Audacity problem

[Marianne]

Hi,

I have a weird problem. I use Audacity for compiling my music. I've done this several times and never noticed anything wrong. But now I have 5 mp3 fragments which I want to join. Whenever I cut and paste the speed of the track becomes different. I can hear it, because the sound gets either higher or lower, but I can also see it because in a track of nearly 4 minutes it can make a difference of up to 5 seconds!

I have tried several things to find out what causes this speeddifference. Thought maybe the different quality of the mp3's were to blame (the bitrates differ from 128 to 320). Tried to convert them to wav before loading them into Audacity, but that doesn't make any difference.

I finally sort of solved it by changing the speed by hand after the pasting. But this takes a lot of time in comparing the original with the final Audacity file. And I don't find it a very accurate way.

I really can't tell if this problem was there before, because I usually don't use too much singing behind my shows. And I think in instrumental tracks the difference will be much more unnoticed.

Any of you soundtechnicians have an idea?

[DaveG]

Try importing each mp3 into Audacity separately and saving again at the same "bit rate" each time?

Then when you try to assemble them at least you can eliminate the likelihood of the different "bit rates" causing the problem.

Incidentally, like JPEGs, repeated saving of mp3 files will eventually corrupt the files to such an extent that distortion etc will creep in.

You should always start with WAV files when doing this just as you should always start with a lossless format such as TIF or RAW when working on images.

DaveG

[Kurt S]

You were on the right track by converting them to waves first but you need to look at the specs of the audio.

What is probably happening is that the MP3's were created from waves that varied in bitrate or sampling rate.

Decode the MP3's to wave and look at the specs. The standard should be 44.1khz, 16 bit, stereo. If the wave file indicates anything different, use the wave editor to change it.

[Ed Overstreet]

I've encountered this problem repeatedly in Audacity not only with some MP3 files but with some WAV files, like those recorded by the "voice annotation" feature on some Nikon cameras and battery grips. With the unusual WAV files I found the simplest solution was to convert them to MP3 using dBPower Amp conversion software, then open the MP3 in Audacity and proceed from there. If I had an MP3 that wasn't playing back properly in Audacity, I'd try converting it to WAV first (using the same sampling settings mentioned above) and then open the WAV conversion in Audacity; that should work.

[Conflow]

Marianne,

When working with (Commercial) Wav or MP3 Audio Files one must always be aware of the following, I shall keep it simple:-

1)...Whether it's Wav or MP3 there is a mandatory 2 secs of 'Lead-In' in each Audio Track.

2)...To the Listener this 'Lead-In' sounds to be a 'silence period' where in fact it contains sub-liminal 128bit Data Signals which are arranged as 'Tags' which imparts to your PC or Player vital information about 'Bit-Rate', Subjective Loudness, Compression Range, Format, Artist Name, Song Title, Cddb Library data, Copyright and so on and so on.

3)...You must understand that Wav and Mp3's are designed to "Fool" the Human Ear into hearing things that are simply not within the Music any more ~ It's complex !

3)...Editing or Cropping out that 'Lead-In' makes Audacity default down to the lowest 'Bit Rate' it had encountered unless you instructed Audacity to save the finished Edited File as XXX Bit Rate in Stereo or Mono. Consequentlly of course the final Edit will usually be shorter in 'Time-Span' that the time span of all tracks in series.

4)...After 'Multiple Edits' and 'Save As' all digitally Compressed Files loose their 'Hi-Frequencies' and super-compress the 'Lowest Frequencies' and flatten out the dyanamic range of the Music but leave the 'Peaks' intact.

This sounds to the Human Ear as a distinct change in 'Pitch' (or speed) due to the loss of the dyanamic frequency range of the Music. (More about this can be learnt from Fletcher-Munsen Curves in any Audio-Acoustics Book and desertations on MP3 Formats).

5)...When Audio Editing always use 'Save' and only use 'Save As' for the final compilation ~ Much like a JPeg ~ this will greatly reduce the 'phantom effect' you hear.

Hope this goes some way in explaining this phenomenon...

Brian.Conflow.

[Kurt S]

"Whether it's Wav or MP3 there is a mandatory 2 secs of 'Lead-In' in each Audio Track."

No, there is a Redbook CD audio format that needs to have a 2 second gap on the first track only. It has nothing to do with wave or mp3 files.

To the Listener this 'Lead-In' sounds to be a 'silence period' where in fact it contains sub-liminal 128bit Data Signals which are arranged as 'Tags' which imparts to your PC or Player vital information about 'Bit-Rate', Subjective Loudness, Compression Range, Format, Artist Name, Song Title, Cddb Library data, Copyright and so on and so on"

Nope. There is no file data in the lead in, hell there is no lead in. The data that the file requires to be played properly is stored in the file header, not in the audio stream.

"You must understand that Wav and Mp3's are designed to "Fool" the Human Ear into hearing things that are simply not within the Music any more ~ It's complex"

Again, wrong. Wave format is an uncompressed format that does not use "psycoacoustics" like MP3's to fool the listener.

"Editing or Cropping out that 'Lead-In' makes Audacity default down to the lowest 'Bit Rate' it had encountered unless you instructed Audacity to save the finished Edited File as XXX Bit Rate in Stereo or Mono. Consequentlly of course the final Edit will usually be shorter in 'Time-Span' that the time span of all tracks in series."

Editing a wave file does not edit the header information or lower the file to it's lowest bitrate, nor do you see a two second gap in the editor that containds the header info.

"After 'Multiple Edits' and 'Save As' all digitally Compressed Files loose their 'Hi-Frequencies' and super-compress the 'Lowest Frequencies' and flatten out the dyanamic range of the Music but leave the 'Peaks' intact."

Nope, Non- lossy WMA, Flac and Monkey's Audio file are all non-lossy compressed and they ca be editied as many times as you like without degredation in quality.

"When Audio Editing always use 'Save' and only use 'Save As' for the final compilation ~ Much like a JPeg ~ this will greatly reduce the 'phantom effect' you hear"

Only true if you are working with lossy formats like MP3. If you are editing wave files, you can do a save and not suffer loss.

I don't like to argue with other forum members, especially members like Conflow who have offered a lot of valuable information in the past but I have to call you on this one. Where on earth did you get this information. Nothing you have stated here is correct.

[Leif]

I don't have enough time to explain ...

I wish you had the time! This sounds very interesting!! So please come back when and if you have the time!

[Conflow]

Kurt,

I stated in my Posting..."That I was keeping things simple for the Purpose of Explaination"....

If you so choose to give a fuller, more Complex Technical explaination about Mp3 Formats and their Coded Headers ~ please do so ~ and while you are at it please indicate where the File header is and why there is a 2sec Intertrack Gap in virtually all Commercial CD-Recordings.

And please,please let the readers know all about the degradation of Mp3's when they are over-edited and over 'Saved As'....and most important what happens to the 'Musical pitch' in these circumstances. That is what this topic is all about....

On other matters...

The 'Redbrook CD Format' to which you refer is a comparitive newcomer to CD-Music Technology - it has nothing to do with the way the WAV is made - it's simply a 'User' information stream about the Artist, Song Title etc,etc. AND MOST OF ALL it's geared towards Copyright and ways of preventing people from copying a CD which they have paid for.

Furthermore with Wav Recordings (pure audio) ~althought they are not digitally encoded, but usually Digitally Recorded in the Studio, and for the most part they are certainly 'highly compressed' usually for the purposes of fitting more Tracks on the CD. If you didn't do that and allowed the full 60db dynamic range you would probably end up with 5 or 6 Tracks on the CD.

But yet you maintain that WAV CD's are pure Audio Streams....If memory serves me correctly that went out with 8, 16, 32 Track Studio Tape Recorders such as Studer, Ampex EMI etc. I know because I was in the Studio Business in those days when we had pure Audio.

Furthermore, I never mentioned WMA (Windows Media Audio) which is a 'Hybrid Format' neither did I mention Flac nor Monkeys Audio ~ two more 'Hybrids' ~ which I am well aware of, being an inveterate user of the "QCD Media Player" and its many Codecs.

Now tell me what have these got to do with "Audacity Sound Editor" and Mp3's and WAV's ?? ~because Audacity can't handle them~ and I don't know anybody who uses these Formats in PTE Presentations.

So instead of taking my simplified Post (selectively quoted by you) out of context, pehaps you would like to offer an explaination to Marianne for the...'apparent change in pitch'...of her edited Mp3 Files ?...I gave her an explaination, whats your's ?

Regards,

Brian.Conflow.

[Kurt S]

You are not keeping it simple, you are totaly wrong on your information.

"and while you are at it please indicate where the File header is and why there is a 2sec Intertrack Gap in virtually all Commercial CD-Recordings."

The file header information is stored in the file but it is not part of the audio stream. The two second gap is built into the CD TOC (table of contents) not the actual file .This is why once you rip the song from the CD there is no 2 seconds of blank audio in the track.

"And please,please let the readers know all about the degradation of Mp3's when they are over-edited and over 'Saved As'"

Yes, on an MP3, WMA, Ogg Vorbis (lossy format compression) it will degrade every time it is resvaed, buth thats not what you were saying. You also indicated waves were this way as well.

"and most important what happens to the 'Musical pitch' in these circumstances".

I already stated why that is probably happening in a previous post. It has nothing to do with editing the first two seconds of auio. Ther eis no file information stored in the audio stream. You can edit it to your heart's content and it won't change the pitch or lower the bitrate. The bitrate info, stero/mono/ bit depth is all stored in the "non-editable" header. A completely separate part of the audio file.

The reason she is experiencing change in pitch is that she is working with audio file that have a different bitrate or bit depth. You have to convert them all to the same rate before you can paste them together.

"Furthermore with Wav Recordings (pure audio) ~althought they are not digitally encoded, but usually Digitally Recorded in the Studio, and for the most part they are certainly 'highly compressed' usually for the purposes of fitting more Tracks on the CD"

"On other matters...

The 'Redbrook CD Format' to which you refer is a comparitive newcomer to CD-Music Technology - it has nothing to do with the way the WAV is made"

First off, CD's don't use wave files. They use raw PCM data. Raw PCM data does not have a header with info about the file the way a wave file does. This is why the redbook standard was invented. The player does not need to know what the bitdepth, sample rate, or number of channels because the CD follows the standard and the player uses this standard to play the CD.

"The 'Redbrook CD Format' to which you refer is a comparitive newcomer to CD-Music Technology "

I can't beleive what your saying, where are you getting this info. The redbook format is not new, it was developed in 1982 and is the standard that Sony/Phillips used for CD audio. It describes the specifications of the CD, the data rate, the storage size, etc of an audio disc. It was invented before the audio CD, not a newcomer.

"it's simply a 'User' information stream about the Artist, Song Title etc,etc. AND MOST OF ALL it's geared towards Copyright and ways of preventing people from copying a CD which they have paid for."

Not true. It isn't just user information, it's specifications on how the CD works. Copywrite? No, there is no copywrite specifications wriiten in the redbook standard. The redbook standard never had copy protection in mind. All the copywrite scemes that have been writen for audio CD are non-redbook compliant. This is why many players are having problems with copywritten discs. They don't conform to standards.

"and for the most part they are certainly 'highly compressed' usually for the purposes of fitting more Tracks on the CD"

They certainly are not compressed. Show me one article that says they are. The only way you get more info on an audio CD is to break the redbook standard of 650 megs and make the disc 700+ megs. There is no compression written into the redbook standard.

"If you didn't do that and allowed the full 60db dynamic range you would probably end up with 5 or 6 Tracks on the CD."

60db? A good quality cassette deck will give you 60db, CD's are up around 105. Besides, you are talking about two different types of compression. Compression/limiting the dynamic range has nothing to do with compressing the file to reduce the size. Two totally different subjects and one has nothing to do with the other. Nor does limiting the dynamic range reduce file size.

"But yet you maintain that WAV CD's are pure Audio Streams....If memory serves me correctly that went out with 8, 16, 32 Track Studio Tape Recorders such as Studer, Ampex EMI etc. I know because I was in the Studio Business in those days when we had pure Audio"

I'm not quite sure how to answer this one. Show me where I said they were pure audio streams. Yes I was also around in those days. That was called "analog recording". This statement is completely irrelevent to our conversation so I'll move on.

"So instead of taking my simplified Post (selectively quoted by you) out of context, pehaps you would like to offer an explaination to Marianne for the...'apparent change in pitch'...of her edited Mp3 Files ?...I gave her an explaination, whats your's ?"

Well, I already explaind it in the third post of this topic. She is trying to mix together audio file with different sample rate or bit depth files.

[Marianne]

Whoop! I didn't mean to raise a row here!

Dave:

Try importing each mp3 into Audacity separately and saving again at the same "bit rate" each time?

I didn't try resaving them because I know that mp3's act the same as jpeg's. Each time you resave them they lose a little more quality. When I work on a project I save the project until I am sure I am ready to resave it as mp3 to use in P2Exe.

Brian:

Editing or Cropping out that 'Lead-In' makes Audacity default down to the lowest 'Bit Rate' it had encountered unless you instructed Audacity to save the finished Edited File as XXX Bit Rate in Stereo or Mono. Consequentlly of course the final Edit will usually be shorter in 'Time-Span' that the time span of all tracks in series.

Whether or not you are right on this one, it can't cause my problem because I also used a whole track without editing anything. So I am pretty sure I didn't remove the mandatory lead-in of the track.

Ed:

convert them to MP3 using dBPower Amp conversion software

I have used dBPower Amp to convert the original low quality (bitrate 128) mp3 to wav. Then opened the file in Audacity. Whenever it is mixed with another file the change of speed is noticeable. So not in the originally opened Audacity file, but only after the copy and paste action! I have compared them to make sure this is really so.

Kurt:

Decode the MP3's to wave and look at the specs. The standard should be 44.1khz, 16 bit, stereo. If the wave file indicates anything different, use the wave editor to change it.

I looked again at the specs. But the conversion to wav is standard set to these figures. I didn't have original wav files, so that is also not a solution.

I am very glad with all of your help. But I am beginning to think this is a matter of slightly corrupted files indeed.

Furthermore, if I use these mp3's and resave them all with a bitrate of 192, what effect does that have on the original, only 128, files? I know that the files which where originally 320 are compressed more, but I don't think there is a way to decompress the low quality files to a level of 192. Or is this like it is with jpeg that I should better use the highest bitrate to resave? But doesn't that make my musicfile too big to handle for P2Exe?

I am not always making my own mp3's, because than I would just use the not lossy WAV. And, come to think of it, when I use tracks of my own cd's, which I convert myself to WAV and then to mp3 I encounter no problems....

Thank you all for your very extensive answers!

Marianne

[Ken Cox]

get your theory here

http://www.cdrfaq.org/

and

http://www.mrichter.com/cdr/welcome.htm

the last one seems to be unavailable right now

but mike and andy are kind of guru's in the cd burning field and am sure they must cover this highly tech music info

ken

[Conflow]

Hi Marianne,

Yes, Kurt and I had a 'bit' of a debate, and in response I am always mindful of the fact that the vast majority of PTE Forum Readers are non-technical so I tend to 'over-simplify' things and use analogies to explain what may be going on.

Kurt on the other hand wishes to have 'things' word perfect and quite rightly so, but one can always 'pick holes' in another's argument - thats O.K for a Debating Society, but I must confess that I steer well clear of such activities when posting to this Forum.

In relation to your 'sampling rate' the adopted Standard is 128kb/sec as set out in ISO/IEC 11172-3 Layer 3.

Major work was done on this in the late '90 by the Fraunhofer Institute at the behest of the ISO and IEC Authorities where it was given official recognition. Sure, there are other sampling rates but it is the considered opinion of those Authorities that nothing substantial (note) was to be gained with the higher rates ~ sorry I won't get into another debate on this.

For those who are interested, a full dessertation on Mp3's can be had from the Fraunhofer Institute and the Lame Organisation to mention but a few.

In closing I must reiterate what I said...You must tell 'Audacity' what 'bitrate' it is to 'Save As'...it does precisely that once told, otherwise it defaults to other settings.

Brian.Conflow.

[Kurt S]

Conflow, my intent was never to start an argument but to possibly inform Marianne of the possiblility that the info you provided was not going to help since it was incorrect. I even tried to compliment you at the end of my post.

I will however say one last comment to you Conflow. You are highly un-informed about audio and CD standards. Your mis-information was serving no purpose other then to confuse Marianne even more. I was try to offer a good reason, not made up speculation such as what you have posted.

Keeping it simple is always the best (KISS) but simplicity does not involve mis-information. A wise man will learn from others. A fool will continue to believe what he wants and ignore facts.

Anyway, I harbor no ill will and I hope you don't either. If I have offended you, I appologize. Debates can be good, you can learn many things.

[Ken Cox]

http://www.mrichter.com/

http://www.mrichter.com/cdr/welcome.htm

is working at 16:20 edst

ken

[Conflow]

Kurt,

I do not understand your Post about an 'argument' ~ I thought we were having a debate ?

Concerning Marianne's problem, I recorded one defective Track from Mariannes Link(s) with 'Audio-Recorder ProV3' and put it through 'Norton Wave Editor' and found a progressive linear Fade-out which quite obviously must have been caused by something...perhaps a consequence of her Camera Batteries running down...I found no evidence of different 'sample bitrates'.

As to offending me, you certainly did not, as to your other allegations about my CD-knowledge, I recently posted a 'Technology Link' to www.imation.com (CD Manufacturers) which went into the whole 'gambit' of Commercial CD-Production which can be verified by many readers of this Forum, so perhaps your comments were a little hasty.

Sure, I agree with you that Debate is good ~ provided personal assumptions are constrained and not made 'de facto'~

Regards,

Brian.Conflow.

[Marianne]

I recorded one defective Track from Mariannes Link(s)

Hm Brian,

Aren't you confusing me with someone else? I never posted a link. Never even posted a show (yet). Where did you get that track?

Marianne