Topic: Audacity vs. Sound Forge post-processing

[EWizard]

I've gotten some awesome vinyl record rips with my Sony PCM-M10 recently.
I'm shooting for peaks of -3db in my recordings.
I have both Audacity and Sound Forge (which came with my recorder).
What I'm wanting to do is trim out the front and rear of my recording, which is basically the beginning and end of an LP side.
Sound Forge will let me do this fine, but the track separation and normalizing do not seem to work as well as in Audacity.
Audacity's downfall is that it appears that I cannot save changes to my .WAV file directly, and have to export at a lower resolution (16 bit, or FLAC, MP3, etc).
Sound Forge will let you save changes directly to the raw recording file.

The post-processing I would like to apply to my files are pretty basic.
1)  Raise the volume of my recordings, so that they peak at 0db.
2)  Trim out the beginnings and ends of the LP side.
3)  Separate tracks

Can someone give me some direction on the best way to achieve this, while retaining my 24bit/48k resolution?

[yug du nord]

IMO..
Audacity is a great free program that can do what you need it to do.
Sound Forge is better.  What version??
Download CD Wave Editor (free) for track splitting.
Download Trader's Little Helper (free) for file conversion.

[bryonsos]

Once you do your editing in Audcity: file > export multiple tracks will do what you want. You'll get a dialogue box where you can select bit rate etc. Audacity is quirky.

[Gutbucket]

Once you normalize the peak level of your files (a little under 0dBFS is better that maxing out right at zero) the resulting dynamic range from any LP will easily fit into that provided by a 16bit file.  There is no advantage whatsoever in storing the files in 24bit format that point, only wasted storage space consisting of nothing but additional low level noise.  If peaking at around -3dBFS during the transfer, you could probably record directly to 16bit without noticing any increase in the noise-floor of the resulting files.  Recording to an higher bit depth could be advantageous if you were not setting levels as accurately and were peaking much lower during the transfer, as the M-10 is probably capable of a real world dynamic range of maybe 18 bits or so when set to write 24bit files.

[EWizard]

IMO..
Audacity is a great free program that can do what you need it to do.
Sound Forge is better.  What version??
Download CD Wave Editor (free) for track splitting.
Download Trader's Little Helper (free) for file conversion.

I have version 9, which came with my Sony PCM-M10.
I need to find some tutorials for how to go about upping the volume, etc.
Thanks for the info.

[chinariderstl]

I wrote the following up a long time ago.  It works well for me.  I grabbed most of it (verbatim) from one of their manuals.


# Sony Sound Forge
------------------------

# Resample to a New Sampling Rate

To change the sampling rate of an audio file, do the following:

1. Process > Resample

2. Set the new sampling rate for your audio file by using the New Sample Rate parameter.

3. Set the Interpolation Accuracy parameter.
    This parameter specifies the accuracy of the resampling process.
    A higher setting provides slower but more accurate processing.
    Unless you have a really long audio file, you probably want to keep this parameter set to 4.

4. IMPORTANT: If you are converting from a higher sampling rate to a lower sampling rate,
    be sure to activate the 'Apply an Anti-Alias Filter' during Resample option.
    This prevents any high frequency content from the file with a higher sampling rate
    from becoming noise in the converted file.

5. Click the Preview button to hear how your file will sound.

7. Click OK.

# Normalizing Audio

To use the Normalize function, do the following:

1. Select the data in your audio file that you want to normalize.
    If you want to process the entire file, select it all by choosing Edit > Select All (or by pressing Ctrl-A).

2. Choose Process > Normalize

3. For the 'Normalize Using' parameter, choose the 'Peak Level' option.

4. Click the Scan Levels button to find the highest amplitude level in your audio data.

5. Adjust the 'Normalize to' parameter by dragging its slider up or down, which sets the highest amplitude level
    to the level you'd like it normalized.

6. Click the Preview button (optional).

7. Click OK.

# Fade Audio

To apply a fade-in or fade-out to your audio data, follow these steps:

1. Select the data in your audio file to which you want to apply a fade.

2. To apply a fade-in, choose Process > Fade > In.

3. To apply a fade-out, choose Process > Fade > Out.

# Adjust Audio Volume - The Volume Function (Adjust Gain or Amplify)

To simply increase the amplitude of a data selection, use the Volume function.

