My Clari-Fi magic box arrived. What do you want to know about it?

how bout just doing some recordings with and without the box. Then run them through wavesurfer http://www.speech.kth.se/wavesurfer/ or some other analysis tool. just like spectrum, SPL ect. Would be nice to check out any quantifiable difference it makes.
 

Record pink noise and a frequency sweep with and without (ideally both from a wav source and an mp3 encoded from that wav), and provide wav files (original, recorded plain, recorded throuh filter) here for people to run analytics on.

That should be enough to figure out the characteristics of the circuit before you crack that sucker open.
 

It would be fair to measure in the analog domain, as going back to digital might re-introduce the artifacts it claims to smoothen.

So I'd run it x y into a scope, for a start.
x straight from the computer cq mp3 player
y via the thingy.
Normally, you'd expect a straight line. Filtering tends to open the figure to an elipse.

A sweep is nice, but sending it spikes and misformed waves and square waves is much more interesting.
 

Something else you can do in the analog domain:

Feed the output of the mp3 player directly and via the Clari-fi into a good mixing console.
Set panning pots to L and R completely on both sets of channels.
Flip the phase of the inputs of the clari-fi.
Make sure the EQ's are off.
Set faders to 0 dB.
Adjust gains precisely so that silence is approached.

This way, you've substracted the clari-fi's sound from the original, leaving you only with the signal the clari-fi has removed/altered.

Judge the output both on a scope and by ear.

phase-shifts will appear as filtered sounds.

Please post any findings on your blog.
:-)

Thanks! I'd be delighted to hear from you.
 

I like where this thread is headed.
 

If it cuts the volume by 10% I couldn't use it no matter how good it sounded.

OT alert:

The output of my ipod nano is pathetically quiet.

I guess there are two reasons:

1) It's designed for listening to modern music that has SUPER LOUD MASTERING and forgets that some of don't listen to that kind of music

2) Fear of lawsuits from people so dumb they can't take responsibility for the food they eat or the level at which they listen to their music and so we all have to suffer as apple turns it down for them.

mini-rant over.
 

I would suggest you run the hammer test. It's the best quality control for obscure plastic things.
 

Unscrew it, note the value of the capacitor and the resistor you find in there and post them here.

And checkout, buy some decent isolating earphones (Shures or Etymotics). The Nano isn't quiet (although your point about mastering is well made) - the Apple earbuds make it perceptually quiet as they let the outside world in; probably on the grounds that people who get run over by cars tend to sue quicker than people going deaf.
 

Square waves compared with and without at a variety of higher frequencies could reveal what it's actually doing. If you record the output of the experiments using a decent microphone then we could all see the actual waveform comparisons.
 

Let's do the coca-cola challenge on this box, I want to clean up my DJ sets:

1. Find an old, badly-recorded rock track, a new, well-recorded & over-mastered rock track, a nice clean techno track, a quiet folk song, and, I don't know, some Trans Am.

2. Record each track through the box, and then when you record the clean version of the track (without the box), record it into the same thing, so that your recording setup colours all of them consistently.

3. Record some pitch-shifted versions so we can see how it deal with aliasing. Maybe some downsampled stuff would be good too!

4. Post both clean and affected versions of each track, but without any indications of which is which. Then we vote on which of each pair is better. Maybe it's great for techno but bad for old recordings... or the other way 'round...

<3 u w/ all osc & lfo max'd
 

Going back to the digital domain does not make much sense, unless it's resampled at a much higher rate then the original (96KHz/24bits).
 

Well, the other analysis ideas on here are more focused on technical fidelity, so of course re-digitizing the signal would mess with them a bit. What I'm interested in is the less quantifiable esthetics of the sound transformation. So I think a high-fidelity, uncompressed recording published in a losslessly compressed format will still give us some idea of how effective it is for cleaning up the sound of compressed audio.
 

and let's not forget the most crucial test of all: will it blend?
 

How would it make up for missing information?

And

If artifacts are spikes, filters wil not help. Limiting the amount steps a signal can rise or fall, will. A bit.
 

I'm wondering if it could be used in the studio as a cheap and dirty saturator/compressor. Put some breaks and drum machine loops through it, increasing the input signal for each pass so that it distorts.