Export thread

This is a neat test on NPR's site. I got 4 out of 6 right. You listen to 3 exact same clips from the exact same song, but each is compressed differently: 128 kbps, 320 kbps, and uncompressed .wav. Then you choose the one you think is the uncompressed wav file.

http://www.npr.org/sections/therecord/2015/06/02/411473508/how-well-can-you-hear-audio-quality

1/6.

I want to see the stats on this test when it's been running a while. My research specialty is signal processing, and I know how the MP3 format works. It is, as the article mentions, pretty impressive. Even lost information isn't totally lost like it is in JPEG or MPEG compression. It's more like it's converted.

I can always pick out which one is the 128mbit sample, but it's hard to tell 320 from uncompressed. Usually the most notable difference is in the high frequency bits... hihats/cymbals and such are usually a dead giveaway.

2/6

I did better once I started clicking the one that sounded the worst.

Didn't even bother taking the test. They all sound the exact same to me.

I was hoping @fade had seen this.
The only one where I didn't pick either the 320 or the WAV was, surprisingly enough, Tom's Diner, which was 100% vocal + reverb.
This was through a pair of earbuds on a stock 16/44.1kHz sound card, I'm wondering whether they would've sounded different through my 24/192 card + ATH M40fs at home.

--Patrick

4/6 on cheap sony plugged into my pc. the other 2 were the 320. There was a slight difference but not very noticeable.

If you're considering trying it, Gas has the right idea. The highs (or lack thereof) are a giveaway.

--Patrick

I was listening specifically for highs and still couldn't really pick it out. Worth noting that listening to quiet highs specifically should be more telling. Loud highs are well represented by the algorithm.

For me, the hint seemed to be the clarity of quieter sounds. Not necessarily lower or higher. Anyway, I thought it was neat and I'd be interested to see what people can generally hear and differentiate.

Clarity of highs was one tell, but depth of range was another. The samples with higher compression felt "flatter" with noticeably less of a difference between quiet and loud passages.

--Patrick

PatrThom said:

Clarity of highs was one tell, but depth of range was another. The samples with higher compression felt "flatter" with noticeably less of a difference between quiet and loud passages.

--Patrick

Click to expand...

Does data compression affect dynamic range compression?

I did terrible, but I also can't hear a lick above 17K.

Also, if you really want to test yourself, I saw someone post this test in one of the comments.

drifter said:

Does data compression affect dynamic range compression?

Click to expand...

...sorta? It's the "bits" that affect dynamic range. 8bit v. 16bit v. 24bit. WAVs are usually 16bit/44.1kHz (at least, that's what "CD-quality audio" means).
@fade and I discussed this sort of thing a while back.

--Patrick

3/6 through crappy dollar store headphones. The 3 I missed were recorded in 320K

The uncompressed tunes tended to have more "depth" of sound. There was just more to them.

But it was really subtle, and I had to listen hard to hear the differences. I didn't think the uncompressed music was significantly better than the compressed tunes. I might've had a different experience with better headphones/speakers/whatever.

What this test has taught me is that 320kbps sounds way better than uncompressed.

At least to me... since that's what every one of my choices were.

1/6.

I got 1/6 when I did it on my standard TV speakers (HDMI audio out from the PC), the second time I did it with my USB headphones and the difference became very apparent to me. The only one that threw me was the Neil Young one.[DOUBLEPOST=1433389777,1433389709][/DOUBLEPOST]It also doesn't hurt that I used to be a DJ and had to run a sound board.

Bowielee said:

It also doesn't hurt that I used to be a DJ and had to run a sound board.

Click to expand...

I also attribute my acuity to my profession. Before I got into radio production, I couldn't give a shit about mp3 compression.

Now, I can always tell when something is 128k and it grates. Lowest I can put up with is 192k, but I prefer 320.

6/6 but I guessed on the classical one. It was between two of them and I couldn't easily discern them. The first four were really easy, I thought.

0 - I listened without headphones and almost always picked the 320 kpbs

listened again tonight on speakers. 5 out of 6.

1/6. Chose the 320 kbps every other time.