1. Select the data in your audio file to which you want to apply amplitude changes.
    If you want to process the entire file, select it all by choosing Edit > Select All (or by pressing Ctrl-A).

2. Choose Process > Volume to open the Volume dialog box.

3. To adjust the amplitude of your data, set the Gain parameter.
    Move the slider up to increase amplitude.

4. Click the Preview button (optional).

5. Click OK.

[EWizard]

I don't know if I'm an idiot or not but I cannot figure how to install plugins into Sound Forge.
I'm looking for the Click and Crackle plugin.
Are there any free ones that are worth a darn?

[ilduclo]

be sure to do most of your edits while still in 24 bit, especially normalize.  I agree that one program doesn't seem to do it all for me. I use soundforge for normalize, resample and bit depth conversion, cool edit pro for trims, fades, cuts, applause reduction (hard limit or envelope) then CD wave editor for track cutting, TLH for flac conversion (works on the 24 bit original archive files, too) and CDWav for mp3 conversion.

[EWizard]

Another question.
For converting LP's to digital, applying Normalization to the digital file it seems is a debated topic.
The debate is whether or not you are "destroying" the dynamic range when applying Normalization.
How is this different than applying Volume?
It appears to me to be the same thing.  Is one better than the other for keeping the dynamic range of the LP?

[chinariderstl]

be sure to do most of your edits while still in 24 bit, especially normalize.  I agree that one program doesn't seem to do it all for me. I use soundforge for normalize, resample and bit depth conversion ...

Precisely, as do I!

[EWizard]

Also, is there any reason not to resample and just export my 24/48 .wav file as .flac via Audacity?
Audacity has a de-clicking effect and an Amplify which seems to be the same thing as Volume/Normalize(the whole file) in Sound Forge.
Audacity will also export individual tracks.
It doesn't seem that Sound Forge will do either of these things (export to .flac or split files or de-click).
Why use Sound Forge instead of Audacity?

[Gutbucket]

is there any reason not to resample and just export my 24/48 .wav file as .flac via Audacity?

That will work.  Primarily, resampling to 44.1kHz and truncation/dither to 16bits reduces the need for excess storage space that isn't providing any additional fidelity, since that additional bandwidth isn't being used.  It can also provide greater player compatibility.  If you don't care about storage space or compatibility, just FLAC the 24/48 files directly.  Likewise, you could transfer at 16/44.1 (or 16/48 if you like), avoiding the truncation (and sample rate conversion) steps, with the same fidelity, given the source material and the good level control you are exercising during the transfer.

A note on file bit depth verses working program calculation bit depth- Most modern audio editors make their internal calculations at a higher bit depth than the bit depth of the file itself.  They are usually working at something like 32bit float internally, regardless of whether the file is 16, 20, 24, 8 bits or whatever.  Keeping the file in a higher bit depth can be important if you are saving and transferring between different pieces of software and making audio changes in each of them.  If you are doing all the work in one program however, or if you are exporting the file and then only doing things which do not require complex calculations (such as tracking in CDWav, FLACing in TLH or whatever) you can go ahead and save at the final target bit rate of 16bits without any penalty.  The editing software will truncate and dither from it's internal 32bit float or whatever down to 16bits or whatever you specify when you export the file.

[EWizard]

Are you saying that a FLAC file created from a 24bit source could be incompatible with some players?

[Gutbucket]

If it can decode FLAC it will probably play a 24 bit file.

16/44.1 WAV is the undoubtedly most universal format, playable anywhere, including being dinosaur disc compatible.

[EWizard]

I sure wish my Sony could play .flac.
Would make my life easier.

[pohaku]

FIIO X1 or X3.  They will play about anything.  I still like the Apple interface the best, but the FIIO players will take micro sd cards (I now have a 128G card in my X3) so the capacity problem is diminished.  I have a 160G iPod Classic that is full.  With the FIIO, I can just add a card and swap them out periodically.  And they sound good - better than the iPod.

[dnsacks]

Apple Iphone/ipad will play up to 24/48 apple lossless files just fine.  It's relatively easy to set up foobar2000 to convert flac to apple lossless in a windows environment (and to have foobar2000 use the sox resampling algorithm to resample/dither files) and very easy to use XLD to do the same in Mac osX -- happy to provide additional details/links to those that are interested.  I'm now tracking/processing/archiving everything at 24/96 and using these methods to generate files that are best for other purposes (i.e. 16/44.1 for burning to cd, 320 bitrate mp3 for mobile listening, 24/48 apple lossless for higher res portable listening, etc.

