128kb/s mp3s - good enough for portable
Sep 6, 2006 at 7:24 AM Post #46 of 75
@jwolf

did you read any of my opening post?

the whole point of this thread is to determine whether 128 is adequate for on the run.

from an abx it is extremely difficult to tell a difference between 128 and 192 with the newest lame encoder.

this is because even at 128 no artifacts are audible.
older codecs used to have artifacts.

assumptions like if 320 mp3 is bad are just not credible.

it is virtually impossible to abx a 192 from cd quality over the long run,
for some music styles it might be possible every now and again, i.e. metal music.

but with an excellent encoder you have got no chance.

8k mono is unacceptable, it is worse than shouting into a tin.

then you proclaim 192 has artifacts

try abxing and it will give you a more objective view on these things
 
Sep 6, 2006 at 9:22 AM Post #48 of 75
Quote:

Originally Posted by cantsleep
i think i need something a little bit better than 192.. something like 256..


For what? If you can tell any sonic difference between high-quality 192's and 256's while being "on the run", I personally will give you seven billion dollars. Really. I will.
 
Sep 6, 2006 at 9:23 AM Post #49 of 75
Quote:

Originally Posted by cantsleep
i think i need something a little bit better than 192.. something like 256..


So does this mean you can ABX 192, with the latest encoder, or are you too just pissing into the wind?

h
 
Sep 6, 2006 at 9:59 AM Post #50 of 75
Quote:

Originally Posted by russdog
For what? If you can tell any sonic difference between high-quality 192's and 256's while being "on the run", I personally will give you seven billion dollars. Really. I will.



Quote:

Originally Posted by mnhnhyouh
So does this mean you can ABX 192, with the latest encoder, or are you too just pissing into the wind?



do i really have to be 'on the run' with my portable setup? can i just enjoy them while im working or something?

seven billion dollars.. is 7 outta 10 good enough?
if so, you owe me 7
evil_smiley.gif
 
Sep 6, 2006 at 10:50 AM Post #51 of 75
Quote:

Originally Posted by Asr
The sound quality difference between 128 and 160 should be noticeable to untrained ears. I used to compose with MIDI (MIDI!) and when I saved it off to a WAV for conversion to MP3, there were stark differences between a 128 and 160 file. Let me say again - it was MIDI. If I could tell a difference using the low-fi samples of a MIDI wavetable, then 128 is too low. If you're listening to ripped music at 128 that's just an insult to the music, regardless of source or headphones.


So what do you say to someone like me who has decent (not audiophile) equipment and good hearing but simply can not reliably identify the difference between 128kbps VBR AAC and 192kbps VBR AAC? You feel that I "should" be able to hear a difference yet it just isn't there. Does that mean your idea is wrong or does it mean I have defective ears?

And I'm not just talking about being unable to do ABXY, I don't even hear what I think is a difference when I know which track is which. I'm listening with a Senn HD595/Portaphile V2 combination (with several hundred hours use on them) and my source is either a iPod 5G or a E-MU 0404 using ASIO drivers from Foobar. If I can't hear it on that rig, then for most intents and purposes it doesn't exist. Wouldn't you agree?
 
Sep 6, 2006 at 11:04 AM Post #52 of 75
@ brent - did you abx?

if yes then you reinforce my results,

the new encoders seem to be really good,

although i disagree with buying lossy music and having no lossless backup,

but if apple is using the newest encoder for the songs on itunes then people are actually getting decent quality files.
 
Sep 6, 2006 at 12:52 PM Post #53 of 75
astranovus, you did the right thing by making a listening test for yourself. I have noticed that the "transparent" bitrate at which to encode mp3's seems to vary a lot depending on the equipment used to listen (headphones, amp), the listener' s training, the type of music and the encoder.

I did some listening tests on my hifi a few years back (not real ABX, just switching) between mp3 CBR128 (old encoder), lame preset extreme (it was new at the time, ~230kbps) and the CD. The results was that I found the 128 cbr a bit flat (not much sounstage, less dynamic) compared to lame extreme and the CD. I couldn't really notice a diference between alt preset extreme and the CD, so at the time I encoded everything in lame alt preset extreme (or the nero equivalent for aac) because I wanted to use my ipod with the docking and my hifi (it was a 3rd gen ipod and dockings were "free" at the time) and quality headphones, but for use with the stock ipod's earbuds 128 cbr would have been ok.

