Decibels, distortion, amplifiers and golden ears - Page 4  

I find it unfortunate that the Cheever thesis has found such promenence given its major flaws
what I really think of the Cheever thesis:
http://www.diyaudio.com/forums/showt...0&pagenumber=5

you may have to give up on the idea of the internet as an authoritative source in many scientific fields, JAES, ASA, in fact most peer reviewed journals don't put their content online for free, not that there isn't some good psychoacoustic info on the web:
fig. 3, 4 and their discussion in the body of the paper treat "harmonic" distortion audibility


even Stereophile occasionally hints at some sophistication in distortion audibility discussion:
http://stereophile.com/reference/406howard/index1.html


given the superior quality of of the above cited articles, (particularly the Czerwinski I cite a few posts down in the diyAudio thread) I'd really, really like to never see Cheever cited again

I agree that the distortion produced by the transducer is the weakest link. But I have to disagree with the rest of the analysis.

Firstly, you assume the test is accurate. I happen to think it's flawed.

1.The sound quality of the samples is far too low (I estimate 96kbps MP3 or worse). This inevitably results in lower SNR and resolution overall, among other things. Therefore the results in this test can't be assumed to be the valid with regard to full quality audio, such as CD or SACD.

2. The interface is badly designed. Concentrated listening produces better results with the eyes closed. There's no way to do the test with one's eyes closed (ie. with the keyboard).

3. The samples themselves are poorly selected. For example, for the Pop test, there is a loud clicking noise at the beginning of the sample. This click has the effect of precipitating a kind of change blindness in terms of audio, where the click serves as the disruption, thereby reducing the listener's awareness of the differences between the currently playing and previously played samples.

4. Looking at the results, it's likely the test software itself has a fault somewhere:



At -45dB, the results should not jump to almost the level of the median. The same jump at -45dB occurs in the Pop test as well.

I for one didn't post my results on that thread because it was obvious that it would lead to one-upmanship. There's nothing stopping anyone from posting any number they wish, and no way to verify the results. So, no reliable conclusions about head-fi can be drawn from those posts.

Simply invoking equations doesn't validate the reasoning behind them. Maybe if a better designed test was created - one in which we could seperate the head-fi results from the rest - we could get a better idea of how well we hear. Until then, just enjoy the music.

This is extremely true. Lowering resolution of the file that much generally reduces the frequencies we are talking of here, which at the same time becomes highly impractical. Majority of solidstate amplification will distort alot exclusively in the high frequency range (how high depends on the bandwidth, and higher quality amplifiers will exhibit alot less of that). When we are talking about low resolution files, we cannot simulate that distortion, and frankly my guess is that by not even covering majority of the frequency range, we are leaving out areas where human ears can be more sensitive. And .01% THD at 1 KHZ is one thing, now being able to say .01% THD at 20-20 is a different thing. Many cheaper amplifiers will skyrocket with measured distortion past 16 khz (hell, some op-amps I was testing for EE labs became useless around 12 khz).

Amen, that is a speculation, I do not know how sensitive ears are depending on the frequency, hopefully someone can correct me here.

Unfortunately the Cheever paper is about the only one which can be found for free on the web, which I suppose why it's referenced so often. The various audio engineering journals are by subscription only and require fees, though in some cases it's well worth the cost for access to the papers.

What's the upshot for the newly initiated here, who you say you are looking out for their interest? Would you be suggesting a Sony Diskman > Boosteroo > HD-580s ?

BTW, would you answer the question previously presented with reference to your medical insurance / health issues, and what that info was presented in support of ?

It is. It's also the reference that Hi-Finthen recycles whenever he wants to make a condescending remark about someone else's gear or hearing ability.

So, why are you posting on Head-Fi? Everyone is perfectly happy with their hobby. You obviously see things a lot different than everyone else. So, why continue to post?

No, not trying to be condescending about less costly gear, however this same old tired subject replays with the upshot that CDPs sound all the same and a Sony Diskman is the suggested CDP; As well as all amplifiers sound the same, may as well buy the least expensive etc... This conclusion is usually made with the same old test your hearing B.S. , when it is not for us to prove our hearing when it is that I do hear differences and improvements inline with the price paid for audio gear... Those who claim not to hear differences in fact are the ears which need testing, is all I'm saying

Where is Hciman77, he usually rolls in about now

I'll take quick and good thanks.