Double blind test 128Kbps vs lossless? I'll be amazed if you can tell much difference

[chengbin]

And please, guys, stay on topic. 

 

[mralexosborn]

You know, it really depends on the song. With SOME songs I can hear the difference between 320 kbps MP3 and FLAC. This does not happen often, but I am just trying to make the point it is not about quality but about what is necessary for the particular song.

 

[JBThazard]

I used my Senn HD598, and I didn't read the entire thread to find out which is which, but to me the 2nd track sounded noticeably clearer (and even a bit louder) than the 1st so I'm pretty confident the second track was lossless. Can someone please confirm for me?

 

[JBThazard]

Double post, but I forgot to add this.
 
When it comes to differentiating lossless and lossy formats, I *usually* can tell the difference. My Senns are open-dyanmic but I use them portably since my 2nd best pair, the Philips Stretch, sounded much worse and I couldn't bear it anymore. I used to use an Ipod Classic with LOD through a PA2V2, but the Ipod broke so now I use a Cowon J3 unamped. Even on a bus I can tell minute but noticeable differences with a lot of the music I listen to (mostly jazz and rock/metal). But just today I compared John Coltrane's "My Favorite Things" and ABX'd myself between FLAC and 192. Couldn't tell them apart. I mean, I guessed correctly, but the fact that I had to guess means I didn't know.
 
I think a lot more factors than most people think goes into being able to differentiate lossless and lossy (some of these are obvious):
 
-Your PMP (hdd based ones like Ipods have trouble processing high bit rates and can break down faster, as opposed to flash based hardware. There must be people out there who disagree but I don't recommend listening to a lot of ALAC on the go....to be safe)
-your cans (HD or now, open or closed, impedence, etc etc)
-amp or unamped?
-the production/mixing/mastering quality and DR (dynamic range) of the particular song/music you listen to
-are your ears dirty/haggard (especially for older people)
-the genre of music itself (simple stuff like pop/hiphop won't benefit so much, complicated music that justifies higher bit rates probably will)
-Do your headphones physically react differently between lossless and lossy? High bit rates push more energy (no matter the cans' impedence) and my Senns will, in addition to providing better soundstage (usually), give a pleasant vibration with high bit rate songs (generally 600-700kbps +), which mimics real instruments. Creates immersion and adds to the whole "sounds like it's right next to you" feeling.....And yes, it's extremely different sounding than brickwalling.
 
I know a lot of this is common knowledge to Head Fi'ers at this point but I've read a lot of lossy vs lossess threads and most people (even the ones who claim to tell the difference) don't seem to mention or consider some factors I just listed. 
 

 

[Head Injury]

Quote:

-Do your headphones physically react differently between lossless and lossy? High bit rates push more energy (no matter the cans' impedence)

 
What?

 

[JBThazard]

So I guess that's a no....
 
I listened to Opeth's Heritage in 320 and then soon after listened to it in FLAC (good example of an album that benefits well in lossless). In 320 my headphones were not doing anything besides playing the music. No noticeable vibrations, weaker soundstage, but that was not the case in FLAC.
 
Though my cans are Open....and Sennheiser has their cornily named EAR (eargonomic acoustic refinement) which may play a role in it. I HIGHLY doubt even IEMs would have this effect....
Quote:

 
What?

 

 

[Head Injury]

Blind test?
 
Lossy compression has next to no effect on the amplitude of the analog signal. You aren't going to get physical vibration with one and not the other unless something is seriously wrong somewhere. Even then I have no idea what would be wrong 

 
Your "high bitrate has more energy" thing is really misleading. It makes no sense. More bits doesn't make the analog signal stronger by any means. What exactly did you mean to imply?

 

[ChipnDalebowl]

I don't plan on reading all pages to tell if its said which one was 128kbps and which was lossless. I couldn't hear any stark differences to really care which was which. Just shows how far 128kbps has come. Plus my hearing quits out at about 17khz on tones, and even then they weren't very loud, so for real music I probably can't hear much beyond 16khz which is the general cutoff for a modern LAME encoded mp3 around that bitrate. You can PM me which it was if it hasn't already been spoiled which track was what.

 

[JBThazard]

Maybe my wording was lazy (I should have said something like REQUIRES more energy, not has more, sorry), but there's no denying that FLAC (for me, 700-1200kbps) takes more energy to play than 320, right? A HD will require more energy to play Song X in FLAC then it would to play Song X in 320, and sometimes I can physically hear the difference between them. Other people have noticed this effect I'm talking about, so I know it isn't some mastering flaw with the music or transducer failure with my cans...
 
