| Audio |
Codecpage News |
Compression
MP3, MP3-5.1
Ogg
Vorbis
MP2
MP4
(AAC)
AC3
Sample
rate conversion and Asynchronity
Squeezing/Stretching
Filtering
with Cool Edit (Adobe Audition)
Magix
Audio Cleaning
Lab
Cutting
and merging MP3 or MP2
Repairing
digital TV audio streams
Other
useful audio tools
Audio
test files
Compression
No video without audio. Compressed, of course. Getting a good audio compressor however is a problem, at least if it should be a codec. Codecs have the advantage that they can be used by several programs for simultaneous video and audio processing.
For capturing or
compression with MS
Media Encoder, MS audio
is recommendable. Good even at 64 kbps (160kbps
is very good to capture TV music programs later to be converted
to mp3). This however refuses to work in file formats other than
ASF and WMV. Both of these formats however tend to stagger at
playback, much more so than AVIs. DivX;-) 3.11 contained a hacked
MS audio engine that would work with AVIs, BTW.
MP3 is not only the most common standard for audio compression, it's also the most used in connection with DivX video of any kind.
Maybe you can get a
software package
containing the discontinued FhG professional mp3 codec that usually
has the file name L3CODECP.EXE. If you have this file, just execute
to install the codec. Audioactive Production Studio contained
it, for example.
For legal reasons, I cannot recommend to use any of the hacks
circulating on the net. The quality of this old FhG codec is also
not so good, it can really produce garbage with certain sources.
The question remains why FhG have abandoned selling any VFW codecs.
Their new engine would be much better and faster than the old
one.
Instead, they license plugins for some 3rd party software. The
disadvantage: you have to split audio and video, then encode audio
only, and merge them again.
Nero Burning ROM from
Ahead
offers an MP3 plugin for as little as
$15 (Nero itself is generally much cheaper in
stores than
at their website, BTW). This plugin is far better and faster than
the old VFW codec (10x real time and more). Uncheck 'allow downmixing'
(this is a lowpass filter that's really too strong).
The new mp3
pro engine for Nero could
be even better but I haven't
yet tried it (available at ahead.de).
With no DirectShow filter available for playback, I'll think twice
before I pay them a second encoder. What really sucks is that
Nero 5.5.x is stubborn in setting 80 kbps mp3pro as default for
any single file now. Just as stubborn as it switches off joliet
and restores all kinds of imbecile burning options at every upgrade.
The
best mp3 option is VBR
done with the FhG engine in CoolEdit2000
(or plugin for Cool Edit Pro) by Syntrillium
(unfortunately, they sold to Adobe;
The latest Cool Edit Pro 2.1 is now almost identical, Adobe Audition
1.0. This is expensive, but Cool Edit also never has been cheap
(more below)). Also very fast. Peculiar that the default low pass
filtering makes files bigger and worse, at least for the average
bitrates. Use none (22050). Switch off intensity joint stereo
if possible. If you do so and set quality to 50%, you get something
like 80...90 kbps at yet an impeccable quality.
CBR 128 kbps (also filter 22050) is also superior to any other
I've heard with this bitrate.
The freeware Lame is also quite good but low vbr rates, but
CoolEdit
is considerably better even if Lame is set to slowest encoding
(5 times slower than CoolEdit or Nero). A good feature of Lame
is the ABR encoding (use Razorlame to set it), it allows to tune
bitrates very low. It can go down to 64k and less, Useful for
already bad soundtracks. At 80-120 k, CoolEdit is clearly better.
Look for Lame on the web and also have a look at the utility software
at divx.com.
Of course, you have to merge the audio with the video after separate
encoding. You can't open mp3 directly in VirtualDub but you can
give it a WAV header, then it works. There seems to be only one
tool on the web that does it quickly and reliable: WaveMP3.
It is very hard to find now and the author's web page is down,
so as an exception from our general policy I have decided to carry
this absolutely legal and free tool, with his permission of course.
Another way is to use nandub
(from nandub.org) to merge video and mp3. This
can even handle variable bitrate (vbr) mp3. A vbr mp3 from CoolEdit
for example can still be excellent for movie soundtracks at about
80 kbps.
Nandub can also make dual language AVIs (2 mp3 streams). For playback, get the Morgan Stream Switcher. With this, Media Player will suddenly offer language selection.
With DivX4/5
being ISO compliant, the development
goes to true MPEG4 streams which will also contain MP4 audio.
