h1

128 or 320 kbps – can you hear the difference?

March 11, 2009

audacity

Okay, we are going to do a quasi-scientific study here with the z=z community, to see if people can hear the difference between 128 kbps and 320 kbps audio streams. Go to mp3 or not, listen to the two sound samples, decide—or guess—which one is at the higher bitrate, and then post your results in the comments, along with any ancillary information (like whether you used speakers or headphones). I couldn’t tell the two samples apart, at least not on my laptop speakers, but I had a 50-50 shot at getting it right, and I did. Let’s see if we can get enough numbers to exclude (or confirm) randomness.

UPDATE (Friday, 10:23 am PDT): We are up to 6 people who got it right and 5 people who got it wrong. I’d love to get some more datapoints. Please share this link and ask people to comment or to send me an e-mail or a tweet.

UPDATE (Friday, 10:49 am PDT): A poll! This is much easier. If you’ve already responded in the comments, please do not vote in the poll.

[via Music Machinery]

44 comments

  1. I guessed correctly but I’m not sure if I really “knew” the answer. I thought I heard a slight difference but it may have just been my brain making shit up.

    For further statistics, I used a set of Bose Triport headphones (I’d say pretty decent/accurate sound quality) but I only have about 20% hearing in my right ear so that might negate everything mentioned above.

    That was pretty fun. I always wondered whether or not I could tell the difference. Regardless, though, I’ll still rip everything at 320. :)


  2. Well, even if everyone individually says, “I couldn’t really tell,” it’s still possible that, cumulatively, we’ll get it right more often than chance. We’ll have to see how it goes.


  3. I got it wrong using the built-in speakers on my MacBook, which are notoriously not very good.


  4. I was right! I thought it was pretty obvious on close listening; low-quality mp3’s have a characteristic sort of higher-register ringing. It really shows up in these samples in the first extended vocal note.

    Maybe I cheated, though; I used reasonably nice headphones (Sony MDR-V600) with an external sound card. I doubt I would have caught it in casual listening; I would test with speakers but it’s late! In any case, I haven’t found myself noticing mp3 noise at any rates higher than 128kbps; I usually encode at 192.


  5. I don’t think using nice headphones counts as cheating! But I wonder if you should follow Justin’s example and rip things at 320.

    3 to 1 and counting!


  6. I got it right using an 7.1 Soundsystem and external soundcard, but didn’t really heard the difference though. Lucky me.


  7. Well i will say the clip is not a good example , they should post a “band” whether it be rock or jazz or whatever , this clip is mostly human voice , it get’s really noticed on cymbals & such .
    But i did pick out the right one with no problems at all .


  8. Result: FAIL.
    Gear: Ears, M-Audio Mobile Pre, Sennheiser HD280 Pro’s, thousands of wasted dollars on a music industry degree.

    I have no excuse, I went back and forth trying to decide which sonic space I liked better and I ultimately guessed B. It seemed a little more open, but I couldn’t really pick any of the flattening or distortion that I expected to hear in either of the clips.

    I do also side with DMK in the call for a clip with more interaction between the sonic ranges. I’d personally prefer a longer clip to get a better feel for the sound of the piece, but maybe I’m just backpedaling to excuse my embarrassing audio fail.


  9. I used my garbage laptop speakers figuring that I would not be able to tell the difference because I would have argued that unless you have decent speakers, there’s no way you could tell. I found that the characteristic that is most affected by bitrate – rapid high-frequency envelope changes – was clearly different between the two. I didn’t particularly think one sounded better, so I had a hard time choosing. If you listen to the castanets (I think those are castanets?), they sound really different between the clips, even on my bad speakers. In the end, I chose the clip where they were “quieter” sounding as I thought this was maybe better frequency resolution. This was probably overthinking as I made the wrong choice. I am guessing that there was some frequency smearing that resulted in the high-frequency portion sounding quieter. Anyway, my point, I guess, is that there was a big difference, even on my bad speakers, but I had trouble deciding that one was better on my bad speakers.

