Another Audio test - need volunteers

[jkeny]

Ooops sorry didn t see where u asked for pm only :-)

No problem, I'm changing the files anyway!

[jkeny]

Ok, new files uploaded & these are different to the previous two :)

http://dl.dropbox.com/u/76284813/File%20w.wav
http://dl.dropbox.com/u/76284813/File%20X.wav
Please let me know if there are any difficulties with downloads.
Doh! It appears that there was a problem with the link not working so I've changed it to the two links above.

[Fran]

On chrome that opens up and starts to play in the browser window rather than downloading as a file. You can just right click on the play bar in the window though and select "save as" to download to your PC.

[jkeny]

On chrome that opens up and starts to play in the browser window rather than downloading as a file. You can just right click on the play bar in the window though and select "save as" to download to your PC.

Yea, thanks Fran,
I tried to find a public link for the folder but can't, doh!.
This is becoming a nightmare as someone tells me that FileW isn't seen in Sonos.

Anyway, as it seems that the reports are no longer coming in so I will tell you what's behind this.

One of the files (I'm not going to say which) has got an attenuated (-39dB) pre-echo of the signal superimposed 20 samples in advance of the signal.

The reason I wanted to test this was twofold. I wanted to test if others heard what I heard.

The way I have processed one of the files is very similar to the way a lot of (most) digital interpolation filters work. You what, Gay? Inside your DAC chips there is a filter & I & others have suspected that it's particular implementation is detrimental to the sound. This digital filter results in an output very similar to my files - superimposing a pre-echo on top & a bit before the actual signal. This is generally considered to be inaudible as it will be masked by the louder signal which directly follows it. I have heard this effect in the treated files, particularly in the fast transients of the rim shots which are preceded by relative silence. The rim shots on one track to me sounds sharp & natural, the other duller i.e. the transient attacks are less distinct & they don't fade as naturally as the other. So this track turns out to be a good test for this effect.

[Ken Moreland]

Hi John, I understand you have superimposed data on an existing track to simulate a digital filter. Does the JKDAC32 not use a digital filter and is it giving an unadulterated output? or are we listening to the superimposed data through a non removable digital filter?
KM

[jkeny]

Hi John, I understand you have superimposed data on an existing track to simulate a digital filter. Does the JKDAC32 not use a digital filter and is it giving an unadulterated output? or are we listening to the superimposed data through a non removable digital filter?
KM

An interesting question, Ken :)
The PCM5102 which is the DAC chip inside the JKDAc32, has two filters which are switchable. There is a very audibly distinct difference between the two filters (FIR & IIR) which is what prompted me down this path. One of the filters (FIR) is the standard filter found in most DAC chips - this filter gives rise to pre-echo on it's output. The other filter (IIR) gives no pre-echo. The FIR filter sounds dull when compared to the IIR filter, much like the differences heard in these files although the effect is actually more audibly noticeable on th eDAC filters than on the processed files I produced. So naturally I have set the filter in the JKDAC32 to the IIR, best sounding one :)

BTW, this filter is pretty much the same as the Apodising filter used in a number of high spec devices such as dCs debussy, Meridian, etc. Just search for "apodizing filter" to learn more about them!