Recently, FhG-IIS
have released their
MP3 Surround format. It can compress 5.1 channels in
160
kbps or above, and the resulting files can be played with older
2 channel decoders as well. An evaluation software is for download
at their site. The encoder is valid until
the end of
2005, the decoder is unlimited.
This is a standard
audio compression
format for MPEG2 and DVD video, at least in PAL countries (for
NTSC DVDs, only AC3 is standardised, but MP2 will work in almost
any case).
A good freeware mp2 encoder is toolame.
You may find it at doom9. Make sure you check "show all audio tools" in
the downloads.
BeSweet+Belight is another very versatile choice (probaly the best).
If you need a sample
rate converter, you may also checkSSRC.
Both Toolame and SSRC are command line programs but can be
automatically
called by TMPGEnc encoder and are reportedly good. There is also
a GUI (Windows frontend) available for toolame.
Another (older) option would be Germanix Transcoder (google for downloads). Sort of 'Swiss army knife' for audio transcoding. Seems only to support 192 kbps with mp2 though, and I haven't yet done quality tests with it.
There also are codec
solutions for MPEG2
The Qdesign codec, for example, was not cheap but super fast and
therefore good
for capturing with slow computers. Bad quality below 160 k however.
Ogg
Vorbis
A very good new audio
format (my favorite
now) is Ogg
Vorbis
Audio
1.0. It's a free, open-source, high compression audio format
that's a real alternative to mp3. Recently, the German c't magazine
has tested several audio codecs including mp3pro, WM9, MP4-AAC,
RealMediaSurround. Ogg Vorbis 1.0 beats them all at 64 as
well
as 128 kbps. Ogg is normally VBR, BTW. The Ogg
DirectShow filter, enables MediaPlayer
to select between multiple audio streams (languages) if so intended.
(MP4 has meanwhile improved a lot and catched up OggVorbis at
least, see below)
I have recently done
my own tests with
OggVorbis and c't is right: Almost incredible, the absolute lowest
possible quality setting, with < 40 kbps, is already really
good for audio tracks that do not contain sophisticated music.
64 kb are good for music, 80 kb are excellent.
I also use this for music archiving now with rates 100-120 kbps.
Well supported by Winamp as well. There are also some CD rippers
converting to Ogg (see Ogg site).
I can really recommend to use OggVorbis audio together with DivX5: No better compression anywhere.
To encode audio, I
use OggdropXpd.
To multiplex with video, OggMux is quite handy. You can mux
as many
audio tracks as desired. For example, different language and additional
comment tracks, all in different bitrates if useful.
The resulting file will play in all versions of Windows media
player, sound tracks can be selected with an icon that appears
in the task bar.
Ogg files can also be
cut after production.
OggCut
is one tool for this.
VirtualDubMod
is more powerful and a must if you want to further edit your .ogm
files. It can also do a LOT more, have a look at their site.
BTW, Ogg media also supports subtitle streams. I will not cover this on this site.
What is also not quite in focus here but remarkable: Tools for 5.1 surround support are under development, and Ogg will support many more formats. Compared with AC3, Ogg compresses several times better.
This is the new audio
standard for MPEG4
program streams, and the newest version (HE-AAC) competes closely
with Ogg Vorbis. In my opinion, it sounds a bit more original
at lowest bit rates.
Nero6 has an AAC encoding module that can also deliver HE-AAC,
and there is also a Winamp AAC plugin. Latest Winamp
versions (5.33) have a prety good built
in HE-AAC encoder and are much easier to use for format conversions
now. Best of all, it's freeware. I guess HE-AAC will be widely
used, other than the MP3pro format that was a complete flop from
the beginning, due to lack of software.
Note that AAC has not
become part of the BluRay standard. It is only found in file formats
like mp4 video.
AC3 (5.1 channel
surround in most cases,
but can also be plain stereo) is the audio format most found on
DVDs. Ifou don't
need multichannel or your video standard is PAL (European) you should
be fine with MP2 audio. If your video standard is NTSC however (US)
and you want to be 100% compatible with vintage players, you may
want to use AC3 at least instead of the other alternative, uncompressed
audio (it may
however still not be so helpful, as DVD players so ancient that
they don't accept MP2 may also be unable to play self burnt DVDs
at all). AC3 is also part of the BluRay standard. Hence, AC3 is most
universal as a format (also if you encode just two channel
stereo).
AC3Filter
is a freeware set of tools that includes Directshow filters and
also a Winamp plugin that can be used for playing AC3 sound.
AC3
encoding
can be done with several free tools, for example ffmpeggui
(comes with an older ffmpeg.exe version).
