advertisement


a downsampling listening test

Werner

Werner

pfm Member
For those interested in the audibility of different approaches to sample rate
conversion and digital filtering I was planning an extensive series of
exploratory listening tests. Alas, I lack the time for this.

However, I did prepare a quick test suite that I will make now available
to listeners.

Send me a PM and I'll return a link to a zip archive with 3 wav files:
A, X, and Y.

A is an excerpt of an original 96kHz / 24 bit recording.

X and Y both have been downsampled to 44.1kHz, and then back up to 96 kHz.
Two different downsampling methods have been used, both with utter care for
mathematical hygiene, both with the filter cut-off at 20kHz.

The result is that the experimental filters will totally dominate your listening
experience, regardless of the type of DAC you employ.

What I ask of participants is that they listen leasurely to A, X, and Y, and try
to determine which of X and Y sounds worse, and why. Then please report in this
thread, along with the DAC and speakers or headphones you used.

Again, both X and Y have been downsampled and contain roughly the same
information. This is not a test of your ears, only of your preferences, you
cannot win or fail.
 
So far only three candidates (not that I expected much more). They should have received their file links by now.
 
I've done a preliminary listen and I will have a serious attempt in a short while. To give it the "best" shot I need to move files onto hi-fi system.
 
Yes, you'll need a capable system. The filters used are state of the art, and differences are bound to be subtle. I think.


Please refrain at least one more day from commenting on the sound, so as to allow the others (more now!) to listen without prejudice.
 
Make sure if you listen on the computer that your system is actually running at 96KHz. Windows runs at a preset sample-rate and converts everything to that.
 
Tenson said:
Make sure if you listen on the computer that your system is actually running at 96KHz. Windows runs at a preset sample-rate and converts everything to that.
Click to expand...

i've sent it to my prism dac set up, will probably only be able to use headphones but i have 4 flavours of those....
 
That's why I transferred the downloaded files onto my ace-o-matic system that does, in fact, tell me that it is playing 96kHz files.
 
i'm am not sure that the song is mastered good enough for the particular test, i struggle to hear difference between the oridginal and any of the worse ones, i have had preference towards X but it might just be a quincidense.
 
10 listeners now. Wait a bit more with your comments.


AK is an audiophile fave. I don't have a lot of hi-res files, but this one sounds better than many.
 
Why on earth is any Windows user going on about how to play a couple of files without the system molesting the content?

If any Windows user wants to try this who does not, as yet, have a known ‘bit-perfect’ playback method, then simply download Foobar together with its WASAPI output module if you are using a recent version of Windows or the kernel streaming module for XP. Using either of these output modules ensures that Windows will not intervene and transmit the audio, unchanged, to the output device.

You can also get a very handy ABX module that will switch, instantly, between the files under test while keeping them at the same running point. It doesn't get any easier.
 
Plutox said:
Why on earth is any Windows user going on about how to play a couple of files without the system molesting the content?

If any Windows user wants to try this who does not, as yet, have a known ‘bit-perfect’ playback method, then simply download Foobar together with its WASAPI output module if you are using a recent version of Windows or the kernel streaming module for XP. Using either of these output modules ensures that Windows will not intervene and transmit the audio, unchanged, to the output device.

You can also get a very handy ABX module that will switch, instantly, between the files under test while keeping them at the same running point. It doesn't get any easier.
Click to expand...

Thanks - to confirm, with the WASAPI module I do not need to manually change the audio device to match file bit depth or frequency?

Richard
 
Dowser said:
to confirm, with the WASAPI module I do not need to manually change the audio device to match file bit depth or frequency?
Click to expand...

Confirmed. On preferences / output, make sure 24 bit is selected, and no dither. I believe this is the default state so you should be OK.

Forgive me if I'm teaching granny to suck eggs, but bear in mind that the mere existence of the WASAPI output module will not put it into use, it must be specifically selected in the 'device' section of the above page. WASAPI (event) followed by the device of choice should do the job.

Having done that, your DAC ought to lock onto whatever signal the computer is putting out. Of course, there are always exceptions ;)
 
An interesting test.

I will post my listening comments after Werner says it's OK to do so, but I think the difference between the two test samples is something along the lines of linear phase & minimum phase filtering.
 
You must log in or register to reply here.


advertisement


Back
Top