from one to the other, do they still sound the same ? If you duplicated them and played either both linear or non linear would you think they are different. I played each file in a loop via mqn and after a while couldn't differentiate between the linear and non linear.

Don't know what mechanism is at play when files are copied or transmitted as they apparently pass through multiple buffers along the way. I can understand how, in a mixed signal (digital & analogue) device like a DAC, that the signal waveform of the digital signal arriving at the input of the DAC can have an effect on it's analogue output even though it is a perfectly acceptable waveform for digital use.

Yes, listener fatigue will render subtle differences indistinguishable.

jkeny wrote:

sbgk wrote: the files are bit identical, what mechanism is at play ? you could copy the contents from one to the other, do they still sound the same ? If you duplicated them and played either both linear or non linear would you think they are different. I played each file in a loop via mqn and after a while couldn't differentiate between the linear and non linear.

Don't know what mechanism is at play when files are copied or transmitted as they apparently pass through multiple buffers along the way. I can understand how, in a mixed signal (digital & analogue) device like a DAC, that the signal waveform of the digital signal arriving at the input of the DAC can have an effect on it's analogue output even though it is a perfectly acceptable waveform for digital use.

Yes, listener fatigue will render subtle differences indistinguishable.

don't think that's the same thing as bit identical files will load into the dac in an identical way. was playing 20 secs of each track in a loop, so fatigue wasn't an issue.

sbgk wrote:

jkeny wrote:

sbgk wrote: the files are bit identical, what mechanism is at play ? you could copy the contents from one to the other, do they still sound the same ? If you duplicated them and played either both linear or non linear would you think they are different. I played each file in a loop via mqn and after a while couldn't differentiate between the linear and non linear.

Don't know what mechanism is at play when files are copied or transmitted as they apparently pass through multiple buffers along the way. I can understand how, in a mixed signal (digital & analogue) device like a DAC, that the signal waveform of the digital signal arriving at the input of the DAC can have an effect on it's analogue output even though it is a perfectly acceptable waveform for digital use.

Yes, listener fatigue will render subtle differences indistinguishable.

don't think that's the same thing as bit identical files will load into the dac in an identical way. was playing 20 secs of each track in a loop, so fatigue wasn't an issue.

Bit identical just means that the files carry the same digitally interpreted signal - it doesn't mean that the signal waveform (the actual electrical waveform) contained in both files are the same. I'm not supporting or otherwise - just suggesting a possibility to explain things that can be bit identical but not physically identical

jkeny wrote: Bit identical just means that the files carry the same digitally interpreted signal - it doesn't mean that the signal waveform (the actual electrical waveform) contained in both files are the same. I'm not supporting or otherwise - just suggesting a possibility to explain things that can be bit identical but not physically identical

it does, though. I've done a diff on the 2 files, opened them in a hex editor and they are identical. So if there is a difference it is in something else. you could argue that the quality of the bits may vary in that they are easier to read in a linear capture file that only makes a difference when playing music. When mqn is loading data into ram it is loading 1s and 0s so little chance of that being an issue.

theoretically you could isolate a byte that gave a different sound, but when you examine the byte in a hex editor they are identical.

do things like playing files from a root directory apply to a ram player like mqn ?

sbgk wrote:

jkeny wrote: Bit identical just means that the files carry the same digitally interpreted signal - it doesn't mean that the signal waveform (the actual electrical waveform) contained in both files are the same. I'm not supporting or otherwise - just suggesting a possibility to explain things that can be bit identical but not physically identical

it does, though. I've done a diff on the 2 files, opened them in a hex editor and they are identical. So if there is a difference it is in something else. you could argue that the quality of the bits may vary in that they are easier to read in a linear capture file that only makes a difference when playing music. When mqn is loading data into ram it is loading 1s and 0s so little chance of that being an issue.

theoretically you could isolate a byte that gave a different sound, but when you examine the byte in a hex editor they are identical.

do things like playing files from a root directory apply to a ram player like mqn ?