Another option . . .   

[OhioHead]

Thank you for the below, I am very much a NEWB when it comes to post production.

Do you suggest "tagging" the songs in SoundForge to make the tracking easier in tracking software?  What do you gain with normalization?

Anybody still you "Magic" audio editing software?

I wrote the following up a long time ago.  It works well for me.  I grabbed most of it (verbatim) from one of their manuals.


# Sony Sound Forge
------------------------

# Resample to a New Sampling Rate

To change the sampling rate of an audio file, do the following:

1. Process > Resample

2. Set the new sampling rate for your audio file by using the New Sample Rate parameter.

3. Set the Interpolation Accuracy parameter.
    This parameter specifies the accuracy of the resampling process.
    A higher setting provides slower but more accurate processing.
    Unless you have a really long audio file, you probably want to keep this parameter set to 4.

4. IMPORTANT: If you are converting from a higher sampling rate to a lower sampling rate,
    be sure to activate the 'Apply an Anti-Alias Filter' during Resample option.
    This prevents any high frequency content from the file with a higher sampling rate
    from becoming noise in the converted file.

5. Click the Preview button to hear how your file will sound.

7. Click OK.

# Normalizing Audio

To use the Normalize function, do the following:

1. Select the data in your audio file that you want to normalize.
    If you want to process the entire file, select it all by choosing Edit > Select All (or by pressing Ctrl-A).

2. Choose Process > Normalize

3. For the 'Normalize Using' parameter, choose the 'Peak Level' option.

4. Click the Scan Levels button to find the highest amplitude level in your audio data.

5. Adjust the 'Normalize to' parameter by dragging its slider up or down, which sets the highest amplitude level
    to the level you'd like it normalized.

6. Click the Preview button (optional).

7. Click OK.

# Fade Audio

To apply a fade-in or fade-out to your audio data, follow these steps:

1. Select the data in your audio file to which you want to apply a fade.

2. To apply a fade-in, choose Process > Fade > In.

3. To apply a fade-out, choose Process > Fade > Out.

# Adjust Audio Volume - The Volume Function (Adjust Gain or Amplify)

To simply increase the amplitude of a data selection, use the Volume function.

1. Select the data in your audio file to which you want to apply amplitude changes.
    If you want to process the entire file, select it all by choosing Edit > Select All (or by pressing Ctrl-A).

2. Choose Process > Volume to open the Volume dialog box.

3. To adjust the amplitude of your data, set the Gain parameter.
    Move the slider up to increase amplitude.

4. Click the Preview button (optional).

5. Click OK.

[chinariderstl]

Note, and for the record, I have updated that document.  If you are using that as a basis, I would suggest using the following revision instead.

Also, you can read about Normalization at the following URL:

http://manual.audacityteam.org/o/man/normalize.html

HTH


Sony Sound Forge
----------------


## Resample to a New Sampling Rate

To change the sampling rate of an audio file, do the following:

1. Process > Resample

2. Set the new sampling rate for your audio file by using the New Sample Rate parameter.

3. Set the Interpolation Accuracy parameter.
    This parameter specifies the accuracy of the resampling process.
    A higher setting provides slower but more accurate processing.
    Unless you have a really long audio file, you probably want to keep this parameter set to 4.

4. IMPORTANT: If you are converting from a higher sampling rate to a lower sampling rate,
    be sure to activate the 'Apply an Anti-Alias Filter' during Resample option.
    This prevents any high frequency content from the file with a higher sampling rate
    from becoming noise in the converted file.

5. Click the Preview button to hear how your file will sound.

7. Click OK.

8. File > Save As (Important: Do not hit Save, this will over-write the original source file).


## Decrease the Bit Rate

To change the bit rate, do the following:

1. Process > Bit-Depth Converter

2. Select the appropriate Bit Depth from the Drop-down box.

3. Select the appropriate (High-pass Triangular) value from the Dither Drop-down box.

4. Select the appropriate (High-pass contour) value from the Noise Shaping Drop-down box.

5. Click Ok.

6. File > Save As (Important: Do not hit Save, this will over-write the original source file).


## Normalizing Audio

To use the Normalize function, do the following:

1. Select the data in your audio file that you want to normalize.
    If you want to process the entire file, select it all by choosing Edit > Select All (or by pressing Ctrl-A).

