• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Critical (Best) Music Tracks for Speaker and Room EQ Testing

unfortunately this does not bear up under the testing done at Harmann and documented extensively on this site (see post #1, this thread :) ). Generally complicated tracks with significant content at a wide range of frequencies seems to be the fastest way to evaluate speakers (when limiting the choice to actual music and not pink noise as recently mentioned). Based on blind listener testing it was shown that these more complicated samples were the best.
Once again: pink noise was shown to be the best.

There is no scientifcally established reason to use anything other than pink noise (on a single speaker) as a listening test.
 
Personally I find Tracy Chapman (and similar, eg Jennifer Warnes) to be far more annoying than pink noise, but that's me. :)

But any actual music will have constantly changing spectral balance (that's why it's music!), which has to be a hindrance in hearing errors in tonality, which is borne out by the Harman data.

Of course the Harman methodology assumes that tonal errors in speakers are the most imporant kind of error. The fact that tonal errors are much less noticeable in stereo than in mono is used to justify listening tests in mono, to make it easier to spot the tonal errors.

The issue I have with this is that it eliminates any possibility of hearing differences in soundstaging that result from the stereo effect that could be important to the listener, more important than the tonal errors that (according to Harman data) are much less noticeable in stereo.

For this I like to use out-of-phase pink noise, and listen for soundstage width.

It is a test that cannot be done in mono (obviously).
 
Personally I find Tracy Chapman (and similar, eg Jennifer Warnes) to be far more annoying than pink noise, but that's me. :)

But any actual music will have constantly changing spectral balance (that's why it's music!), which has to be a hindrance in hearing errors in tonality, which is borne out by the Harman data.

Of course the Harman methodology assumes that tonal errors in speakers are the most imporant kind of error. The fact that tonal errors are much less noticeable in stereo than in mono is used to justify listening tests in mono, to make it easier to spot the tonal errors.

The issue I have with this is that it eliminates any possibility of hearing differences in soundstaging that result from the stereo effect that could be important to the listener, more important than the tonal errors that (according to Harman data) are much less noticeable in stereo.

For this I like to use out-of-phase pink noise, and listen for soundstage width.

It is a test that cannot be done in mono (obviously).
Largely agreed. I find testing with one speaker and with two speakers is optimum.
 
I doubt 'complicated' was a criterion at all.

· Tracy Chapman, "Fast Car", Tracy Chapman
· Jennifer Warnes, "Bird on a Wire", Famous Blue Rain Coat
· James Taylor "That's Why I'm Here", “That’s Why I’m Here”
· Steely Dan “Cousin Dupree”, “ Two Against Nature”
· Paula Cole, “Tiger”,” This Fire”
· “Toy Soldier March”, Reference Recording

These are not 'complicated' works of music by any standard definition.
I'm not sure that "complicated" is the correct word but when you look at the large table on post 1, the top three (after pink noise) are "female pop rock", "orchestral", and "male pop rock", while the least useful music is either "solo..." or "jazz..." or "strings". So in other words, you could say that the best music is "wide spectrum sound". If you want to word-smith that into another word feel free. But if you don't agree that the songs above have "wide spectrum sound", then what would you say are the characteristics of the music that gave the best ability to distinguish speakers?
 
But any actual music will have constantly changing spectral balance (that's why it's music!), which has to be a hindrance in hearing errors in tonality, which is borne out by the Harman data.
Even pink noise has though, independently if its random or periodic pseudo random (which is the most correct), just the periods are shorter. ;)
 
I'm not sure that "complicated" is the correct word but when you look at the large table on post 1, the top three (after pink noise) are "female pop rock", "orchestral", and "male pop rock", while the least useful music is either "solo..." or "jazz..." or "strings". So in other words, you could say that the best music is "wide spectrum sound".

Well, yes, that is a better term..Amir called it 'rich spectrum' sound in post 1. Its musical complexity isn't a correlate. String quartets for example can be very musically complex but won't serve well for this purpose.
 
