320 kbps MP3 vs. normal audio CD listening Sound quality - Page 4

 

It's very interesting. What are the perceptual differences between a rubidium clock and a cheaper one? How did all those audiophiles describe the differences? 

 

The peer research done to date indicates that jitter is normally just not a problem and no amount of hand-waving , ultra-high-end clocks or sighted anecdotes will really alter that. If you have access to better empirical data that suggests otherwise we would all be (certainly I would be) interested to read it. To date jitter below absurd levels is just not a problem. You should look up Benjamin and Gannon, 1998 who are Dolby labs researchers and Ashihara et al 2005 (NHK researchers) for the best empirical data to date. One lot did it with sinusoidal jitter the other with random jitter, in no cases did peak-to-peak jitter below 20ns cause any issues in musical listening. 10ns was just audible in single frequency 11Khz test tones.

 

For theoretical information chaps like Julian Dunn and Malcolm Hawksford are better sources (no axe to grind) but they do not do empirical listening tests only provide mathematical models and often these models are based on very pessimistic assumptions such as listening to music 120db above the noise threshold. For all intents and purposes any $200 DAC competently engineered will give you sufficiently low jittter so as not to lose any sleep. For pathological cases the McIntosh music server is the poster child but that is a real outlier. 

 

If you want to worry about jitter be my guest but it is a largely wasted effort and no number of sighted anecdotes, obsessional Stereophile reviewers ( the sort who who will allow 1% speed variation in turntables through without batting an eyelid and tut-tut about 500ps jitter)  and industry dog and pony shows will alter that...Now if dCS had some level matched blind tests properly proctored that contradicted what we know so far that would be very interesting but another manufacturer white paper with no empirical backup is insufficient evidence to be taken seriously round these parts.

WindowsX, have you tried the

 

 

It also works with CDs and improves the sound to beyond what you can imagine. 20 audiophiles approve this message.

 

I'll chime in.

 

I can tell the difference between Mp3 and lossless / CD Source when flipping back and forth, but in only ONE way. It sounds tarded' but CDs are definitely "louder". As in, you can detect the improvement in dynamic range. Other than that, the recording doesn't matter, everything else sounds the same to me. I would probably flunk a blind test on a CD I wasn't familiar with.

 

Its a little easier to detect low bit-rate Mp3s (like 128) to a 320 / lossless. I did it a couple times today for fun. But it wasn't anything too distinct. There was some echo in the recording which was rendered more smoothly in the higher bit-rate recording. All around, high bit-rate sounded a little smoother, but it wasn't easy given the few instruments in the recording. The complexity of the recording doesn't seem to matter either, since I think a bad speaker will tend to render complex music badly no matter the bit-rate.

 

The only way to really start screwing things up is with sampling rate. The higher it goes, the better I think things sound, the lower, the worse.

I'm not sure, but I think encoders knock the volume level down to prevent clipping with hot mastered tracks. My AAC files are always a hair quieter, but if I boost the volume, they sound the same.

 

Do the test on any CD you are familiar with using Foobar 2000's abx comparator (ie completely blind) - make sure you use the volume matching setting.  Use a lossy encode of either 256aac or 320mp3, and compare it to the original lossless rip from the CD.  Use a CD you are completely familiar with.  Perform 15-20 abx iterations.  Post the log.

 

I'd be 99.99% sure you won't be able to tell the difference if you set up correctly.

 

In what you said above - you described the difference to be 'loudness'.  So you weren't using same source?  And it definitely wasn't volume matched.  This has nothing to do with dynamic range - and most humans will immediately spot differences in volume.  That's why for a valid comparison, the two comparative tracks have to be volume matched.

 

Try it under correct conditions - if you're game  - it really is enlightening.

 

Lol. Under those kinds of controlled conditions, I am sure I would DEFINATELY flunk.

 

But this sort of goes to show that, under my "real world' flip back from CD to Mp3 and so on, there is some influence over the sound coming from somewhere. Is that the "compression" effect generating the slight volume difference?

 

I just tried the test 15 times. I scored 8/15. How do I find the log?
 

I can. I rip my CD's to Wav and much prefer the sound, for my ears it is more organic and able to express the emotions of the artist. In comparison 320k or less feels flat and lifeless to my ears. More importantly It does not engage my interest. I get fed up quickly with compressed music, it seems to irritate me where as uncompressed leaves me feeling more relaxed.

I don't really care that the flat earthers will bang on that there is no difference, I simply don't think they know how to listen to music!

 

It won't have anything to do with compression - its more likely to be the volume set via the transcoding tool.  It'll be volume that's influencing you - nothing else.

 

 

 

8/15 = purely guessing (ie half the time you got it right).  When you hit the exit button to finish the test, you'll get the option of saving a text file.  That has the log.

 

It's actually quite liberating to really find out your limitations isn't it?  Will also be amusing for you when the next "can you tell the difference" thread appears - to see all the people who've never conducted a proper test talking about night and day differences, or perceived flaws 

 

Or perhaps we're just open enough to actually run a blind test (which is pretty easy to set up) and truly find out for ourselves .  Placebo can be a hugely powerful influence.  A blind test mitigates that.  Then all that is left is the truth.  Whether you want to find out that truth or not is up to the individual.

The difference between compressed and uncompressed isn't an overall thing like "flat and lifeless". It's artifacting at specific points in the file. If you are detecting a "veil" over the sound or some sort of overall lack of energy, you can pretty much bet it's all in your head and would instantly dissappear with a fair test.

See, this is what i mean, pointless even going on as you guys just refuse to go beyond what you perceive, I think this is where I exit and hope you enjoy the music in the way you do, I will continue to enjoy it the way I do, after all, music is an art, not a science

Offer a clue, then comes the grand exit.

imo the difference between 320 and uncompressed/lossless is that the flac one sounds a little smoother and full bodied, while the 320 sounds little edgy
but then again these differences are so extremely small i confuse some tracks, i cannot hear the differences, i only can feel them

Smoothness and full bodiedness are not what the difference between compressed and uncompressed sounds like. Listen to extremely low bitrate compressed and then listen to a little higher and a little higher. The artifacting at the point where it crosses over into transparency is smaller and more infrequent, but it's exactly like low bitrate artifacting.

At some point the artifacting becomes so it's only in isolated spots. When there is suficient bitrate to render the sound, it's transparent. You only get artifacts in the complex areas that are harder to render. It isn't an overall thing.

Your description of only being able to "feel" the difference and not hear it is a classic description of sighted expectation bias.
Edited by bigshot - 10/24/12 at 10:25pm