Yes, that's the bit I can't explain - how the 1s & 0s in RAM could be any different as a result of where they are being loaded from?
I can understand how the quality of the signal waveform (i.e the electrical representation of those 1s & 0s - the digital signal) arriving at the DAC could have an effect on the analogue stage of the DAC & as a result, the SQ.

nige2000 wrote:just done a audirvana bells whistles etc versus r2 + mqn on the macbook pro
mqn and r2 left mac OS X and audirvana in the dust
most noticeable is the higher noise floor then the lack of detail and dynamics
im not talking about a little difference its pretty huge
so i really don't think I've got it wrong:)

and yes i think audirvana is doing a good job on OS X especially with all the magic settings

your hard to convince:(
thats why i use mac os for emails and internet related stuff

I am, but you are doing a great job! But only in an investigative way trying to find 'The Holy Grail'. Like you.
I respect your devotion and sheer hard work in trying stuff out, which others like me can benefit from. The whole purpose and joy of forums like this.

I still harbour the feeling that Apple's Linux-based audio stack is superior to Windows. If only everything else surrounding it - especially the render loop - could be optimal, I reckon it would come out the winner.

At great danger of outstaying my welcome, I think you have got a lot closer to comparing Apple v. apples (well someone had to say it!)
Both on the same machine (albeit a laptop which could be contributing more noise than your proper PC-based rig)

BUT - as we know the software is key. Were you comparing an AO optimised R2 with a plain vanilla Mavericks? That would def not be a fair fight as we all know. OS/X is able to be optimised along the same lines as R2 and would no doubt level the playing field.

I am actually encouraged by your findings. MQn is still clearly the winner, but others are catching up. If detail and dynamics are on a similar level, then the render loops must be close in performance. Noise level could be more hardware related. This is great news for all CA enthusiasts and Gordon can take a lot of the credit for forcing the competition to catch up with his pathfinding activity.

Gordon - 'MQn for Mac' next?!

Jonathan

P.S. I have an iMac 21" at home which is our main family email/browsing machine and I love it to bits (!)

ok
only one way to solve this
you can set the conditions and the testing method
you can pick any player, pc and os I've got

so heres what I've got

ivybridge desktop pc with win 8.1/osx Yosemite
macbook pro Yosemite
macbook air mavericks
i can boot anything into either win server r2 std or AO 1.3

at a bit of a push i can install OS X on to my audio pc (but thats a bit of work not to mention a waste of time because OS X is a noisy os)

i can do a headphone test with std win 8.1 and Yosemite on my desktop with a player of your choice fairly easily

point begin no matter what the hardware is tweaks etc OS X will sound worse on the same pc/mac in comparison to win 8 upward

trust me theres none more disappointed that macs sound worse

other than that ill have to catch a flight to london for home demonstration

btw i hear the iMac is particularly bad at audio might be because they put the main board in a emi/rfi hell behind the monitor
beautiful device though

nige2000 wrote:ok
only one way to solve this
you can set the conditions and the testing method
you can pick any player, pc and os I've got

so heres what I've got

ivybridge desktop pc with win 8.1/osx Yosemite
macbook pro Yosemite
macbook air mavericks
i can boot anything into either win server r2 std or AO 1.3

at a bit of a push i can install OS X on to my audio pc (but thats a bit of work not to mention a waste of time because OS X is a noisy os)

i can do a headphone test with std win 8.1 and Yosemite on my desktop with a player of your choice fairly easily

point begin no matter what the hardware is tweaks etc OS X will sound worse on the same pc/mac in comparison to win 8 upward

trust me theres none more disappointed that macs sound worse

other than that ill have to catch a flight to london for home demonstration

btw i hear the iMac is particularly bad at audio might be because they put the main board in a emi/rfi hell behind the monitor
beautiful device though

Blimey! And to think I thought Superdad was thorough! I am feeling slightly guilty. But definitely up for your very kind offer/challenge.
Wonder though if it is a bit off topic for others and so will PM you.

We can always publish the findings on completion.

Cheers
Jonathan

8.42 phew, no double blind trolls were harmed in it's production. A different sound, anyway.

sbgk wrote:8.42 phew, no double blind trolls were harmed in it's production. A different sound, anyway.

8.41 and 8.42 have vinyl like ticking on 4416.