THD measurements show sound quality?

[Azzkiller]

http://www.gedlee.com/downloads/phan08.wav
http://www.gedlee.com/downloads/phan09.wav
http://www.gedlee.com/downloads/phan03.wav
http://www.gedlee.com/downloads/phantom.wav

I see a lot of people comparing the THD specs on recievers, amplifiers or soundcards in order to compare one another as to which sounds better. For those who arent familiar, THD stands for total harmonic distortion, so a smaller number is better right? Absolutely. However, the problem is that it requires a rather high level of THD to become audible. Another problem is that the THD spec does not encompass all types of audio distortion.

If you guys don't mind, I'd like you to download all the above files and listen to them. They are modified to have varying levels of distortion.

My question for you is which of the files sound like they have a lot of distortion? Which files sound like the original? Can you hear the difference between them?

[zabomb4163]

[url] Can you hear the difference between them?

08 - sounds like it was recorded off the radio. notes are too soft.
09 - sounds like it is raining in the background
03 - almost sounds like an echo. instruments sound a little muddy and the notes are again soft.

[stratcatprowlin]

The first and last sounded very similar.
The second had a lot of noise in it and the 3rd got distorted at the when it reached a high point in the recording level.

[stratcatprowlin]

Oh, I forgot to answer the last question!

Its impossible for me to tell which is the original because I am not familiar with the original.It could be the one with all the noise for all I know and could have been cleaned up digitally.

[Azzkiller]

The phantom is the original, sorry.

[stratcatprowlin]

Cool test! btw what was the difference in the first and the original?

[Azzkiller]

Cool test! btw what was the difference in the first and the original?

SPOILER, DO NOT READ UNTIL YOU HAVE LISTENED TO THEM ALL!
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
http://www.gedlee.com/distortion_perception.htm The first file as you can see on Dr. Earl Geddes site has 9.6 percent THD and 22.4 percent IMD. Yet it sounds (at least in all the people I know who have done blind tests) identical to the reference. While the second file sounded terrible but only has .1 and .3 THD and IMD respectively.

The point of the thread is, dont look at the THD figure on a product to determine how good it will sound. Be it .001 percent or 1 percent, its almost certainly going to be inaudible. The order of the harmonic distortion, the frequency, and ESPECIALLY the type of distortion (as evidenced in Phan09 with .1% THD) all play a large factor on whether it will be audible or not.

[zabomb4163]

receiver buying guide

[johan851]

Phan 09 sounds awful. There's tons of fuzz in the background. The main melody doesn't sound different from the original - it's as if fuzz was added, but the rest stayed the same. There were occasional burps though.

Phan 08 doesn't sound too bad. Some of the background detail is lost, maybe a hair softer in general.

Phan 03 sounds like the actual melody (or whatever you want to call it) has become distorted. It clips on the highs and crackles and whatnot.

[johan851]

Ah...interesting.

[magick_man]

8 is ok but doesnt sound very good
9 is very bad, sounds like fm with weak reception
3 is just horrible

~Magick_Man~

[stan03]

8 sounds pretty close to the origional.
9 sounds alright, but there is horrible static on the background
3 just sucks the treble and bass sounds fake and crackly.

[Ryan T]

The point of the thread is, dont look at the THD figure on a product to determine how good it will sound. Be it .001 percent or 1 percent, its almost certainly going to be inaudible. The order of the harmonic distortion, the frequency, and ESPECIALLY the type of distortion (as evidenced in Phan09 with .1% THD) all play a large factor on whether it will be audible or not.

I agree that THD isnt the most useful in the way of sound quality but I think there useful with amplifiers, and home theater receivers, etc. Lets say you have two different receivers one states that it can output 25 watts of power and .07% THD the second can output 25 watts of power at 1% THD. That tells me that the first receiver has more CLEAN power. Meaning if you bumped up the THD levels it could output more watts at the cost of the THD levels increasing. Most of the cheaper receivers claim to have a ton of power but they rate them at high THD levels. Then you jump to middle class stuff and its usually rated at low THD levels. From my personal experiance the cheapo receivers i've had vs the middle class ones there was a BIG difference in power. What I always look for in a receiver is low distortion full bandwidth power with all channels drivin. A lot of companys (especially low end ones) rate one channel with "high" distortion at 1Khz. That is really worthless in telling how much actualy power a receiver has. So I wouldnt discard THD completly. I do think that measurments do get relied at a little too much though. The best way to see pick out speakers and such is realy to just go listen for yourself not just stare at some spec sheet.


Ryan

[Azzkiller]

I agree that THD isnt the most useful in the way of sound quality but I think there useful with amplifiers, and home theater receivers, etc. Lets say you have two different receivers one states that it can output 25 watts of power and .07% THD the second can output 25 watts of power at 1% THD. That tells me that the first receiver has more CLEAN power. Meaning if you bumped up the THD levels it could output more watts at the cost of the THD levels increasing. Most of the cheaper receivers claim to have a ton of power but they rate them at high THD levels. Then you jump to middle class stuff and its usually rated at low THD levels. From my personal experiance the cheapo receivers i've had vs the middle class ones there was a BIG difference in power. What I always look for in a receiver is low distortion full bandwidth power with all channels drivin. A lot of companys (especially low end ones) rate one channel with "high" distortion at 1Khz. That is really worthless in telling how much actualy power a receiver has. So I wouldnt discard THD completly. I do think that measurments do get relied at a little too much though. The best way to see pick out speakers and such is realy to just go listen for yourself not just stare at some spec sheet.