Well, yes, that is a better term..Amir called it 'rich spectrum' sound in post 1. Its musical complexity isn't a correlate. String quartets for example can be very musically complex but won't serve well for this purpose.

Of course. Spectral complexity rather than musical complexity is what is relevant in this discussion. I would not rely on Harman to be a definitive guide on how measured and subsequently graphically illustrated music translates into subjective measures of complexity. There is an almost infinite interplay between fundamental, harmonics, and room interactions in large scale classical music. Moreover the degree to which a track is revealing also heavily depends on the listener’s experience. The piano, being my instrument of choice, is most revealing to me. Hence it makes sense that many listeners, whose musical experience is limited in scope, would prefer the aforementioned glitter-image tracks and conflate pleasant sound with revealing attributes when in reality the simply like what they hear.

In certain cases the Harman chosen test tracks show that while technical expertise is high the musical is expertise and how complex music relates to their finished product, is lagging behind. Some employees, I am sure, are very musically experienced. Luckily, it seems like their methodology seems to work across the spectrum.

Lastly and hypothetically, if Harman put out two test tracks. One being “Fast Car” and the other A DGG remastering of Beethovens 5th by Karajan we all know which one will garner the most attention as a test tracks by main street. It should also be obvious which one is more complex, demanding, and revealing.
 
Of course. Spectral complexity rather than musical complexity is what is relevant in this discussion. I would not rely on Harman to be a definitive guide on how measured and subsequently graphically illustrated music translates into subjective measures of complexity. There is an almost infinite interplay between fundamental, harmonics, and room interactions in large scale classical music. Moreover the degree to which a track is revealing also heavily depends on the listener’s experience. The piano, being my instrument of choice, is most revealing to me. Hence it makes sense that many listeners, whose musical experience is limited in scope, would prefer the aforementioned glitter-image tracks and conflate pleasant sound with revealing attributes when in reality the simply like what they hear.

In certain cases the Harman chosen test tracks show that while technical expertise is high the musical is expertise and how complex music relates to their finished product, is lagging behind. Some employees, I am sure, are very musically experienced. Luckily, it seems like their methodology seems to work across the spectrum.

Lastly and hypothetically, if Harman put out two test tracks. One being “Fast Car” and the other A DGG remastering of Beethovens 5th by Karajan we all know which one will garner the most attention as a test tracks by main street. It should also be obvious which one is more complex, demanding, and revealing.
I don't know that I'd select Beethoven's 5th (any recording) as particularly good test fodder. In contrast, I'd pick a orchestral work with either a soloist (or soloists) or a work where the concert master plays a significant solo role, e.g., R-K's Scheherazade. Being able to track and follow the concert master (soloist) within a dense orchestral texture is a good test of linearity and channel matching. It's kind of like tracking a small boat in a storm-tossed ocean of pink noise. If one must have Beethoven and Karajan, I would suggest the Missa Solemnis, the Benedictus in particular. (same sort of challenge)

I'm not convinced of pink noise's efficacy as a tool for speaker evaluation. It's rather a blunt instrument without supporting instrumentation, e.g. REW. This may be one of those rare occasions where Pink Floyd might be preferable to pink noise.
 
Of course. Spectral complexity rather than musical complexity is what is relevant in this discussion. I would not rely on Harman to be a definitive guide on how measured and subsequently graphically illustrated music translates into subjective measures of complexity.

I don't recall Harman researchers trying to do any such translation. It's us doing it here. We need not belabour it. What matters is spectral content.

There is an almost infinite interplay between fundamental, harmonics, and room interactions in large scale classical music. Moreover the degree to which a track is revealing also heavily depends on the listener’s experience. The piano, being my instrument of choice, is most revealing to me.

Solo piano did relatively poorly in the Harman work. Perhaps it depends on how far the work ranges up and down 88 keys, and how often it does that during the work.

