Re: [MLUG - DISCUSSION] "optical digital sound cards" OR "what's the best way to play MP3 files on stereo system?"

[Jack Smith <EMAIL:PROTECTED>]

On Sat, 2007-09-01 at 18:14 -0500, Mike Miller wrote:
> On Sat, 1 Sep 2007, Stephen Montgomery-Smith wrote:
> 
> > Some people claim that they can hear the difference between MP3 and WAV 
> > (i.e. CD) recordings, because with MP3 compression there has to be some 
> > kind of loss of data.  So maybe by using MP3 you have already lost any 
> > advantage that optical output might bring.
> >

Another option is to use some sort of USB speaker system. The signal
from the computer to the speakers is digital and the speakers have an
internal ADC that they use before they can output the signal. I wouldn't
think there is any quality difference between digital data sent over a
USB cable or an optical line as both are digital. 


> I can tell that you are new to this, but what you've said about lossy 
> compression is definitely true.  The thing is, MP3 is not just a single 
> way of encoding audio data, there are lots of choices.  One of the most 
> important choices is the bit rate.  If you have a high enough bit rate, no 
> one will ever be able to distinguish MP3 from audio CD.  With a low bit 
> rate, they are distinguishable, and if you go low enough with the MP3 bit 
> rate it is trivial to tell the difference.
> 
> A lot of everyday MP3 encoding is done at 128 kbps.  That will be 
> excellent for many audio tracks on most playback equipment, but sometimes 
> you'll hear a bit of phasing artifact (that's what I call it, not sure of 
> the technical term).  It's a distinctive sound and it can be annoying. 
> For me, I use a variable bit rate in the range from 160 kbps to 224 kbps 
> and that seems to work great -- great because I have never been able to 
> tell the difference between the copy and the original and because it is 
> still reasonably good compression.  I've never heard of a blind test where 
> anyone could tell original CD audio from MP3 encoded at 320 kbps, so 320 
> kbps seems like a good choice but the files are much bigger than I need. 
> Most people who can claim to tell differences might be imagining things -- 
> this is the sort of thing where strict blinding and randomization is 
> essential.

Most music encoded in 192 kbps MP3 and decent speakers sounds close
enough to the original CD that I can't tell the difference. 128 kbps is
all right, but it can sound a little "fuzzy" at times. Anything below
128 kbps is crap. I use Ogg Vorbis much more than MP3 and Ogg is a
little better, with a 160 kbps Ogg Vorbis file being pretty much
indistinguishable from thee original CD to me. 

> Another popular format is FLAC (free lossless audio codec) which, as you 
> can infer from the name, is lossless.  It's a good encoding scheme but it 
> takes up about four times the volume of my MP3s.  It is true that storage 
> costs continue to plummet, but there is more cost to the FLAC than that: 
> I find that if I play back a FLAC file while doing other things on my 
> computer, it will stick a little because the CPU can't keep up with all 
> the work, but with MP3 that never happens.  Again, I can't hear a 
> difference between my 192 kbps MP3 and the FLAC (except when the FLAC is 
> sticking and sounding bad), so I just use MP3.
> 
> Mike

I have most of my music ripped as FLAC. It's not that I can really hear
a difference between a 160 kbps+ Ogg Vorbis and FLAC but that I can
recreate the original WAV file from the FLAC files and burn a perfect
copy of the original CD if I lose, scratch, or otherwise maim it. Any
lossy codec like Ogg or MP3 will not give you back your original data,
so a CD burned from Ogg or MP3 files loses quality if it gets ripped
again, like doing serial copies of cassettes. 

FLAC files are large, but I don't see my computer lagging at all when
playing them. My old and now-dead 5-year-old laptop with a 2.2 GHz
Pentium 4-M didn't use more than about 10% of its CPU while playing FLAC
files and my desktop with a 2.2 GHz Athlon 64 X2 uses only a few percent
of one core. I think your problem might stem from the fact that FLAC is
a much higher bitrate and thus needs to read much more from the HDD than
MP3, so if you're fighting something that is HDD I/O heavy, you'll hear
that much more in FLAC than MP3 audio. My old laptop had a 4200 rpm HDD
in it and I certainly saw the same behavior, but a quick look at the
disk activity in gkrellm showed why. I'd give that a check if I were
you- FLAC shouldn't take too much CPU to play unless your CPU is very
old and slow or has about no FPU power. 

Now video is another story altogether- some of the more advanced video
codecs like H.264 or xvid4 do eat some significant CPU when they are
playing. My new 1.067 GHz Core 2 Duo laptop is about tapped out playing
720x540 video files encoded in H.264 but plays the same video encoded at
the same 1.00 bpp quality setting in xvid4 with about 40% CPU usage. I
tested the effect of the laptop's 945GM IGP versus discrete graphics by
playing the files on my desktop, which has an ATi R580 GPU in it. The
CPU usage on the 2.2 A64 X2 was about 45% for the H.264 and 15% for the
xvid4. 

--Jack


_______________________________________________
discussion mailing list
EMAIL:PROTECTED
http://mlug.missouri.edu/mailman/listinfo/discussion