[taymat]

I was wondering what the real differences are between mp3 bit rates and the answer is coming to me in blind tests:
I identified the better quality mp3 4 out of 5 times between 128 and 320, 4 out of 5 times between 198 (alt preset standard) and 320, and I'm listening to 233 (alt preset xtreme) and 320 right now, to get more used to them before I do a blind test. Why? Because it's getting hard, really hard. Maybe 256 cbr/alt preset xtreme is the limit for most listeners, what are your thoughts?
Edit: No I really can't hear a difference. Out of about 15 times I only got the right answer about 7 times! I'll be downloading at alt preset xtreme in future not paranoid 320 cbr.

 

[breez]

I'd suggest using more trials at a time to reduce the error margin. There is still a big probability of getting 4/5 right by guessing.

Do you use some kind of plugin for Winamp to do the testing? Foobar comes with an ABX comparator which makes testing easy and straightforward.

 

[Shosta]

When i decided to buy a mp3 player i made comparisons between a lot of compressions levels with my laptop and my Sennheisers HD600. I couldn't tell diferences upper CBR 224 kbps.
When I adquired the player I repeated this tests. The same track at different compressions (and MP3 versus WMA too). Philips HDD120 >> Cmoy and Senn-HD600 and I could tell subtle differences not important upper 256 kpbs.

In my tests (my ears and configuration):
- CBR at 256kbps MP3 and upper, diferences were difficult spot.
- WMA had a different quality (my player? the decoder? the wma encoder?) a slight boost in lower frequencies????.
- I hear VBR (dbpoweramp-LAME:224 kbps to 320kbps) equal than CBR 320Kbps. At least with that hardware.
- My laptop have a bad output (hiss and detail lost).

 

[jpr703]

To me it's a tradeoff between file size and sound quality. How sensitive you are to that size depends on your MP3 player. I have a 40 gig iPod and would rather err on the side of quality since the file size differences aren't all that significant (for 40 gigs worth of storage) from 224 to 320. I draw the line at lossless since that's where the file size takes a big jump.

 

[morphie]

Don't know what gear youre blind testing with, but I can note a big difference between each gap (including 256--->320) and lets not even talk about lossless which i use 95% of the time now...

 

[Alu]

Your setup plays a major role in this, as you might already know. As I want "original" CDs as source anyways I'm using flac and ape, lossless, only through my asio transit, 16bit of course, all for the originalness.

Even if I would not be able to distinguish the lossless from a very HQ lossy, say mpc, I'd still feel very uncomfortable not having the most perfect possibility. (for cd compression at least) I never tried distinguishing hq lossy and lossless, and I have no intention of doing that either, because I don't want to downgrade. HDD space isn't an issue here anyways.

 

[taymat]

Quote:

Originally Posted by breez I'd suggest using more trials at a time to reduce the error margin. There is still a big probability of getting 4/5 right by guessing. Do you use some kind of plugin for Winamp to do the testing? Foobar comes with an ABX comparator which makes testing easy and straightforward.

Well maybe there's a significant probability of guessing the right one, but with 128/320 comparisons I heard slightly less presence/detail, especially when just listening to intros which are easier to remember. When the music was more complex I thought I heard thinner, less full bass and metallic-like harmonics to high frequencies.
Alt preset standard/320 comparisons were hard, and I'm not sure of the exact differences I just got a gut feeling that certain background parts of the song didn't sound right, they lost ambience and decay. To elaborate they sounded like they stopped a bit soon, started a bit abruptly, and were not as smooth IMO.
The alt preset xtreme/320 I couldn't hear any difference, and when I thought one existed it didn't. Also the results were completely random, it was 60/40, 40/60, all the time so I'm sure there's no difference.
I'll try foobar to use abx to try and make the blind testing more random and use more results, but I'll admit just playing the song normally I couldn't spot differences, only with 128/320 comparisons. If you use pause in winamp though and make a mental note and then do the other one and make a mental note of what to look for and then wait for that point in the song when you click randomly for 30secs or whatever to shuffle playback, it becomes easier.
Edit: Maybe the reason I can't spot differences between alt preset xtreme/320 cbr as others can is that I'm using akg k240s.

 

[Tim D]

Sounds about right...funny thing I was using AKGK240S as well and made similar comparisons. The test I took was pretty darn blind and set up by someone else though. It was a recording I've never heard spliced and different sections were encoded differently and you had to identify where these transition points were and to what!

For portable use lame aps is good enough, and for the average noisy air-fan computer setup as well. In fact CBR encodings really annoy me compared to VBR, but I guess HD space is cheap.