Hence it makes sense that many listeners, whose musical experience is limited in scope, would prefer the aforementioned glitter-image tracks and conflate pleasant sound with revealing attributes when in reality the simply like what they hear. In certain cases the Harman chosen test tracks show that while technical expertise is high the musical is expertise and how complex music relates to their finished product, is lagging behind. Some employees, I am sure, are very musically experienced. Luckily, it seems like their methodology seems to work across the spectrum.

Again, I doubt any of this mattered to them. And yes, trained listeners were involved in at least some of the studies. Have your read the papers?

Lastly and hypothetically, if Harman put out two test tracks. One being “Fast Car” and the other A DGG remastering of Beethovens 5th by Karajan we all know which one will garner the most attention as a test tracks by main street. It should also be obvious which one is more complex, demanding, and revealing.


The B5 is certainly among the most well-known orchestral works, so don't think it would be obviously discounted -- in fact orchestra without choir did rather well in the Harman work.
 
I don't know that I'd select Beethoven's 5th (any recording) as particularly good test fodder. In contrast, I'd pick a orchestral work with either a soloist (or soloists) or a work where the concert master plays a significant solo role, e.g., R-K's Scheherazade.

Are you guys actually looking at that graph in post 1? Or better yet, seeking out the papers?
 
Are you guys actually looking at that graph in post 1? Or better yet, seeking out the papers?
Agreed to all the above, specially in mono testing.
The unfortunate thing is that the art of depth in the recordings is somehow lost so testing it in stereo (obviously) so to test gear and room requires some of the ones which conclude it.

Reiner's Scheherazade for example shines at it. As many Lew Laytons works. ( I have excluded it from my test works though, I love it too much to wear it out) .
Anything can be useful if used for purpose.
 
Are you guys actually looking at that graph in post 1? Or better yet, seeking out the papers?
Yes to both. I still don't see much value in pink noise for subjective evaluation other than to differentiate (possibly even rank) multiple transducers under controlled conditions. The results shown in the papers are hardly surprising. Unfortunately, most real world situations in which end-users audition speakers, little resemble those described in the tests.

Pardon me if I misunderstood the intended direction of this thread, but its title suggested that this was a discussion of music tracks for subjective evaluation, not test signals, which I still believe, absent test equipment, have limited end-user utility.

If, however, the test regime is to include instrumentality (e.g. REW) I see little point in using music tracks, except to define test limits.
 
The lesson the Harman work teaches is that music that most closely approximates pink noise in terms of spectral content and spectral balance is probably the best for evaluating speaker quality.

Amir: "I have had the fortunate luck of sitting through a couple of their blind tests and can attest to the efficacy of the tracks used"
 
The lesson the Harman work teaches is that pink noise itself is the best signal for evaluating speaker quality.

Or at least those aspects of speaker quality that can be ascertained by listening to a single speaker.

Which, I continue to maintain, are not all aspects of speaker quality.

Please note that this claim is not contradicted by any Harman data.
 
It's not. Harman data shows that flaws that are perceptable in mono are less percepatble in stereo, and that relative rankings are usually (but not always) the same in mono and stereo. (See Figure 8.12 in Toole's book for an example of rankings that reverse in stereo vs mono.)

This is only possible if there is something about stereo sound that is not present in mono sound.

And it is (or should be) completely obvious that there are aspects of stereo sound that are not present in mono sound: sound localization being the primary one.
 
I assume this last one is The March of the Toy Soldiers from Tchaikovsky’s Nutcracker? I did a google search and nothing came up. Does someone have orchestra or conductor info?
From the article referenced:

DWS – Dallas Wind Symphony Conducted by Frederick Fennel / Tchaikovsky’s Parade of the Wooden Soldiers from Dallas Wind Symphony
Sampler/ Reference Recordings, 2009, B002AG2NN4.

Description: Classical with Brass/Wind Orchestra

The reference is actually incorrect in the article...since the composer is actually Léon Jessel.

