Home Theater Geeks 527 Transcript
Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.
Scott Wilkinson [00:00:00]:
In this episode of Home Theater Geeks, I discuss the results of a blind test conducted by Diyaudio.com and the results were surprising. Stay tuned. Podcasts you love from people you trust. This is Twit. Hey there, Scott Wilkinson here, the home theater geek. In this episode, I'm going to talk about an experiment that was conducted on a forum called diyaudio.com. the moderator there goes by the name of Pano, and he conducted a really interesting experiment that I wanted to share with you. He took 30 second clips from a bunch of different recordings from a wide variety of genres and he created four versions of those clips.
Scott Wilkinson [00:01:12]:
The first one was the CD audio version, a WAV file, uncompressed pcm. The next three he recorded in what's called a loop recording. So he played the file on his computer through a USB audio interface. It went out the outputs of the interface through three different media and then back into the input and recorded the resulting file in one of them. He used professional interview interconnect cables, regular copper cables, and looped the output of the audio interface back to the input, recorded the input into a new file. He then took the audio output from the audio interface and ran it through 20cm of wet mud, which you can see here, the desktop system that he used. And the right and left channels went through two different trays of mud. And in the second picture, you can see from an overhead view that, that they were two trays of mud.
Scott Wilkinson [00:02:42]:
And in the. That was the third version. Fourth version was the same thing. Taking the outputs of the USB interface and sending it through two bananas, which you can see here now, the. The interface to the mud and the bananas. He took some old speaker, sorry, some old microphone wire, and he soldered a penny onto the hot conductor of that, of that wire, of those two wires. And there you can see how he soldered these pennies onto the hot conductor of the wires. The shields of those wires he connected to the aluminum foil that was used to line the plastic containers where the mud and the bananas were.
Scott Wilkinson [00:03:43]:
He said he used the aluminum foil because without it, the noise was much too high to make a good test. And with the foil lining those containers, the noise floor was actually very low. So there you go. So he had these four files and they were played from the computer, passed through the different media and sent back into the computer and recorded. And the recording equipment was very common gear. A Dell laptop running Windows 10. The audio interface was an M Audio Fast Track Pro, 2 channel, 2 times 2 audio interface. He used some software called Gold Wave 7.0 for file trimming and and level correction.
Scott Wilkinson [00:04:39]:
He used the Reaper DAW or Digital Audio Workstation for simultaneous playback and recording and the REW RTA function Real Time Analysis for finding the optimum input and output levels for the M audio interface. Now the four versions of these files were uploaded into Google Drive folders in both FLAC lossless audio coding and wav uncompressed and labeled with a random code. So the participants would be able to identify each one separately, but they wouldn't know which was which. And so they could download whichever ones they wanted and play them on their system at home and try to determine which one was the original, which one was sent through the wire, which one was sent through the mud, and which one was sent through the banana. Now there's on diyaudio.com There's a nice thread about the test and we'll put a link in the show notes to that. There's also a link to the actual test itself. And Pano has upgraded the the, the test, the first iteration is as I described it. People would download the files and determine, listen to them and see which one they thought was which.
Scott Wilkinson [00:06:11]:
In this new version, it's what's called an ABX test and it's done entirely online using an audio engine that was custom developed by acid. Test IO does not rely on the browser for audio reproduct or audio rendering. It does it itself. And it's of a known high quality. So the only variable is the listener's system that they have connected to their computer. Also in an AB test you should know that you can choose whether or not you want to try to determine which one is the wire, which one is the mud and which one is the banana and what you will get. You choose one of those, let's say mud, just for example, it will play you the original file, it will then play you the MUD file and then it will play one of them and you don't know which. So you know which one's the original when it's played.
Scott Wilkinson [00:07:17]:
You know which one's the MUD file when it's played, but then it plays plays it a third time and you don't know which is of it of those two is being played. And you have to say is it the MUD file or is it the original file? And the new test keeps track and keeps score for you and, and whether or not you were correct or not. So it's a, it's a pretty cool test actually. Now I first became aware of it on A website called Tom's Hardware, which is really nice technical website. I go there often and we'll put a link in the show notes to that article as well. And that article gave us the bottom line. And the bottom line is the listeners who participated in this test didn't do any better than random chance. Overall, the participants couldn't tell the difference between the original file, the interconnect cable, the bananas and the mud.
Scott Wilkinson [00:08:27]:
According to the results published by Pano, there were only six correct responses out of 43 guesses. That's a 13.95% hit rate. Tom's Hardware went farther and calculated there's a 6.12% chance of getting the same or fewer correct answers if, if the participants had randomly guessed, which is only slightly above the 5% significance that statisticians use in such an analysis. Now, in other coverage that I've read of this experiment, there have been some thoughts that perhaps it's because there's water in mud and in the bananas that act as a good conductor. And that seems reasonable to me, possibly aided by minerals in those substances, especially salts. So, okay, that seems likely to me. Now what I'd love to see is the same experiment conducted with quote unquote normal interconnects, I.e. monoprice or something inexpensive, followed by really expensive cables that can cost thousands of dollars per meter.
Scott Wilkinson [00:09:54]:
So would people be able to tell the difference? I would bet money that they wouldn't. That again, the results would be no better than random chance, which would indicate to me that you don't need to spend gobs of money to get good sound. And I've been saying this my whole career, decades. I've been saying you don't need to spend a ton of money on really expensive exotic cables because you won't hear the difference. And such a test would would be a good indicator that that's the case. I hope maybe Pano considers it. But anyway, I thought this was a really interesting experiment and I am glad I was able to share it with you. Now if you have a question for me, send it on along to HTGWIT TV and I'll answer as many as I can right here on the show.
Scott Wilkinson [00:10:54]:
And if you have a home theater you're proud of, send me some pics. Maybe we'll get you on the show to talk about it. Until next time, geek out.
