Recording WAV vs. MP3

It seems my Tascam dr5 recordings sound better when recorded as MP3 (???). Is it possible Tascam has settings that propel the sound quality of these proprietary MP3 recordings? They do inform that one can not sell these (MP3s)… without acquiring some kind of license or something… The WAVs (44 kHz/16 bit) by comparison sound rather flat (even muddy) and generally of lesser quality… even comparing to 128k MP3s (44 kHz/16 bit)… Is this possible? Or are my ears failing/playing tricks on me?

Is maybe WAV some sort of dud format to record with? Even being surpassed by MP3s?

16-bit 44100 Hz WAV format is a “higher quality” format than MP3 (regardless of the MP3 settings), in that it is a more accurate representation of the audio data. MP3 encoding always discards some of the data, which causes “compression artefacts”. For high bit-rate MP3s (around 190 kbps or higher) the loss of quality is negligible / barely audible. At bit-rates of 128 kbps or less (stereo music), most people can hear the difference between MP3 encoded audio and the original uncompressed data. Some people prefer the sound of lower quality MP3s, just as one may prefer junk food to good quality food made from fresh ingredients, or prefer instant coffee to “real” coffee.

If they have gone through the trouble of providing a proprietary mp3 encoder, it could very well be that it applies some tricks, like bass boost to make it sound “better”.

You see that with some audiophile players too.

Going on other Tascam equipment that I’ve used, I doubt it. However, if Black Dog Bluez posts a couple of short samples of the same (identical) music, one recorded as WAV and one as MP3, then we can easily see what differences there are.

Okay, will post samples ASAP. I’ll have to record these from recordings to control that everything is equal… I figure… If I perform live, of course each time would be different. And I figure best to use some already well-known established recording, copyright aside, as this will not be for profit, and I’ll only sample part of a song.

Though in continuation, what is it that happens to an MP3? Just compression? With the possibility of this MP3 process possibly a sort of edit designed to improve or rather maybe remove all the (only known to German scientists) things a song doesn’t so much need or may even sometimes be better without?

Also, considering, the MP3 process available free (LAME) may be a lesser type then available when one has paid for some device or program containing MP3 features/technology. Tascam (a German company) seems to be tight-lipped on all but the basic info of their device (tascam dr5) with not much support either, short of emailing I assume, which I have not pursued.

Tascam, or TEAC, is a Japanese company…

As for software, their main suppliers are in France or Switzerland.

what is it that happens to an MP3? Just compression?

It’s not “just compression.” The main problem with MP3 is the sound damage increases every time you edit the work. Ever had the experience of downloading music and making a dance or party mix, but the mix sounds a little more like a cellphone than you would like? That’s multiple-pass MP3 compression.

WAV is a perfect format and doesn’t have that problem. The audacity default sound format is 44100, 16-bit, Stereo which is Audio CD format.

You can make an MP3 (or other format) from a perfect WAV file, but if you have an MP3 master, the sound damage is sticky and rolls forward. MP3 was ever only intended as a space-saving format for your music (or video) player.


There have been many advances in psychoacoustic modelling (the main area where audio data compression improvements are possible) over the last decade or so. The commercial Fraunhofer encoder was effectively frozen in the mid '90’s. As far as I’m aware, LAME is the only MP3 encoding technology that is still being actively developed and is generally considered to be the best MP3 encoder available at medium to high bit-rates. In particular, VBR encoding in LAME is significantly better than the reference design.

For more information about MP3 and related technologies, the Hydrogenaudio forums have a lot of good information.

To avoid possible legal/copyright issues, have a look at “Creative Commons” licensed music:

Here are the (Tascam DR5 recorded) samples:

WAVE 16 bit 44 kHz mono (705 kbps):

MP3 ?? bit 44 kHz mono (128 kbps):

The WAVE was trimmed with Audacity (2.1.2) (set to 16 bit, not 32 bit float). MP3 trimmed with MP3DirectCut; no volumes or other changes were made. Recordings exact same unless any outside noises entered. possible outside noises currently are neighboring AC units intermittently powering on and off which do effect my recordings.

I’m not sure why I thought Tascam was German, maybe I’m getting this mixed up with the MP3 technology, which I think heralds from Germany.

Here is the Tascam MP3 disclosure:
, which I’ve yet to explore and fully understand, and assume if I release MP3’s from WAV with Audacity that this is not an issue — or is this an issue with all MP3’s?

BTW I’m not selling anything but just don’t like the idea that if I create with an MP3 it is not really mine! This, assuming there is no such issue with Microsoft’s WAVE or any other format.

Also, per MP3 processing/creation: What is this “multi-pass compression” they use? I assume some other kind of compression other than the usual in audio editing.

The difference I notice in MP3’s compared to WAVE … is maybe this compression (regardless of what kind it is). Obviously a compression working not like the usual… and a curiosity to me. Point being, the possibility of MP3 technology being utilized to improve the sound of all formats, as a possible effect… … thinking maybe not all the technology the MP3 utilizes is necessarily bad — if it’s cancelling bad sounds for good to any extent?

If I’m on to something: I don’t think there are any such tools. As I’ve tried with Audacity to “clear” my WAV recordings but it always seems to just throw the balance off, taking away “necessary” parts. I theorize MP3, in taking away, chooses wisely — to an extent anyway — and ‘takes away’ by such means that are not known and or available to me.

The two sound very similar to me. I can hear a tiny bit of “pre-echo” in the MP3 version, which is to be expected (example, listen carefully to the onset of “C” at 12.3s and you may hear a slight smearing of the high frequencies).

I’d also expect the MP3 version to have a gate-like effect on noise in extreme high frequencies, though I can’t hear that on my loudspeakers and I don’t have headphones available at the moment. This is clearly visible in the track spectrogram view.

If I want to gate high frequencies (or any other audible change), I’d rather do it intentionally with an effect rather than have it done automatically as a consequence of data compression. Ideally I would want the exported sound to be as similar as possible to my original, and WAV does that better (more accurately) than any lossy compressed formats.

All “lossy compression” formats (MP3, Ogg, MP2. AAC …) discard “unimportant” audio data so as to reduce file size. The lower the bit-rate (smaller files), the more is thrown away. High bit-rates discard less audio data. There are subtle differences between different formats as to how they decide which audio is “important” and which is not. At very low bit-rates, the losses are obvious and, for music, they all sound “bad”. At very low bit-rates, the differences between compression formats are clearly audible, for example SPEEX sounds much less bad at extremely low bit-rates than Ogg or MP3, especially for speech, because SPEEX is optimised for low bit-rate speech, whereas OGG and LAME MP3 are optimised for medium to high bit-rate music. (The Speex codec has been obsoleted by Opus)

MP3 licensing has been a nightmare in terms of licensing and IP rights for decades, and in my opinion, one of the best (worst) examples of how patent and intellectual property (IP) law can be abused by big business, hence my personal recommendation to avoid it.

Mp3 is too lossy and sounds worse than the wav. WAV is for recording and editing.