Chris
 
Last edited:
I post the list of tracks Harman uses after literally decades of research to detect fidelity of speakers and room Auto Equalizers in another thread but thought it deserves its own thread. I have had the fortunate luck of sitting through a couple of their blind tests and can attest to the efficacy of the tracks used:

---------

AES Paper, The Subjective and Objective Evaluation of Room Correction Products
Sean E. Olive, John Jackson, Allan Devantier, David Hunt, and Sean M. Hess

JW - Jennifer Warnes, “Bird on a Wire”
TC - Tracy Chapman, “Fast Car”
JW - James Taylor, “That’s Why I’m Here”



AES Paper, A New Listener Training Software Application
Sean Olive, AES Fellow
Harman International Industries


· Tracy Chapman, "Fast Car", Tracy Chapman
· Jennifer Warnes, "Bird on a Wire", Famous Blue Rain Coat
· James Taylor "That's Why I'm Here", “That’s Why I’m Here”
· Steely Dan “Cousin Dupree”, “ Two Against Nature”
· Paula Cole, “Tiger”,” This Fire”
· “Toy Soldier March”, Reference Recording
· Pink Noise (uncorrelated)



AES Paper, Differences in Performance and Preference of Trained versus Untrained Listeners in Loudspeaker Tests: A Case Study*
Sean E. Olive, AES Fellow


James Taylor, “That’s Why I’m Here” from “That’s Why I’m Here,” Sony Records.
Little Feat, “Hangin’ on to the Good Times” from “Let It Roll,” Warner Brothers.
Tracy Chapman, “Fast Car” from “Tracy Chapman,” Elektra/Asylum Records.
Jennifer Warnes, “Bird on a Wire” from “Famous Blue Rain Coat,” Attic Records.


And this from a 1992 research at NRC on genre of music and its revealing nature in this regard:

Program+Influence+on+Listener+Performance.png


A bit about the science, the suitability of track is a matter of statistics. Colorations in speakers are only revealing if there is significant content/energy in that part of hearing spectrum. Rock music tends to have such rich spectrum. Classical music as a general rule does not. Hence the domination of rock/pop music in the top most critical list.

In both this space and audio compression with which I am intimately familiar with, high fidelity of the music recording is not an aid and if anything a distraction. A "pretty" sounding track sounds pretty on many systems because we are drawn to it by its good substance. Critical test clips on the other hand tend to be uninteresting and force you to pay attention to the task which is to analyze equipment with your ear.
Bit off topic but could it not be interesting to test FIR based DSP solutions versus IIR based DSP solutions from a measurment. point of view an subjectivly listening point of view an to do that with panel speakers that as far as i know phase coherent time alignt versus box speakers. Out come could be that let say both sort of speakers favours or FIR or IIR based DSP solutions. That could be a start to favour a certain DSP solutions based on the build of your speakers. Don't know if i make sense :facepalm:
 
My reference albums to check components when I'm buying a new one, either integrated amps (mine is a Marantz PM 6007 which I got three years ago), or more recently, that was July 2024, an SMSL D400 PRO and KEF'S Q550 speakers, are:
-Van Halen 's For Unlawful Carnal Knowledge, the 2005 Forever Young Japanese CD (tracked to analogue tape, mixed down to digital) or the 2023 96/28 Digital download (when Eddie Van Halen started to mix down to digital, he also did an analogue master, but bet IS that when the time came to do the final mix, he got both a digital recorder and an open reel Studer tape recorder plugged to his mixing board to capture that final mix to both anslogue and digital) from 2023, which used the full analogue master.
-Telarc's The Film Music of Jerry Goldsmith SACD, a native DSD recording.
-Analogue Production's The Planets SACD. The recording is from 1973, SACD release IS from 2018 if memory serves me well.
-Roxy Music Avalon, Japanese SHMSACD.
-E.T. The Extra-Terrestrial 20th Anniversary SACD, stereo layer.
 
Back
Top Bottom