Latest Thread Images

If I mostly listen to mp3's, does it make sense to spend more than $100 on headphones?
Apr 11, 2011 at 10:30 PM Post #121 of 151
Explosions are always awesome, regardless of reason.  x2 on the blowing the thread up.
 
And, besides, if you're going to upgrade your headphones, that seems like a good reason to upgrade your music too.  That way you won't have excuses NOT to upgrade gear in the future, thereby sentencing your wallet to eternal damnation.
 
Apr 11, 2011 at 11:04 PM Post #122 of 151
alright, files are uploaded now, heres what is included in the zip, a 30 second song full of all your transients, dynamics and complex harmonic detail you requested has been used, i have included a file containing just the information removed by the encoder, and, part of the song where the random parts of it are compressed and others are lossless. i want to know if you can tell me which parts are compressed, and which arent. all you need to do is tell me from this time to this time (ex. 0:15-0:20) its also must be noted the information removed from the file was way below the volume level required to even hear it against the rest of the music, in fact, its about 70dBFS down. interestingly enough to my surprise, there is a 10 second increment where nothing is removed at all from the song (which you will hear the silence in the file containing the removed information.) please use your absolute BEST headphones 
smile.gif

 
http://www.mediafire.com/?ojo5tvkavookoo7
 
Apr 11, 2011 at 11:30 PM Post #123 of 151


Quote:
yepimonfire said:
alright, files are uploaded now, heres what is included in the zip, a 30 second song full of all your transients, dynamics and complex harmonic detail you requested has been used, i have included a file containing just the information removed by the encoder, and, part of the song where the random parts of it are compressed and others are lossless. i want to know if you can tell me which parts are compressed, and which arent. all you need to do is tell me from this time to this time (ex. 0:15-0:20) its also must be noted the information removed from the file was way below the volume level required to even hear it against the rest of the music, in fact, its about 70dBFS down. interestingly enough to my surprise, there is a 10 second increment where nothing is removed at all from the song (which you will hear the silence in the file containing the removed information.) please use your absolute BEST headphones 
smile.gif

 
http://www.mediafire.com/?ojo5tvkavookoo7
Click to expand...


huh? i hear one long file of noise. that's the stuff that got removed?
and then one short clip of "music"?
what are we supposed to do with them?
 
 
Apr 11, 2011 at 11:33 PM Post #124 of 151
the noise is the information removed by the encoder, the other file is music where there are random points in the song that are compressed and lossy, i wanna know where compression starts and stops throughout the song. if you cant tell me, you cant hear it :p
 
Apr 11, 2011 at 11:37 PM Post #125 of 151


Quote:
yepimonfire said:
the noise is the information removed by the encoder, the other file is music where there are random points in the song that are compressed and lossy, i wanna know where compression starts and stops throughout the song. if you cant tell me, you cant hear it :p
Click to expand...



well, the problem for me is i don't even hear a song!  :)
 
but seriously, yeah, i can't tell where the compression starts and stops.
 
Apr 12, 2011 at 12:09 AM Post #126 of 151
so then this should thoroughly answer your question.
 
Apr 12, 2011 at 12:16 AM Post #127 of 151


Quote:
yepimonfire said:
so then this should thoroughly answer your question.
Click to expand...



oh, it was answered a few pages back. i was under the mistaken impression that mp3s didn't warrant high quality headphones. now i now better. and i will be buying some more expensive cans soon, to hear for myself.  :)
 
Apr 12, 2011 at 2:25 AM Post #128 of 151
@ the guy/gal who could really use a bucket of water
 
Clearly, there is a huge loss of dynamics, resolution and whatchamacallits(italian mobster voice) at roughly 5-10 and 25-30.
 
The fact that we could actually hear anything from what the 320kbs music test, which was all the stuff the encoding removed, meant that there were clearly audible sounds that were cut off. So if we could hear the noise that was removed (which could make the difference audible), would it be too unreasonable to say that little Timmy-boy out there might actually be able to hear the difference between lossless and mp3?
 
