Thursday, June 20, 2013

RTA not working on Samsung Galaxy S4

Literally, the application named RTA does not work correctly on the Galaxy S4.  It is unable to show an interesting graph of pink noise response from my living room audio system (pictures below, sorry they got turned upside down somehow automatically).  Furthermore, the response shown doesn't vary much with orientation of the phone from the listening position.  Whether I point the front of the phone at the quite directional electrostatic speaker or not makes very little difference.  (That lack of directionality might be good for some things, like measuring actual noise exposure, but bad for measuring specific sounds or system responses.)  In case you are prepared to argue the problem is the ultimate SPL limit, which may well be around 90dBA as it is for other Android phones or so I have read*, I tried measuring at two different system SPL levels, which were around 75dBA and 65dBA, and they showed the same uninteresting picture, simply having an enormous peak around 8kHz and rolling off dramatically above and below that by 10's of dB's.  (Tthe left is 75dB upside down and the right is 65dB turned left.  And, strangely enough, it matters little which way the phone is turned, just as the pictures might suggest, but actually I had the phone laying flat in the same direction for both pictures.)

In contrast, the application named RTA on my iPhone G3 shows a very interesting and plausible frequency response of my living room system, broadly flat through the mid to highs but with a few peaks and dips below 600Hz, and then rolloff at the very lowest and highest frequencies.  I'm sure the rolloffs at the top and bottom are iPhone limitations, I have 16Hz-25kHz nearly flat response, but I think the rest is fairly accurate, or if not, it's simply the result of 4 and a half years of smartphone wear or careless measurement technique resulting in minor reflections.  The picture changes in interesting ways depending on orientation--notably the extreme highs get flatter when the bottom of the phone is pointed right at the speakers.  That's what I mean by an 'interesting' picture, when you make changes, you can see their effect reflected in the graph.

A different Android app running on the Galaxy S4 gives an equally plausible but different picture.  This is from AudioTool, the paid version:

Because AudioTool does show plausible response, I suspect the problem is that the default calibration of the free version of RTA isn't good for the Samsung Galaxy S4, and unfortunately you can't change the calibration on the free version.

(*I don't know why Android phones are limited to 90dB.  Tiny capsule microphones usually have a much higher limit, something like 110-140dB.  The iPhone SPL app can measure up to about 105dB, but that could be because the app is automatically compensating for known microphone compression.)

Thursday, June 13, 2013

Types of Listening

1) Annoyed listening (e.g., neighbor's dog barking).

2) Enjoyable Background Music Listening

This is fine, even if audiophiles consider it inferior to the point of condemning it altogether.  One can't always devote full attention to music, in fact, for most of us it's rare to have both the time and inclination together.  So having pleasant background music can be one of the greater joys in life, if done well.  I prefer the absence of talk altogether, but can listen to non-commercial radio also if there is minimal talk.  One of the radio programs I used to love to hate was "Adventures in Good Music" which seemed to be more talk than music.  Whole house music systems like the Sonos I use are helpful, as were auto-reversing tapes, and FM radio.  Internet radio based on lossy compression--I don't like it much.

Background music listening can't provide the intense pleasure that serious listening can.  But because it is something can can be done over a relatively larger proportion of one's time, and I'd say in principle the more the better, it can constitute a large part of audiophile enjoyment.  It should not constitute ALL of the enjoyment, for that would be saying that NO serious listening was being done.  But taking up about 95% of one's total system playing time, and providing about 50% of the total enjoyment, sounds like it would be about right for a working person with many other interests.

3) Serious Listening

You can't be doing much else while doing this.  Drinking a wine or beer, maybe, but even that's pushing it.  But rather than concentration, the goal is relaxation, relaxation into the music, not thinking about much else, letting thoughts go away generally, including thoughts about the qualities of music systems.

4) Comparative Listening

Listening to more than one system variation with the aim of determining differences and establishing relative or absolute quality levels.  This is most definitively "Comparative Listening" when done in one listening session or a planned series of sessons.  But long term comparison is possible also--so basically what determines whether listening is comparative or not is the degree to which thoughts about audible differences arise in the mind at the time.  So while I do little but long term comparison anymore...I'm still guilty of comparative listening, though to a lesser degree than some.

This is what audiophiles are known for.  However, contrary to some self-appointed categorizers, there is no essential need to do this to be 'an audiophile.'  As stated in an earlier post, 'an audiophile' need only be enthusiastic about music reproduction, not necessarily dedicated to the labor of improving audio systems through comparative listening, component substitution, and/or tweaking with accessories or special treatments.  I'd suspect that too much of those latter things, and one is well on the way to audio burnout.

Besides which, comparative listening as it is usually done should not be taken seriously.  At minimum, comparative listening should be done Double Blind, where even the experimenter doesn't know which equipment has been chosen.  And level matched.  Any results other than those from a level matched double blind test should not be taken seriously at all, ignored, or not even thought of if possible.  It is very easy to be wrong, our expecations and pre-conditioning (including, the previous playback) can easily dwarf actual equipment differences, and usually do IMO.  Unfortunately, most audiophiles I have met seem to lack the mental compartmentalization to doubt their own hearing sufficiently, in which case I'd suggest they'd be better off not doing comparative listening at all, because any judgements made are likely wrong, and after awhile a large set of wrong or random judgements may cascade to very wrong audio configuration or (just as likely) audiophilia nervosa, when one can't enjoy listening to anything anymore because of the feeling it needs improving somehow.

The reference is double blind testing (which includes ABX type) in a full trial of at least 30 trials, testing to a 95% confidence interval.  This is a lot of work and very hard to do well.  I have conducted 3 full blind tests in my lifetime on a well regarded audio guru.  Not one of those tests came close to the 95% confidence interval.  The audio guru chose the tests because he felt he could not fail on even one choice.

I make Comparative Listening a special category of listening is fundamentally flawed.  Thinking of the need to make comparative judgement not only takes out much of the inherent pleasure in listening to music, it makes itself impossible.  A correct comparative judgement can only be after the fact of listening.  It need not be long after, it could be mere seconds, and that may be the best for some kinds of tests.  But once comparative listening is the goal, it becomes impossible NOT to think about the ultimate conclusion, and such thinking makes it impossible to feel the inherent pleasure, which is what the comparative judgement should be all about.