Now I recently came across files my sister encoded to 128 CBR aac in itunes and they sound much better than the old 128 cbr mp3's that I had tested at the time, definitely good enough for on the go listening. With modern encoders (lame 3.97b and the latest nero) I encode in preset standard (lame) or q5 (~190kbps in both case) intead of extreme because I can't honestly say I hear a diference (again, not ABX, just listening on my hifi and with good headphones).

On the other hand, I know someone who's a former professional musician, use very revealing headphones and finds even "modern" 128 cbr mp3's unlistenable and 256 cbr "not perfect but OK" compared to CD, so apparently training and setup can have a huge impact.

To each his own bitrate....
 
Sep 6, 2006 at 1:58 PM Post #54 of 75
Quote:

Originally Posted by astranovus
@ brent - did you abx?


Well, I tried. Gave up pretty quickly since I was getting chance (or worse) results. So I backed off and realized that in most cases I didn't hear a difference in A and B to start with...makes it kind of hard to get X right, you know?

I've downloaded a few things from iTunes including a couple of my favorite tracks where I paid the 99 cents for an iTunes track that I already had in a WavPack file. On the iTunes stuff I think I can hear a bit more clarity in the WavPack but I still can't pick X correctly in a blind test.

But a Nero-encoded AAC that averages out to similar bitrates (VBR with a -q .3 switch, IIRC) it sounds just exactly like the original CD on my equipment. I'm not really suprised since a lot of the music I listen ends up with certain tracks only using a very small bitrate in VBR. Heck, the WavPack compression can get some of this stuff down to 600-700kbps so it's obviously not very complex audio.

I've pretty much started ordering iTunes music whenever the mood strikes (i.e. when it's cheaper than Amazon and my local record store doesn't have it in stock). Most of my listening is at work with an unamped PX200 headphone or in the car with a cassette adapter. If I can get it locally I don't mind paying $12 or $16 and then I have a permanent copy of the music in Redbook format. But if it's a special order I figure what the heck, iTunes is $10 and the quality is good enough for my normal usage.

So far, my collection is about 100 albums for which I have the CD and another eight or nine that I only have iTunes downloads. That's probably the ratio I'll continue with over time. I also have a couple dozen individual tracks from iTunes which is a cheap way to acquire those classic pop music songs whenever you take a notion.
 
Sep 6, 2006 at 10:51 PM Post #55 of 75
Quote:

Originally Posted by mnhnhyouh
How old was the mp3 encoder?

h



I don't see how that's relevant but at the time I was using the version of LAME that was included in CDex 1.40.

Brent, I'd suspect you can't easily differentiate probably because of the difference in encoding formats and possibly the music as well. When I was referring to 128 and 160, I meant CBR MP3. VBR AAC will give better results for sound quality.
 
Sep 7, 2006 at 6:38 AM Post #58 of 75
I can definitely notice the difference between 128k and 192k+. Its mostly just slightly better treble detail but my enjoyment of music remains the same. A bad song is still a bad song if you know what I mean
etysmile.gif
 
Sep 7, 2006 at 6:43 AM Post #59 of 75
Quote:

Originally Posted by hardstyler
I can definitely notice the difference between 128k and 192k+. Its mostly just slightly better treble detail but my enjoyment of music remains the same. A bad song is still a bad song if you know what I mean
etysmile.gif



Wize words indeed.
 
Sep 7, 2006 at 7:24 AM Post #60 of 75
I've done some blind testing with the Lame encoder, this is what I found

128kbps
- sounds much better using joint-stereo than regular stereo. Test it out yourself. This was with the latest stable version of Lame
- good for the "on the run" type sound someone was talking about earlier. I use this encoding bitrate for songs I dont really care for sonically, but would like to fit on my mp3 player, in other words I'll trade the best sound quality for decent quality to save space for the songs I dont care for AS MUCH as other songs. I wouldnt use anything lower than this, and I would use joint-stereo. For fine listening I can tell a difference, but it doesnt ruin it for me (luckily)

192kbps
- pretty good, most of the time I can't tell the difference, and if I did it wouldnt matter to me cause it still sounds pretty good. I used to use this for songs I liked and wanted great quality/size ratio.

VBR Joint Stereo presets
- love them, dont use anything else now. If I used a v-5 preset I think it was, the average bitrate is supposed to be 132kbps (between 110kbps to 150kbps), the file ends up the same size as a 128kbps CBR file, yet it sounds better because it has a higher bitrate up to 150kbps I think for the parts of the song that need it, and lower than 128kbps for the silent parts that dont need a high bitrate. This is what VBR does. I personally use Lame preset standard V-2
- joint-stereo sounds better than regular stereo, but dont be fooled its not combining two channels into one (a common misconception)
 

Users who are viewing this thread

Back
Top