    I’ve babbled a lot, but I think this is kind of the wrong example for discussing bitrate. Typically, (current especially) rock music lacks dynamic and would likely be difficult to differentiate in a test like this, but music with a lot of pauses and rapid changes in envelope is not well encoded in MP3 format. If you listen to any symphony, for example, it sounds terrible as an MP3.


  10. Wrong! I drove myself nuts going back and forth between the two. They sounded so similar to me, even on good headphones. I’m in agreement with dmk that something with cymbals – or even piano – would have made it more obvious.


  11. I got it and felt reasonably confident that it wasn’t just chance (used 2.1 speakers). I’d love to try more, though, and see if I do better than random chance on a larger sample.


  12. Yeah, Jon – there’s two separate things we could be testing.

    Can any individual discriminate between the two bitrates? For this, you’d want to present that person with a number of music samples to see if they could select the correct one at a rate higher than chance, or;

    As a population, can we discriminate between the two bitrates? This is what we are investigating; it doesn’t say anything about whether a given individual can tell the difference.


  13. I think this may actually be inspiring a real study. I looked around and only found a few, none of which was very thorough. I’ll have to do some more searching, but I have a couple of students really interested in doing some music experiments.


  14. Oh, I failed to say above, but perhaps this is a collaborative effort…


  15. I would so love to see real data on this!


  16. Comment explosion!


  17. Comment explosion – whoo!


  18. I’m clearly days late, but found this relevant: “Young People Prefer Sizzle Sounds of mp3 Format

    (Incidentally, I could detect a slight difference on earbuds but got it wrong. Maybe I’m hiss-inclined?)


    • Thanks, Katie – there’s a post on just that topic lined up. Stay tuned!


  19. [...] Interested in whether you prefer 192 or 320 kbps MP3s? Check out an earlier z=z post here. [...]


  20. Got it right, but I just bought new Shure headphones and I really had to think about it while listening. At first I actually liked the lower sound quality, but then I noticed the same thing Tim Smith did with the fuzziness on the first long note. Highly doubtful I would notice a difference on speakers or in a car. I will stick with my low quality storage purely for storage reasons… too bad, I would love to say I prefer quality over quantity


  21. Wow, I’m never ripping at 192 again. I didn’t even have to listen to more than a couple of seconds to determine which was which. I’d always sort of wondered if there was a difference, and now I know for sure. I used headphones of pretty mediocre quality. Even for people who can’t tell a difference consciously, there could be a difference in the subconscious enjoyment of the music. Thanks for this post.


  22. [...] sound quality in music. The first and second posts are up.  If you haven’t yet, you can test yourself to see if you can hear the difference between MP3s at 128 and 320 [...]


  23. Got it right, but my choise was definitely affected by some of the other comments I read. Extremely slight difference through my Koss portaPro-headphones. Need to poins out that in more complex songs the difference is way more evident I think though, naturally because all the music has to “fit” in the bitrate. In the clips there wasn´t much anything. 320 kbps or lossless is my way, even though I failed miserably on a blind-test I made for myself using two tracks, same song but one was 320 kbps and other was lossless. There was perhaps a slight difference though.


  24. well i did tell it in one shot . it was easy for me
    i m using intex headphone (very cheap around 5 U$)still. i think i m sensitive to music


  25. [...] 128 or 320 kbps – can you hear the difference? Test yourself to see if you can distinguish between low and high bitrate MP3s. [...]


  26. I got it correct but that’s based on the load times of each piece =S


  27. The difference on .mp3s is consequential, especially since most of us are listening to music in the car, on iPod, etc. But burn the tracks to a CD, and the difference is STAGGERING. 256kbps is CD quality. 224kbps is decent, but anything under that is pretty dreadful.