Ryan

I never said to discard THD. Infact, I agree totally that it is usefull in measureing maximum output power. However, it irks me when I see people tell others that one reciever is better than another because of the THD rating they see is lower on one. Now, depending on how the manufacturer tests, the reciever with equal power output yet higher distortion may in reality be more powerfull than the other. We need much more than just 100w into 4ohms 20hz - 20khz 1% THD to determine how the amplifier's output is being tested.

[stratcatprowlin]

Azzkiller,you obviously have made your point,A very good one at that.

Now I cant help but wonder,What shall we look for then?
And do manufacturers supply this information?
If they do not,what are we to do?
Are we left in the dark to decide by price,weight?material?name?and THd? hehe.

[Azzkiller]

Azzkiller,you obviously have made your point,A very good one at that.

Now I cant help but wonder,What shall we look for then?
And do manufacturers supply this information?
If they do not,what are we to do?
Are we left in the dark to decide by price,weight?material?name?and THd? hehe.

You would have to test the reciever yourself to get accurate information. Like I said, the testing methods differ (even between the same brand).

I would never judge a reciever by its price either. Weight sometimes can distinguish cheap recievers from expensive ones, but look at the new Panasonic recievers with switching power supplies. They get rave reviews, yet are the size of dvd players and very light as well.

If the manufacturer uses the FTC power output rating, you can use it to determine the true power output. However I dont think Ive seen any cheaper recievers use the FTC rating except for in stereo mode.

Honestly name is one of the main buying factors I use. It's obviously not the best though. Im really partial to HK's recievers, I dunno I just really like them. And the funny thing is, they are usually rated to deliver half the power or less than the competiton. Go figure.

I suggest a person buys a reciever in this fashion. Figure out what features you need to have in a reciever. Then set a price ceiling. Then go to as many places as you can find, judge the asthetics, and find the one that will do everything you need. Then buy the unit that costs the least that fits all your requirements.

[stratcatprowlin]

Yeah I guess thats the only thing we can do.

[JigPu]

So what was the difference in the THD and IMD introduced between 8 and 9 (other than amount)? I can't conciously make out any difference between the original and 8 (though subconciously for some reason the original sounds better... but that may just be because my brain knows it's the original :D). 9 and 3 suck though.

So what was the difference?
JigPu

[Azzkiller]

So what was the difference in the THD and IMD introduced between 8 and 9 (other than amount)? I can't conciously make out any difference between the original and 8 (though subconciously for some reason the original sounds better... but that may just be because my brain knows it's the original :D). 9 and 3 suck though.

So what was the difference?
JigPu

I talked to an online buddy about this and he said that the THD on #8 was primarily second order. Our ears have more difficulty hearing the distortion the lower the order it is. In other words, 9% 2nd order harmonic distortion may be inaudible, but 3% 7th order harmonic distortion may be audible. The THD spec doesnt account for this though. Thats why people like tube amps so much. Many have terrible THD figures, but when you consider that a lot of it is second order, it may be much harder to detect. The second order distortion sounds pleasing to many people if its in high amounts. As for the IMD, Im not really sure. My guess is that the IM distortion is in a frequency range harder to hear.

#9 sounds so bad because the distortion you hear is crossover distortion. Even in miniscule amounts, you can stiill hear crossover distortion. So even though its THD and IMD were so low, you still heard nasty audible errors.

Yes, they should be totally equal in terms of sound quality (#8 and original that is) so you probably were imagining the difference. However, if you want to know with a better confidence, you can download ABX software which will allow a blind comparison between the two. Let me know if you want a link to the software and an explaination on how to set it up.

[stratcatprowlin]

I 'd be interested in that software you mentioned,is it difficult to set up?

[Azzkiller]

I 'd be interested in that software you mentioned,is it difficult to set up?

Well the one i downloaded was. I had a different version that was much easier to use, but Im still trying to find it. It's late, but hopefully tommorow sometime I'll find the easy one.

But I HIGHLY reccomend it, especially if you would like to test to see what bitrate of mp3 compression is transparent to your ears, or to analyze distortion and other things.

[stratcatprowlin]

Cool, ill check back tomorrow then TX.

[Azzkiller]

http://www.kikeg.arrakis.es/winabx/winabx.zip
There is the ABX comparator I was talking about.

Download the zip, then extract the .exe. Run the .exe and select phan08 and phantom as the two files (it doesnt matter which order). Now, on the righthand side of the screen you see several options under "Mode". Click the "ABX" button. Also click "disable quick switch". Now, click the "play A" button to get a feel for what it sounds like. Then click the "play B" button to get a feel for what it sounds like. After that, click "play x" (at this point the testing has begun). If the song you are playing sounds like "A" then click the "X is A" button. And vise versa if it is B. Then click "next trial", and then "play x" again to start the next trial. You can watch your progress near the bottom of the screen.

I reccomend doing as many trials possible, even if you get them wrong once or twice. usually 10 trials is a good number, but the more the better. :)

Lemme know if anything isnt clear.