I think not, and that's precisely what some of the members here are trying to say ("You probably won't be able hear a difference"). I don't think those same members are trying to catagorise people into camps like 'Good listeners' and 'lol, u bad', they are simply accounting for the very real(albeit small) possibility that you could in fact hear a difference, because there is a difference. And this has been proven, because a previous member mentioned something about Timmy posting a positive ABX result from a lossless v mp3 test. For the masses out there like myself, it's not 'wrong' to 'fall' for the trickery of mp3, because that's the whole point of mp3!
 
So it's a little bit too much of a stretch to say things like ~"You WOULD NOT hear a difference between high bitrate mp3 and lossless, because the noise removed is inaudible"(but it's clearly not inaudible though, as you have demonstrated for us).
 
I absolutely share your perspective though, mp3 is really flippin' awesome, and I myself could not hear the difference, even down to something like ~120kbs on some songs. And I like to think I'm using decent headphones, so I'm not missing too much of the action out there. I also think some of the claims out there are a little....bonkers(to put it nicely), but you can't slam them down for accomplishing something that should be possible. It's just their opinions and experience, which are as worthless as yours and mine.
 
But your experience does not mean that there is no audible difference, because unfortunately, little Timmy would like to have a word.
 
Apr 12, 2011 at 10:12 AM Post #129 of 151
it is inaudible because the noise removed is about 70 dBFS down, that means if you are listening to a song at 80 dB (a common loudish volume.) the removed data is down at around 10dB, the brain will and does completely throw it away.
 
Apr 12, 2011 at 10:20 AM Post #130 of 151
also keep in mind this was at 320kbps, at something like 160 or 128, the data removed is audible and i will certainly not argue that.
 
Apr 12, 2011 at 10:55 AM Post #131 of 151
peskypesky: hey man, FWIW I think there is a point getting higher end 'phones. The bitrates you described are great, and nobody (not even the debating yepimonfire and ScareCrow) will deny that. I think there are reasons for getting better headphones, and they're independent of whether you can audibly ID an mp3 from lossless.
 
Let's just take "better" to mean more accurate and less boosting of frequencies. In that case, your SR80 might be bested by, say, a HD600, because those Grados have an emphasis on high frequencies, while the Sennheisers don't (just for e.g.; I'm not too familiar with either model). The 600 also has a rep for throwing a fairly wide, cohesive soundstage--in this case, that might also be cause to upgrade, since the SR80 is upfront in its presentation of music. 
 
Build quality could also be a good reason to get pricier headphones. It's not always true, but perhaps there are headphones with a similar sound to what you have, but less of the design "flaws". When I owned the SRH 440, one thing that bugged me about them was the little wire on the outside of the headband, connecting the band to the cup. I always had the fear of those wires getting snagged; there are similarly-styled monitor headphones without this weakness. The 440s are also kind of heavy; in contrast, Sennheiser's HD 25-1 is lighter and certainly no sonic inferior. On that note, the HD 25-1 isolates better than the 440--isolation could also be a reason to upgrade without suffering loss in sound quality.
 
Dang. I sound like some kind of Sennheiser fanboy.
 
Apr 12, 2011 at 3:43 PM Post #132 of 151
Apr 12, 2011 at 4:31 PM Post #133 of 151


Quote:
eric_c said:
peskypesky: hey man, FWIW I think there is a point getting higher end 'phones. The bitrates you described are great, and nobody (not even the debating yepimonfire and ScareCrow) will deny that. I think there are reasons for getting better headphones, and they're independent of whether you can audibly ID an mp3 from lossless.
 
Build quality could also be a good reason to get pricier headphones. It's not always true, but perhaps there are headphones with a similar sound to what you have, but less of the design "flaws". When I owned the SRH 440, one thing that bugged me about them was the little wire on the outside of the headband, connecting the band to the cup. I always had the fear of those wires getting snagged; there are similarly-styled monitor headphones without this weakness. The 440s are also kind of heavy; in contrast, Sennheiser's HD 25-1 is lighter and certainly no sonic inferior. On that note, the HD 25-1 isolates better than the 440--isolation could also be a reason to upgrade without suffering loss in sound quality.
 
Click to expand...


I hear you, no pun intended. The SRH440s are a bit heavy. And I know others are bugged by that exposed wire, although I'm not. In any case, I'm seriously considering some more expensive headphones and also a DAC/amp. Want to see what all the hype is about. Let's just say I'm skeptical. But open-minded. If something sounds better, I'm down with it.
 
