View Full Version : MP3 quality poor due to low volume levels, notcompression.

Simon Turner
2004-01-04, 15:53
When I started this thread I did not really expect any answers, but have
returned from a few days away to find a wealth of information to trawl
through. Thank you to everyone who replied.
I don't really want to restart the thread but I have to clarify a couple of
things, add a couple of things I omitted and post the results of some (very
unscientific) testing I did yesterday.

First off. I originally started using MP3Gain because many of my MP3's
sounded quite distorted when played via headphones on my iPod. I found that
the distorted files always registered as clipped by MP3Gain (incidentally
the distortion was most noticeable on modern highly compressed and loud
music such as psytrance). When I used MP3Gain to reduce the db level to that
below clipping the distortion disappeared.

Secondly.. when I said ripping, I usually meant encoding. Sorry. It is the
encoding that appears to introduce the clipping, not the ripping.

I can't hear the distortion via the line-out of any other devices (e.g cheap
CD based MP3 player, iPod, Squeezebox, PC sound card) unless I increase the
gain considerably above clipping level (about 6db or 8db or more perhaps).

When I say the sound is thin and weedy (I think Jason asked) I meant the
sound is vapid, lacking oomph (or boll**** as we might say in the UK), a bit
powerless, not very dynamic. The amplifer I am feeding the analogue-out of
the Squeezebox into is an integrated amplifier (i.e. with both pre-amp and
power-amp) and i assume that the input (pre-amp?) is not being run
correctly. As an indication.. normally i might have my amp power know at
about thirteen minutes to the hour, with the Squeezebox i need it just past
the hour. If I used the CD at just past the hour setting I wouldn't be able
to hear the telephone and after a while the neighbours might complain.
I am not using a receiver, as someone assumed. For me a reciever is a radio.
Due to Jason's post I am now wondering if what I mean by weedy is due a
reduction in dynamic range caused by reducing the volume of the file to
below clipping level (as indicated by MP3Gain, not necessarily 89db).

I tried three different integrated amplifiers, a Nad, a very old Trio and a
Harman Kardon. The results were the same on all so I've not differentiated
below. I used a Harman Karden CD Player as reference.

I used a CD of a couple of particularly kick drum driven techno songs.
I ripped these songs to MP3 and made copies of them, adjusting their volume
levels in MP3Gain (in steps of 1.5db). I then played these back via the
Squeezebox and the iPod to see how the volume levels compared with the
original CD.

The original file did not register as clipping by MP3Gain. An adjustment of
1.5db introduced clipping
To achieve the same volume level as that produced by the CD player I had to
adjust the volume level of the files:
iPod: +4.5db
Squeezebox: +6db.

The iPod does not have any volume (or eq) adjustment for it's line-out. The
Squeezebox was set to max by the remote.
The file with the 6db above clipping volume sounded horrible in my iPod
headphones and also a bit yucky in the Squeezebox headphones. However, no
distortion was noticeable via the line-outs.

For me format is not an issue. MP3 is the one I am stuck with.
Using the Digital-out is not possible as I have decided that I will buy a
good quality hifi system (not an AV system, it'd be stupid as I don't really
watch any TV, videos or DVDs!) and that the price of DAC's is too high to
consider buying one.
I an still not really any nearer resolving my problem than I was at the
start when I originally made my post despite all the useful and illuminating
I am beginning think I may increase the volume of all my MP3s and sell the
iPod (although it looks nice, when you have most of your MP3 collection
ripped using 320 CBR or --alt-extreme the battery life is laughable, only 4
cds sometimes!). Hopefully I'll find another personal MP3 player with a
similar audio quality that will not distort (anyone tried the Dell one?).

Anyway. Thanks again for all the replies.

BTW. I've never considered myself an audiophile before. I just like loud
banging music!

Brighton UK

-----Original Message-----
From: discuss-bounces (AT) lists (DOT) slimdevices.com
[mailto:discuss-bounces (AT) lists (DOT) slimdevices.com]On Behalf Of Rob Studdert
Sent: 02 January 2004 15:08
To: Slim Devices Discussion
Subject: [slim] MP3 quality poor due to low volume levels,

On 2 Jan 2004 at 8:17, T wrote:

> Well, as I previously stated, there is NO clipping introduced (verified by
> various measuring equipment), despite the fact that mp3gain claims there
is. If
> it can't be measured, it's not there. And, if the equipment is designed
> correctly (and I've been doing it for 5+ years), a 0dBFS (or any other
> signal) will exit the decoder at exactly that level with no clipping.

Of course nothing exits the digital I/O at greater than 0dBr, it's the
calculated dynamic overshoot that's clipped.

> Actually, the ARE designed to accomodate the 'technicians top forty', as
> that is what they are primarily used for in the professional world (i.e.
> program distribution from studio to transmitting station).

I've only been in the game for 20 years and freq runs/test tones are oft
referred to as "the techs top forty", not real world dynamic audio.