EDIT: I did a blind test comparing Opeth's "Haxprocess" between 320 and FLAC (616kbps) and phyisically, it was very distinguishable (basic SQ, however, not as much but that's obviously a different argument).
 
Quote:

Blind test?
 
Lossy compression has next to no effect on the amplitude of the analog signal. You aren't going to get physical vibration with one and not the other unless something is seriously wrong somewhere. Even then I have no idea what would be wrong 

 
Your "high bitrate has more energy" thing is really misleading. It makes no sense. More bits doesn't make the analog signal stronger by any means. What exactly did you mean to imply?

 

 

[Head Injury]

Quote:

Maybe my wording was lazy (I should have said something like REQUIRES more energy, not has more, sorry), but there's no denying that FLAC (for me, 700-1200kbps) takes more energy to play than 320, right? A HD will require more energy to play Song X in FLAC then it would to play Song X in 320, and sometimes I can physically hear the difference between them. Other people have noticed this effect I'm talking about, so I know it isn't some mastering flaw with the music or transducer failure with my cans...
 
EDIT: I did a blind test comparing Opeth's "Haxprocess" between 320 and FLAC (616kbps) and phyisically, it was very distinguishable (basic SQ, however, not as much but that's obviously a different argument). 

No, there's plenty to deny there. It depends on the player. A hard drive-based player might spend more energy reading the larger FLAC file because a lot of their power goes to the hard drive and larger files require more seeks. Flash-based players might spend more energy decoding the MP3 because its compression is more complex for the CPU.
 
That said, neither of those have any effect on the "energy" of the analog signal. You can't hear the energy required to decode a file, sorry.
 
What were your blind test results and methodology?

 

[TMRaven]

Lossy does effect dynamic range which could lead to the perception of more energy in playback.

 

[Head Injury]

Quote:

Lossy does effect dynamic range which could lead to the perception of more energy in playback.

Could you link me to information on this? Every lossy track I've ripped has had almost identical ReplayGain volume levels to the original lossless, within 0.1dB.

 

[JBThazard]

Well lossy does affect DR but only marginally, and usually not noticeably. What I'm talking about is not related to the soundstage (which didn't really differ much between 320 and FLAC), I'm speaking more about how "bouncy" my HD598s got, just straightforward, more phyisical vibrating action going on.
 
 
Quote:

Lossy does effect dynamic range which could lead to the perception of more energy in playback.

 

 

[TMRaven]

Quote:

Could you link me to information on this? Every lossy track I've ripped has had almost identical ReplayGain volume levels to the original lossless, within 0.1dB.

 
That's because most cds don't really have a dynamic range over 60db or so, so the higher noisefloor of the lossy doesn't come into effect compared to the lossless.  However when dealing with something low like 128kbps vs something like 320kbps or lossless, I can tell a difference in the dynamics of percussion in the files.  The 320 or lossless file has louder peaks in it compared to the 128kbps version.  As regards to the whole visceral aspect though, I have a hard time believing.  

 

[Head Injury]

Quote:

 
That's because most cds don't really have a dynamic range over 60db or so, so the higher noisefloor of the lossy doesn't come into effect compared to the lossless.  However when dealing with something low like 128kbps vs something like 320kbps or lossless, I can tell a difference in the dynamics of percussion in the files.  The 320 or lossless file has louder peaks in it compared to the 128kbps version.  As regards to the whole visceral aspect though, I have a hard time believing.  

Noise floor has nothing to do with peaks. In order for the peaks to be quieter, the low bitrate file would have to be dynamically compressing them, such that quieter sounds are louder and louder sounds quieter. This would be apparent in ReplayGain's analysis. I've not heard of MP3 dynamically compressing anything, but I could be wrong.
 
I just encoded "Blowin' in the Wind" by Bob Dylan to V5 MP3 (~153kbps) and ReplayGained it. Both the FLAC and MP3 have -3.70dB gain, the MP3 has a peak of 1.217572 and the FLAC has a peak of 0.998932. That's weird, but it doesn't suggest any dynamic compression.
 
I could open them both up in Audacity if you want. I probably will anyway 

 
Yeah, I did anyway. Both have the same amount of dynamics, but the MP3 clips in several places. Probably because of the variable encoding, or compression artifacts.