Apr 12, 2011 at 6:15 PM Post #134 of 151
Quote:
skarecrow77 said:
 
Last I checked, there are still killer samples that even the newest revisions of LAME can't encode transparently no matter what the bitrate, the psychoacoustical model is not perfect and still screws some things up. Granted, we're at the point where we're talking about the 1-out-of-a-thousand special case here, but it's still there. There are still tracks that you can ABX even 320 cbr or -v 0 tracks from lossless on even modest equipment. The TRUE shining benefit of today's lossy compression is that the other 999 songs out of that thousand are transparent. If that was your point, you need to stop talking in absolutes. Also, the psychoacoustical model is just that, a model. It is not a perfect mapping of the aural capacity of even a single human being, let alone the entire human race. There is always room for a given person's ears to be "outliers" from the statistical bell curve. That sort of thing happens FAR less frequently than the users in this forum tend to claim, but it's still there, and will always be there as long as there is lossy compression. Humans are not a single model with concretely defined capabilities. My wife can usually ABX music easier than I can, and in turn am usually able to pick out video artifacts in stronger levels of h.264 compression before she can.
 
Oh, also, your earlier statement that WMA and AAC at 128kbps are the equivilent of MP3 at 256kbps is WAY off. newer codecs are inherently better designed than MP3 is, but the counter to that is that MP3 (LAME specifically) has received far more tuning and attention than the other codecs (yes, even Apple's baby AAC) due to the market penetration. That is, in simple terms, more people care to make mp3 better because it is supported in damn near everything made in the last 5-10 years. Today's LAME is almost equal to AAC and WMA at equivilent bitrates, It really comes down to the specific encoding which is better. Also, you'd do well to check out Ogg Vorbis if you haven't already, its performance at medium/low bitrates (~96kbps to ~160kbps) is probably the best of the bunch, especially the aoTuV b5.7 version of Vorbis. I am unable to reliably ABX -q 6 (~118kbps) from lossless on either my HD-650s or my DT880s (let alone my MS1s) on the vast majority of my music.
 
By the way, just a general FYI people, generally 320cbr or even -v 0 are a waste. If you can't tell the difference between 320 and lossless, you probably can't tell the difference between -v 2 and lossless either. -v 2 has recieved the strongest tuning and tweaking and is the sweet spot of the codec. That said, LAME is so strong nowadays that on a portable setup in an environment with some noise, etc, even the much-maligned 128kbps is a very strong contender.
Click to expand...

At one point I was wondering about -v2 quality for that reason.  Then I took a couple lossless files I had of actual music (relatively complex so more challenging for the encoder, but not "killer" samples) and encoded them to -v2 and -v0 with LAME 3.98.  I was able to ABX -v2 from the original but not -v0 from the original.  It wasn't just loss of detail--the -v2 had some funky high-pitched sizzle artifact that I could usually pick out.  Though -v2 should be good enough on most music, I found it wasn't quite good enough.
 
Since I don't feel like doing ABX tests all the time and testing every single CD I get, personally I encode to Ogg Vorbis aoTuV 6.02 (newer and better than 5.7) -q8 these days, which is about 256 kbps average and thus at least smaller than just running 320 kbps mp3 all the time.  There might be a better codec for that kind of bitrate though, I dunno.  I've had more success with aoTuV 5.7b compared to LAME at similar bitrates in the past, so I figure this new version isn't going to be worse.  That pretty much should be transparent no matter what, unless you're using a killer sample.
 
I'm not familiar with all of the technical details of the encoders, but I looked at the difference signal between original and mp3 vs. original and vorbis for one track, at similar bitrates.  The mp3 difference signal actually had less average power over the track than the vorbis difference signal, but the FR of the two were quite different.  Vorbis produced pretty much white noise, while the mp3 had more power in some frequencies than others.  Particularly, it had more power at high frequencies, where humans hear better.  I guess this is why the vorbis design is superior.
 
Apr 12, 2011 at 7:54 PM Post #135 of 151
i just did an ABX test with a piece of choral music. A was ALAC (Apple lossless codec). B was 320kb mp3.
 
the first time. my accuracy score was 40%.
the second time, 20%.
i was just guessing really.
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top