  28. Why not just instead of A & B have A & many B’s so only one of them is the better one then there would be less chance of hitting the right one by luck


    • There’s no question that this isn’t a well-designed experiment, Jon. If you read up in the comments, you’ll notice that some of us talked about how this could be redesigned to more accurately measure people’s ability to discriminate between the sounds.


  29. interesting test … i got it wrong. i can tell the difference between 128 and 320, but sometimes it takes awhile … and it’s most noticable in the drums, cymbals and high tones, which this track doesn’t have much of … would be interested to see it with a bigger range of (longer) tracks


  30. got it wrong, tried on both 2.1 speakers, and headphones.


  31. I did hear the difference, but it was very very subtle, enough to not hear the difference when not trying to detect it. I wore headphones and closed my eyes to listen.


  32. Was fairly clear to me, but i suggest for a proper test try using a full instrument heavy rock track, this track was obviously selected to make it harder to tell the difference but you hear the difference very clearly when the percussion kicks in.


  33. I got it wrong lol. I couldn’t tell the difference but if I was listening to the whole song or some music I liked..maybe then I could tell the difference lol.


  34. [...] (With an average speed of 22 Mbit/s for an 802.11g network, you can download a 5 min song at a bitrate of 128 kbit/s in just a couple of seconds, so you don’t need to spend long in a hotspot.) But having a [...]


  35. I got it right. It was a noticeable difference. You could hear the singer’s vibrato clearly on the 320 but was kind of dull on the 128


  36. got it wrong, was using v-moda’s vibe headphones at the time

    i think maybe the reason so many of us got it wrong was because none of us knew what the “right” track sounded like; without a real-life example to compare them against, i think the majority of us who were borderline (like i was) picked the track they liked better, not the one with higher quality


  37. I had very little trouble hearing the difference somehow. However, I would have a hard time describing what the difference was exactly. Maybe it was more resonant. I mean, it sounded like 320 kbps was smoother. Anyway, when I listened to the unknown clip, there was no question that it was 128 kbps. I should probably listen to more clips to make sure, but I don’t think I would need a statistical test to prove the difference to myself. Still, thanks for putting this up to a test; it’s an interesting question.

    (I was using UE triple.fi 10 earbuds)


  38. Admittedly, the difference wasn’t night and day for me. After 2 listens of each A & Band of the sample, I got it right, mostly because of the final chord (i.e. in the 320 kbps sample, the trumpet—or cornet or whatever brass it was—punches the last note very cleanly, whereas this is muddled into the sound in the 128 kbps). Also, in the 128kbps, it is harder to distinguish which instrument or section was sustaining the leading note. There weren’t any clear differences in vibrato for me, although I suspect that after hearing the higher quality sample and recognizing there was vibrato, I would’ve “expected” it (or my brain filled it in for me).

    It would have been better if we were not told which one was of which bitrate, if we were introduced the lower bitrate sample first, and if we were asked which one was of higher quality on top of which one matched the third sample.

    You guys should try some of Joe Hisaishi’s work in FLAC (you can actually hear sheet music being turned, and equipment being adjusted in One summer day from Spirited away OST) as opposed to the regular CD quality rip (192 kbps ish is it?).

    Of interest, I’m using ATH-M50 (supposedly a DJ monitor set), Macbook source, no amplifier. Anyhow, thanks for the share!


  39. I heard good with my old altec lansing speaker.The high “trashing note(i think its cymbals and orchestra hits)” can give better result by listening the sharpness.


  40. Nailed it.

    I could definitely tell the difference, but I usually only encode 192 anyway. Oh well.

    I used some clock/radio speakers, so apparently they’re decent. Nice!

    I also work in radio and do audio editing and such for us, so that might be why I can tell. Who knows.


  41. Have AKG 702s, a DAC, and an amp…couldn’t tell the difference for the life of me


  42. I was right 5 times!
    I wonder if there’s a difference between 192k & 320k..



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

%d bloggers like this: