Problem: (SOLVED) MP3's inconsistency in lengths.

[goddi]

Greetings,

I am having a problem with an mp3 music file that I want to use in a slideshow. It is so strange, that I am not sure how to explain it except to just show what I have been experiencing with it. The main problem is that I see different lengths (in minutes and seconds) for this file when comparing them in MyComputer, Audacity (after Exporting it) and in PTE.

I took this one mp3 into Audacity only to look at it to trim off any dead end spaces but in the end I only Exported it. The original file had a bitrate of 21, so I Exported it at a bitrate of 24 to get as close to the original bitrate as possible. So only the file size changed slightly. The two file names are KhoaSok-1-21.mp3 (the original mp3) and the other KhoaSok-1-24.mp3 (the original file Exported at 24 bitrate using Audacity).

Here are 6 observations of these 2 files that make no sense to me (let's call one '21' and the other '24'):

1- In MyComputer, '21' shows a length of 10:31 and '24' shows a length of 8:52.

(See attached: 1-KhoaSok-MyComputer.gif)

To add to the confusion,

Media Player Classic: '21' shows 14:13; '24' shows 8:52.

Windows Media Player: '21' shows 10:32; '24' shows 8:52.

Quicktime: '21' shows 8:52; '24' shows 8:52.

2- In Audacity, '21' and '24' both show lengths of 8:52.

(See attached: 2-KhoaSok-Audacity.gif)

3- In PTE, '21' shows a length of 14:13.

(See attached: 3-KhoaSok-21-Timeline14-13.gif)

4- In PTE, '24' shows a length of 8:52.

(See attached: 4-KhoaSok-24-Timeline8-52.gif)

5- In this mp3, there is a drum solo that starts somewhere in the music file.

In PTE's Timeline, the drum solo starts at 8:08 in '21' but in '24', it starts at 5:05.

(See attached: 5-KhoaSok-21and24-DrumSolo.gif)

6- In '21', the PTE Timeline flatlines at about 8:51 all the way to the end, at 14:13 (but the drum solo is playing until the end at 14:13, even though flatlined).

(See attached: 6-KhoaSok-21-Timeline8-51.gif)

If anyone can shed some light on these inconsistencies, I would appreciate it. I have attached all the gifs and will try to attach the original mp3. It looks like the original file, '21' has some 'problems' and that Audacity 'fixed' them when it was Exported? Just really curious why I see all these differences in lengths.

Gary

KhaoSok-1-21.mp3

[Lin Evans]

Hi Gary,

If you look very closely at your mp3 in Audacity, you will notice that there is a "repeat" programmed near the end. One of the features of Audacity is to allow the user to repeat a segment and the drum solo, near the end has this programmed. All mp3 players are not sensitive to this, but PTE is and can perform this function. I removed the repeat from your mp3 and it should read the same and work properly now.

Here's the link:

http://www.lin-evans.org/gary/gary.zip

Happy New Year,

Lin

[goddi]

Hi Gary,

If you look very closely at your mp3 in Audacity, you will notice that there is a "repeat" programmed near the end. One of the features of Audacity is to allow the user to repeat a segment and the drum solo, near the end has this programmed. All mp3 players are not sensitive to this, but PTE is and can perform this function. I removed the repeat from your mp3 and it should read the same and work properly now.

Here's the link:

http://www.lin-evans.org/gary/gary.zip

Happy New Year,

Lin

============================

Lin,

I have looked 'closely' but I really don't see what apparently you see. Can you be more specific as to what to look for in the original '21' file? I looked at the very end, at 8:52, of the original mp3 but I don't see anything.

Thanks... Gary

ADDED LATER: I noticed that your file was saved at 64kbps (4Mbs) as opposed to 24kbps (1.6Mbs) , as mine was. Was this just 'automatic' or do you really hear the difference? I just want a smaller file so I used 24kbps.

[Guest Yachtsman1]

Hi Gary

Don't know if this helps?

yachtsman1

[goddi]

Hi Gary

Don't know if this helps?

yachtsman1

==========================

Yachtsman1,

I think you are referring to the bitrate part of my question. I know you can set the bitrate when you are doing the Export process. However, you seem to be pointing to the Default Sample Rate set in the Quality tab. Not sure what the Sample Rate really is but I don't think it has to do with the bitrate that you can change when Exporting an MP3.

I'd just like to know what Lin sees that indicates there is a 'repeat program' when he viewed my mp3 file.

Thanks... Gary

[cgbraggjr]

A general comment: in my experience mp3s with low bit rates perform erratically. For instance, I've had mp3 players and software that simply refuse to play anything recorded below 32K. Since space is cheap these days, why not experiment by having Audacity generate 64K files?

[nobeefstu]

Gary,

If anyone can shed some light on these inconsistencies, I would appreciate it.

The duration inconsistencies are most likely attributed to File 21 being encoded in VBR and the File 24 being encoded in CBR.

VBR files often need scanning and partially decoding of the entire file to obtain accurate estimates of the bitrate to best approximate the duration because the metadata maybe tagged differently in the file. Audacity decoded the entire file for full editing and thus reflected the correct duration(s) for both files.

Stick with encoding your files to CBR ... since this seems to be the best for all players and PTE to accurately reflect the proper duration.

[goddi]

Gary,

The duration inconsistencies are most likely attributed to File 21 being encoded in VBR and the File 24 being encoded in CBR.

VBR files often need scanning and partially decoding of the entire file to obtain accurate estimates of the bitrate to best approximate the duration because the metadata maybe tagged differently in the file. Audacity decoded the entire file for full editing and thus reflected the correct duration(s) for both files.

Stick with encoding your files to CBR ... since this seems to be the best for all players and PTE to accurately reflect the proper duration.

=============================

Nobeefstu,

I didn't know what VBR was so I poked around in the program that I used to create the File 21. I had used a program called All Sound Recorder Vista. I had used its default settings to make the recording off of what was playing on my PC. I looked into its Options menu and found that it has a default setting for recording MP3s, which is 'Standard (VBR)". In that drop-down menu, it also shows "insane (CRB)". I think that would make me not want to chose that setting.....

But I see that it also has the setting choice of "Use CBR". Maybe that is the one I should use. It defaults to 128 kbps. I will try that setting out next time I use the Recorder program. Is "128" the appropriate choice?

Now, how could you tell from the File21 that VBR was used to encode it????? I don't see it in any of the Properties items.

Also, does Lin's comment on a "repeat" programmed into the file still something to do with the variations in lengths? Still not sure what he meant.

Please see attached gifs for better understanding of the above posting.

Thanks... Gary

P.S. This really is fascinating. The more I screw up, the more I learn!!!

[fh1805]

Gary,

You asked about 128kbps. I use 192kbps for all my MP3s. I can hear a difference in quality between 128kbps and 192kbps when played back on a good quality pair of speakers.

regards,

Peter

[Guest Yachtsman1]

HI Gary A long time in another galaxy etc, Peter produced an Audacity guide, which has a number of points that refer directly to the problem you have created. See page 5 of Peters guide.

Yachtsman1

audacity tech.pdf

[Guest Yachtsman1]

Gary,

You asked about 128kbps. I use 192kbps for all my MP3s. I can hear a difference in quality between 128kbps and 192kbps when played back on a good quality pair of speakers.

regards,

Peter

Hi Peter

Can you clarify how you do that.

Regards Eric

Yachtsman1.

[nobeefstu]

Gary,

In the attachment file I just used Media Player Classic - Home Cinema which many users have to view the information.

As you compare the two files ... the orig file has a very limited information table as compared to the converted file . You will also notice the orig file VBR information table has no direct reference to the time duration where the converted file CBR does.

I too have used All Sound Recorder Vista. I have attached my settings used. I also prefer 192kbps over 128kbps.

[fh1805]

Can you clarify how you do that.

Hi Eric,

In Audacity, as part of File...Export processing. After giving the file its desired name and specifying the file type to be MP3 rather than WAV, click on the Options button (bottom of the three) and then use the drop-down selection for the Quality. N.B. These instructions are based on Audacity v1.3.11 beta. Note also that this is where you can set the Bit Rate (Constant - which I use, Variable, etc.). A couple of years ago, I saved the same music clip under each bit rate from 128 through to 320. I could hear a difference up to 192 but could not hear the difference between that and 224. I therefore decided that 192 would be sufficient for my needs.

regards,

Peter

[Guest Yachtsman1]

Hi Peter

Unfortunately I can't seem to follow what you are suggesting, I use 1.3.12 beta, which I assume differs from yours. Also I use MP3's as my source mainly. The only time I have imported WAV's is from Soundsnaps & a sound effects CD I have. I have included a couple of screen shots of my audacity export pages, one using Gary's MP3 & one using one of my WAV files.

Regards Eric

Yachtsman1

PS Gary, I've also included a SS of your clip illustrating the Hz rate on yours which I believe should be 44,000.

[davegee]

Options?

DG

[Guest Yachtsman1]

Hi Peter

I've found the options button now see ss, I've also created a couple of MP3 exports, one standard one upgraded. I know the samples are not suitable subjects for detecting improvments, I did it just to illustrate how.

Regards Eric

Yachtsman1

Airplane, Biplane-start up 128br.mp3

Airplane, Biplane-start up after raising bit rate.mp3

[goddi]

Gary,

In the attachment file I just used Media Player Classic - Home Cinema which many users have to view the information.

As you compare the two files ... the orig file has a very limited information table as compared to the converted file . You will also notice the orig file VBR information table has no direct reference to the time duration where the converted file CBR does.

I too have used All Sound Recorder Vista. I have attached my settings used. I also prefer 192kbps over 128kbps.

=======================

Nobeefstu,

Super. I did not realize that I could get that info using the Properties of Media Player Classic. Very nice.

And thanks for the the screen shot of your settings for All Sound Recorder Vista. That nails it down. Use CRB + 192.

This little 'problem' that I had with this one little MP3 really has revealed a lot for me.

Sincerely, Gary

[goddi]

...PS Gary, I've also included a SS of your clip illustrating the Hz rate on yours which I believe should be 44,000.

============================

Yachsman1,

I did not notice that. I'll keep my eye on that setting too.

Thanks... Gary

[goddi]

Hi Eric,

In Audacity, as part of File...Export processing. After giving the file its desired name and specifying the file type to be MP3 rather than WAV, click on the Options button (bottom of the three) and then use the drop-down selection for the Quality. N.B. These instructions are based on Audacity v1.3.11 beta. Note also that this is where you can set the Bit Rate (Constant - which I use, Variable, etc.). A couple of years ago, I saved the same music clip under each bit rate from 128 through to 320. I could hear a difference up to 192 but could not hear the difference between that and 224. I therefore decided that 192 would be sufficient for my needs.

regards,

Peter

================================

Peter,

I did a quick test using different setting for the same MP3 file. I did not do a listening test, but I just wanted to see the file size differences.

For the same file:

1- CRB + 192 Kbps + 44.1 Khz = 12,482K

2- CRB + 64 Kbps + 11.025 Khz = 4,162K

3- CRB + 24 Kbps + 11.025 Khz = 1,561K

Big differences in the final file size. I'll have to try this on a good MP3 to see if I can tell the difference in sound quality.

Gary

[fh1805]

Guys,

I sense it is time for some basic facts about digital sound files...

I'd like you to visualize a simple sine wave showing on an oscillascope: comprising just one sweep from the baseline, up to maximum positive amplitude, down through the baseline to minimum negative amplitude and back to the baseline. Also, assume that this wave form lasts for just one second of duration. I have deliberately chosen the oscillascope for this example because it illustrates one very important point that is easily overlooked: sound isn't digital, it's analogue! There is no such thing as digital sound: there is only digitized sound. To store the sound as a file in our computer, it has to go through a process to convert it from an analogue to a digital representation of that analogue sound. This is done by an analogue-to-digital converter (ADC) routine.

Using our one second of sine wave as our source, the ADC process cuts this up into slices (think slices of meat at the deli counter). The more slices it cuts, the higher the quality of sound reproduction will be. This is called the Sample Rate. CD quality reproduction is achieved at 44,100 slices per second (44.1KHz). So Gary's pieces at 11.025KHz will not be of the same sound quality as the piece at 44.1KHz.

Having sliced the sound file up, the ADC now has to encode each slice. It has a choice of how many bits (binary digits - 0s and 1s) to use to encode each slice. This is called the Sample Format. CD quality sound files are usually encoded at 16-bit. Note that the files on a commercially-produced CD are encoded as WAV files. A WAV file is the sound equivalent of a TIFF image file. Note also that Audacity supports Sample Formats of 16-bit, 24-bit and "32-bit float". Arguably the 24-bit and "32-bit float" options should give even better sound quality. But now we come up against the human factor.

The human ear can distinguish sounds in the range 20Hz-20KHz (approximately) provided that there is no impairment to the hearing (e.g. from an excess of loud disco music in your youth or a natural loss of performance with age). The combination of a Sample Rate of 44.1KHz and a Sample Format of 16-bit is more than adequate to capture accurately the range of frequencies that are audible to the human ear.

Whatever software you use to do your soundtrack assembly, I encourage you to set the sample rate and sample format to the values I've indicated: 44.1KHZ and 16-bit.

So what about the 128kbps vs 192kbps debate? These are "compression" values for MP3 files (for an analogy, think JPEG compression in your image editor). The only time you set this value is as you save the output file. Because MP3 compression is a "lossy" compression just like JPEG compression, there will be some noticeable degradation of the final output compared to the original.

regards,

Peter

[davegee]

Peter,

I agree, but I go a little further.

MP3 - 256 - 320 Kbps (I don't only use MP3 for PTE!).

JPEGs - 100% Quality

No compromise!!

DG

[Ken Cox]

YEARS AGO

Ronnie West was trying to teach me about sound - -my solution was to go out and buy a sound meter with digital readout

-- record the same track/album at different bit rates

-- then with sound meter and different cd's, i took all to my truck and ran tests with truck running and played each album

-- proof was in the readings -- big difference from 128 to 320

Ronnie was correct )

i rip at 256 - good enough for my ears and equipment

ken

[Big Kev]

Guys,

I sense it is time for some basic facts about digital sound files...

I'd like you to visualize a simple sine wave showing on an oscillascope: comprising just one sweep from the baseline, up to maximum positive amplitude, down through the baseline to minimum negative amplitude and back to the baseline. Also, assume that this wave form lasts for just one second of duration. I have deliberately chosen the oscillascope for this example because it illustrates one very important point that is easily overlooked: sound isn't digital, it's analogue! There is no such thing as digital sound: there is only digitized sound. To store the sound as a file in our computer, it has to go through a process to convert it from an analogue to a digital representation of that analogue sound. This is done by an analogue-to-digital converter (ADC) routine.

Using our one second of sine wave as our source, the ADC process cuts this up into slices (think slices of meat at the deli counter). The more slices it cuts, the higher the quality of sound reproduction will be. This is called the Sample Rate. CD quality reproduction is achieved at 44,100 slices per second (44.1KHz). So Gary's pieces at 11.025KHz will not be of the same sound quality as the piece at 44.1KHz.

Having sliced the sound file up, the ADC now has to encode each slice. It has a choice of how many bits (binary digits - 0s and 1s) to use to encode each slice. This is called the Sample Format. CD quality sound files are usually encoded at 16-bit. Note that the files on a commercially-produced CD are encoded as WAV files. A WAV file is the sound equivalent of a TIFF image file. Note also that Audacity supports Sample Formats of 16-bit, 24-bit and "32-bit float". Arguably the 24-bit and "32-bit float" options should give even better sound quality. But now we come up against the human factor.

The human ear can distinguish sounds in the range 20Hz-20KHz (approximately) provided that there is no impairment to the hearing (e.g. from an excess of loud disco music in your youth or a natural loss of performance with age). The combination of a Sample Rate of 44.1KHz and a Sample Format of 16-bit is more than adequate to capture accurately the range of frequencies that are audible to the human ear.

Whatever software you use to do your soundtrack assembly, I encourage you to set the sample rate and sample format to the values I've indicated: 44.1KHZ and 16-bit.

So what about the 128kbps vs 192kbps debate? These are "compression" values for MP3 files (for an analogy, think JPEG compression in your image editor). The only time you set this value is as you save the output file. Because MP3 compression is a "lossy" compression just like JPEG compression, there will be some noticeable degradation of the final output compared to the original.

regards,

Peter

Excellent explanation peter - it is a good reminder to all of us although when I explain it I usually use slices of cake !!!!!

By the way - have tyou ever had a problem with the waveform on one of the stereo channels being not centred on the 0db line??

I think the problem is being caused by this extremely old computer with a not very good sound card - any ideas???

John

[Big Kev]

Guys,

I sense it is time for some basic facts about digital sound files...

I'd like you to visualize a simple sine wave showing on an oscillascope: comprising just one sweep from the baseline, up to maximum positive amplitude, down through the baseline to minimum negative amplitude and back to the baseline. Also, assume that this wave form lasts for just one second of duration. I have deliberately chosen the oscillascope for this example because it illustrates one very important point that is easily overlooked: sound isn't digital, it's analogue! There is no such thing as digital sound: there is only digitized sound. To store the sound as a file in our computer, it has to go through a process to convert it from an analogue to a digital representation of that analogue sound. This is done by an analogue-to-digital converter (ADC) routine.

Using our one second of sine wave as our source, the ADC process cuts this up into slices (think slices of meat at the deli counter). The more slices it cuts, the higher the quality of sound reproduction will be. This is called the Sample Rate. CD quality reproduction is achieved at 44,100 slices per second (44.1KHz). So Gary's pieces at 11.025KHz will not be of the same sound quality as the piece at 44.1KHz.

Having sliced the sound file up, the ADC now has to encode each slice. It has a choice of how many bits (binary digits - 0s and 1s) to use to encode each slice. This is called the Sample Format. CD quality sound files are usually encoded at 16-bit. Note that the files on a commercially-produced CD are encoded as WAV files. A WAV file is the sound equivalent of a TIFF image file. Note also that Audacity supports Sample Formats of 16-bit, 24-bit and "32-bit float". Arguably the 24-bit and "32-bit float" options should give even better sound quality. But now we come up against the human factor.

The human ear can distinguish sounds in the range 20Hz-20KHz (approximately) provided that there is no impairment to the hearing (e.g. from an excess of loud disco music in your youth or a natural loss of performance with age). The combination of a Sample Rate of 44.1KHz and a Sample Format of 16-bit is more than adequate to capture accurately the range of frequencies that are audible to the human ear.

Whatever software you use to do your soundtrack assembly, I encourage you to set the sample rate and sample format to the values I've indicated: 44.1KHZ and 16-bit.

So what about the 128kbps vs 192kbps debate? These are "compression" values for MP3 files (for an analogy, think JPEG compression in your image editor). The only time you set this value is as you save the output file. Because MP3 compression is a "lossy" compression just like JPEG compression, there will be some noticeable degradation of the final output compared to the original.

regards,

Peter

Apologies for the none capitalized Peter and the odd typo - I have a dyslexic keyboard too!!

John

[fh1805]

Hi John,

No, I've never had an off-centre waveform. The Normalize Effect will allow you to re-centre to 0dB. But I guess you already new that!

Peter