, 
, 
or the Mac version of Microsoft Internet Explorer in order to view the rollovers correctly.Recently, after setting up Audiograbber1.82.02 to use Audioactive Production Studio 2.0.4j as an external mp3 encoder I realized that that program has serious flaws. Not with the encoding, but with mantaining settings and command line parameters. It simply refuses to obey certain commands and will constantly change its encoding options unless you manually enter everything.This makes it horrible if you want rip a cd and make mp3's with correct id3 tags as everything must be done by hand. So I set about my journey to find a newer compressor based upon the Fraunhofer codec
What I discovered was a bunch of out of date pseudo-scientific comparisons comparing discontinued programs with settings I don't use or informal listening tests which are not focused at determining what compression scheme is being used. I found the best collection of semi-intelligible information at www.ff123.net. But that mostly consisted of stuff like " As you can see here by this frequency analysis of test signal g24365s viewed at -40 decibals of half inverted cosine transform B generated using maker.dll 3.23.1343 not 3.23.02 there appears to be attenuation at 17546 Hz." This really didn't help as I wanted to know three things.
The comparison consisted of the following 6 programs based on the Fraunhofer encoding engine.
This is intended to be a simple anyone can do it and prove it for themselves exercise that consists of five easy steps.
The following test samples were used in the creation of this Guide:
, , or the Mac version of Microsoft Internet Explorer in order to view the rollovers correctly.
The Radium documentation states that these two programs use the exact same Fraunhofer encoder and that the radium codec (see below) is basically that encoder which was hacked out of these two programs. This test was just to verify that Audioactive and Producer Pro really are the same. They are :).
The above picture represents files encoded with audioactive at 128 Kbps and 256 Kbps respectively. Even at 128 Kbps it encodes frequencies above 16000Hz leaving not much of a difference between 128 and 192. This is why some people supposedly hear artifacts as it is encoding a wide range of frequencies at low bitrates.
The documentation accompanying this file is quite detailed and repeatedly states that it is the same encoder as Audioactive and Producer Pro here is a brief excerpt:
Q : How does the 'Radium codec' quality compare to other encoders?
A : Audioactive, Opticom Producer and the Radium codec all essentially
use the same engine - they will all create the same quality of output
if given the same input and options. These Fraunhofer codecs are
widely recognized as being the best mp3 codecs available.
Exactly the same (the hecklers in the crowd will say wait a minute I see a little something in the left channel), but remember this is an external acm codec that has been hacked out of an .exe file and then edited with assembly to improve the encoding speed. After encoding some more random files from my cd collection and comparing them I never had this happen again. So out of all the files to encode I picked the oddball. There has been some speculation that the radium codec boosts the volume or has a different cutoff frequency. This is not so. The above is as much of an inaudible difference you will ever see. Try it yourself and see.
Look at the spectral graph above, do you see the bar on the right side of the screen? Thats all the information that the radium codec deleted! After I pasted the radium encoded file on top of the one encoded with mp3 producer, the waves cancelled out except at the end because the radium encoded file was that much shorter. I definitely don't want any part of my music deleted no matter how small so this codec is out the door.
The encoder in this program is completely different than the Aaps/Mp3 Producer/Radium/Kristal ones above. It's also one of the newest. Here is a pic of the above encoders vs. the one included in the newest version of cool edit 2000, fht.flt (dated november 2000). Note: be very careful when setting your encoding options, depending on how you select your encoding rate, the maximum bandwidth and encoding quality will fluctuate for the same given bitrate. As stated above in the comparison section, I used a bitrate of 128 and high qualiy mode which corresponds to what cool edit 2000 and cool edit pro call (alternate codec high.)
Note 2:In Cool Edit 2000 the "alternate codec medium" section is based on the "alternate codec high" resulting in pretty much the same output (including the dropouts). The Fast codec is very different and does not share the alternate codecs characteristics.This is different in Cool Edit Pro so read that section as well
Completely different. As you can see this encoder clamps the bandwidth to ~16000 Hz at 128 Kbps and extends it from that line as you raise the bitrate. Not shown here, but true is that after 160 Kbps the Hq Mode gains the advantage over the fast mode and encodes more high frequency and at higher quality. Below that point the high quality mode encodes more of the higher frequencies, and generally does a better job. At 128 Kbps they are about the same.
See those glaring white spots right in the middle of the dark bars? That means there is absolutely no sound there, it also means that this is one of the easiest mp3 artifacts to here. Its glaring, it sounds like a cdplayer skipping; and its loud. Anybody can notice if it happens. This isn't something that recquires supersonic hearing or an army of audio engineers to test so I'm glad its fixed. Notice anything else different; like a drop in frequency? Keep reading.
The newsest versions of Fraunhofer found in these two programs are closely related. The one shipping with Cool edit pro dates April, 18 2003 which is almost three years newer than the one bundled with cool edit 2000. The main differences are:
In Cool Edit Pro the medium setting output more closely matches that of the fast option not the high quality option.
| Fast codec comparison | High Quality (alternate) codec comparison |
|---|---|
 The fast codec is quite similar in both of these revisions. A quick visual inspection will show that the output is nearly identical as far as frequency response goes. However if you were to invert and paste you would see that the output has changed significantly although not nearly as much as the high quality codec. |
 The High quality codec has definitly changed in this version. It also appears to cut off the end if the clip, like radium This clipping occurs often but not all the time which is strange. Ce2000 High quality version never clips. While the fast versions of both programs occasionaly clip. Changing the padding padding doesn't help. I have no idea why this happens. If you mouse over the above image, you can see that the ce2000 mantains a straight line at 16000 Hz while Cool edit pro reaches to 16000 but likes it around 14500 Hz. |
| 256 Kbps Fast | 256 Kbps High |
 As the bitarte rises the fast codecs continue to look alike, the newest version encoding a bit more of the high frequencies. |
 As stated above, the high quality codec regulates the sound to 16000 Hz while the prevous version encodes all the way up to 21000 Hz. This cutoff seems very drastic especially at this high bitrate. |
To be honest I can't tell the difference between any of these encoders. I tried having someone play random files to me and then try to guess which ones where what, but I failed. No, I'm not deaf, tone deaf or hearing impaired but theirs nothing really audibly wrong with them either.Objectively I would say the following.
I haven't tested Music Match, but I've heard that it only utilizes the fast codec and I'm not sure if it uses the latest one or not. I don't know what other program use the latest versions and I'm sure the ones that might are more audio centered like sound forge and don't really integrate well with stand alone rippers and id3 taggers, like audiograbber and would be a real pain to have to manually do each step for each track by hand.
If you use Exact Audio Copy you can use the codecs that are supplied with either Cool Edit Prov2.1 dated April, 18 2003, or the one from Cool Edit 2000. Just copy the fht.flt file from either product and paste it in your EAC directory. Voila! Now you can rip encode and id3 tag using the latest Fraunhofer technology.