Probably
the best free Solution is BeLight: this newer GUI to BeSweet is easy
to use and allows to transcode many audio formats (the old BeSweet GUI
was quite bizarre). You need to download BeSweet
and then BeLight
and copy the BeLight files into toe BeSweet folder.
AC3Filter
is a freeware set of tools that includes Directshow filters and
also a Winamp plugin that can be used for playing AC3 sound.
Nero also offers AC3 encoding in some upper versions. OEM versions
of Nero (called Nero Express) don't have the audio conversion
function (be aware that in the US, these versions are nevertheless
often sold way overpriced and with a description that tries to
pretend that you get the full version).
Sample rate conversion and Asynchronity
48/44.1 kHz conversion often introduces asynchronity. Why?
Well, 48/44.1=1.0884354.....
Hard
to implement in
a conversion routine. An accuracy of about 5 digits is mandatory
to avoid any time deviation. So the reason for this problem is
quite obvious. Even a very small inaccuracy will cause the sound
to deviate from the picture after some time.
(Some people stated the soundcard was the problem and a PCI soundcard
would solve the problem. Nope. No soundcard hardware is ever used
by any of the conversion programs.)
Correcting the problem:
1)
use a good conversion
program. For example, if you encode something with FlaskMPEG,
always stick to the source bitrate and use something better, like
CoolEdit, to convert later on.
There is a very command line conversion tool, SSRC, that can be used with
TMPGenc
or with BeSweetGUI.
An easy way to convert sample rate and file format in one
step, is with Germanix
Transcoder (google for download sites).
BeSweet+BeLight is the latest and probably also the best choice for
most tasks. You
need to download BeSweet
and then BeLight
and copy the BeLight files into toe BeSweet folder.
2) do a direct stream copy with VirtualDub and set a little different video frame rate. Not the most elegant method but it works. May cause you problems if you want to do further work on the file later on. Disencouraged.
Conversions from cinema frame rate (24fps) to PAL TV (25 fps) are usually done by simply playing the movie faster. The speed difference can hardly be noticed, but the audio cannot simply be accelerated because the change in pitch would easily be heard by many people.
For Video, only the framerate byte in the AVI header has to be changed, either with an avi tweaking tool (instantly, but be careful with some tools before you tested them thoroughly) or by copying or converting the file with Virtualdub (in most cases, frame size also has to be changed, which requires a recompression).
The solution for audio is a special conversion, changing speed while maintaining pitch.
There are 3 tools I
can name for this:
- Adobe Audition (former Cool Edit)
- Germanix Transcoder (google for downloads)
-
BeSweet+BeLight also offer some speed change options (download BeSweet
and then BeLight
and copy the BeLight files into toe BeSweet folder).
I haven't done tests with GermaniX in that discipline, but for Cool Edit I can say that with "medium" quality setting, results are very good.
Filtering with Cool Edit (Adobe Audition)
Cool Edit was way the
best audio software,
however it has been taken over by Adobe, as 'Adobe Audition',
and now there is no Cool Edit 2000 version anymore. Audition is
pretty expensive, but it's street price is only half as much as
Cool Edit Pro was, and only little more than the almost 140$ that
Cool Edit 2000 with only one add-on (filters, really necessary)
had been. This is still pretty pricey for amateur use, however
I fear that there is really no alternative if you really want
some professional audio filtering and cleaning. With newer version
available meanwhile, you may bargain for the 1.0, which is entirely
adequate.
The peculiar part of it: Cool Edit was only known and sold to
a few people, but since it's with Adobe, everyone wants it.
Hum
A typical problem
with TV recordings
is intercarrier hum. TV cards tend to have this, and it's especially
nasty because this is like a square wave, with lots of harmonics.
I've recently seen a comparison of audio tools and the guys really
tried to use a notch filter to eliminate hum. As such a test is
zilch, I can't tell from it which products are recommendable.
I know however that
CoolEdit (now Adobe
Audition) is good, so here is how to use it on hum:
Use the 'noise reduction' tool.
Mark a silent part anywhere in the soundtrack and 'get samples
from selection'
Set reduction level to 100% below 1 kHz and let it drop to below
50% above 1 kHz.
Hiss
You could eliminate
hiss and hum in one
pass with the noise reduction tool. However, the part above 1
kHz may reduce hiss very well, but there is great danger of getting
warbling sound artifacts (Many professionally remastered CDs also
have this).
Experiment with the preview function; use good headphones to determine
the best settings.
You may get better results using 'noise reduction' on hum only
(up to about 5oo Hz) and tackling noise with a threshold filter
(named 'hiss reduction' in CoolEdit). Again, sample a silent part
and then set the threshold carefully.
Another good approach is to use several steps of moderate 'noise reduction' in sequence. I got some very good results with this but it's time consuming.
Clicks/Crackles
CoolEdit has a
terrific declicker as
well, and its MP3 encoder is the best that I've seen. The ideal
tool to rescue your old vinyl albums.
For declicking, again sample a silent part, that only contains
clicks, then let it do a 2-pass run. If you get equal or more
'rejected' than 'corrected' clicks, you're OK.
With the MP3 encoder, never allow any low pass filtering (always set filter to 1/2 of sample rate), and never allow intensity joint stereo. Both make things worse, at any bit rate. Mid-Side joint stereo is always OK.
The best cheaper
alternative to the almost
perfect Audition is Magix Audio Cleaning lab. It has easy to use
and very efficient filters for noise and crackle, as well a numerous
other functions.
Easy to use and you can't do much wrong trying it at this price.
Cutting
and merging MP3 or MP2
MP3trim is
a tool that can cut MP3 without recompression and, unbeknownst
to most, also does MP2. We open the file with it, select to cut
of frames from the beginning, and click until we have selected
enough frames to roughly equal 480 ms (in this case). Now save
the file, and quickly and without recompressing anything, the
job is done.
Should we encounter a positive delay, another tool called MP3merge
also is capable of MP2, and we only need some silence encoded
with the right bitrate (toolameGUI) to cut and paste some of it
before our audio file.
MP3directcut
is a freeware tool that also cuts MP3 as well as MP2, but other
than the above it has a graphical wave display and many more functions.
Repairing
digital TV audio streams
Some audio problems
that may cause error
messages with really picky applications as GUI for dvdauthor or
DVDstyler, can be corrected with mp3val-frontend, that even does
many files
at once, or, if it's AC3, with Delaycut. Delaycut
can also synchronize audi files by adding or cutting parts at
the beginning and end.
With mpeg2 DVB recordings, it is advisable always
to apply these tools, as a precaution. With HD recordings, the
procedure may cause asyncheronity, here it should be regarde as a last
resort.
Another issue: sometimes one appliation opens the audio file and another won't. This could be a problem with the file header. Then it may help to open the file in MP3trim, cut off just one frame (0.02s) from the beginning, and save the file. This may repair the header and it does not cause any visible asynchronity. Again, mp3trim can open mp2.
Some digital TV
transmissions may have
nasty audio streams (fortunately this has become very rare): Switching
between stereo and AC3, for example.
Stoically reading over all the bumps is CoolEdit or Adobe Audition,
if you choose multitrack mide, right click on a track and select
"insert audio from video".
Save as wav and recrompress.
Here are a few
freeware
tools
to maintain an mp3 collection:
MP3gainGUI can adjust the volume of multiple mp3 files or folders without recompression, reversible any time and obeying physiological loudness. A real must-have !
Rname-it (popup infested homepage) is a very easy and powerful multiple file renamer that can be handy for any image or sound collection.
ID3master comes in handy to manage mp3 file info tags.
MP3tag offers a lot more and cannot only handle mp3 tags including cover images, but also tags of ogg, mp4, or flac files.
These are samples for testing. You may use them to check the frequency response of your amplifier, sound card or mp3 encoder, for example.
Test sweep from 20 Hz - 20 kHz
Perfect mp3 version of the above (this is entirely sufficient to download if you want to use an mp3 player or if you have a software that converts mp3 back to wav).
Copyright
(C) 1998-2011; all
rights reserved. All
materials in these pages are presented for scientific evaluation
of video technologies only. They may not be copied from here and
used for entertainment or commercial activities of any kind.
We do not have any relation to and do not take any responsibility
for any software and links mentioned on this site. This website
does not contain any illegal software for download. If we, at
all, take up any 3rd party software here, it's with the explicit
permission of the author(s) and regarding all possible licensing
and copyright issues, as to our best knowledge. All external download
links go to the legal providers of the software concerned, as
to our best knowledge.
Any trademarks mentioned here are the property of their owners.
To our knowledge no trademark or patent infringement exists in
these documents; any such infringement would be purely unintentional.
If you have any questions or objections about materials posted
here, please e-mail us
immediately.
You may use the information presented herein at your own risk
and responsibility only. We do also not guarantee the correctness
of any information on this site or others and do not encourage
or recommend any use of it.
One further remark: These pages are covering only some aspects
of PC video and are not intended to be a complete overview or
an introduction for beginners.