2. Choose Process > Normalize

3. For the 'Normalize Using' parameter, choose the 'Peak Level' option.

4. Click the Scan Levels button to find the highest amplitude level in your audio data.

5. Adjust the 'Normalize to' parameter by dragging its slider up or down, which sets the highest amplitude level
    to the level you'd like it normalized.

6. Click the Preview button (optional).

7. Click OK.

8. File > Save As (Important: Do not hit Save, this will over-write the original source file).


## Fade Audio

To apply a fade-in or fade-out to your audio data, follow these steps:

1. Select the data in your audio file to which you want to apply a fade.

2. To apply a fade-in, choose Process > Fade > In.

3. To apply a fade-out, choose Process > Fade > Out.

4. File > Save As (Important: Do not hit Save, this will over-write the original source file).


## Adjust Audio Volume - The Volume Function (Adjust Gain or Amplify)

To simply increase the amplitude of a data selection, use the Volume function.

1. Select the data in your audio file to which you want to apply amplitude changes.
    If you want to process the entire file, select it all by choosing Edit > Select All (or by pressing Ctrl-A).

2. Choose Process > Volume to open the Volume dialog box.

3. To adjust the amplitude of your data, set the Gain parameter.
    Move the slider up to increase amplitude.

4. Click the Preview button (optional).

5. Click OK.

6. File > Save As (Important: Do not hit Save, this will over-write the original source file).

[gratefulphish]

Another question.
For converting LP's to digital, applying Normalization to the digital file it seems is a debated topic.
The debate is whether or not you are "destroying" the dynamic range when applying Normalization.
How is this different than applying Volume?
It appears to me to be the same thing.  Is one better than the other for keeping the dynamic range of the LP?

I know that I am responding late, but you know, better late . . .

As I understand things, and I think I understand them, normalization is a process of raising the volume of the entire track/file up to the limit set by the loudest sound in the whole file.  So, if you wanted to keep your max level at -1db, and your loudest sound on the recording was at -6db, then the program would first find that loudest sound, subtract the max limit, and then raise the volume of the entire track by 5db.

Volume is similar, but potentially slightly different, as you are then specifically raising the volume of an entire track by a set number of dbs.  Theoretically, if you didn't see a particularly loud note, and added too many dbs, you could then create a clip where there was none on the original recording.  But, as long as you stay below the 0db limit, the net effect is essentially the same.

In neither case is the dynamic range affected, as the same amount of additional volume is applied to the entire recording.  Think of it as turning the volume up or down on your listening device.  It is just making it louder or quieter, but it is not affection the total range.  You are just raising the overall threshold level, so that the volume can be louder, particularly if you run your levels low when recording.

What does affect dynamic range is compression, which a whole other topic, which too many people don't understand IMHO, of course, and avoid, when it is a practical necessity in reproducing what really happened in the room.  Our ears are not microphones. 

[Gutbucket]

^ Steve is correct. (Hi Steve)

And provides a good summary of standard "peak normalization", which which increases the level of all content in a file until the highest level peak reaches your target threshold.  It only throws away the "empty space at the top", and brings up the noise-floor at the bottom along with the rest of the content.

Normally (pun), normalization will not reduce the dynamic range of the file, but in some cases, processes which are called "normalization" can do so.  It all depends on the specifics of the tool applying it and what it's doing.

Confusing things somewhat are some alternate normalization tools, sometimes called "RMS normalization" or the like, which in contrast to "peak normalization" raise the RMS or "averaged" level of the file up to some target threshold.  Because it's acting based on longer-term average levels, the shorter-term peaks may clip, or depending on the algorithm, they may be compressed or limited above the threshold to avoid clipping. That's actually two things, normalization + limiting, both done by a single tool.  The noise-floor at the bottom is raised along with the content, but there may not be enough room at the top so things get either chopped off or squeezed above a threshold to fit.  So RMS nomalization or the like will reduce dynamic range whenever the compression or limiting cuts in.

It would be better if those tools were more descriptive and didn't simplly call themselves "normalization", but that's the real world.

Caveat emptor / caveat lector.

[gratefulphish]

^^^

Hi Lee, you are right as well, but I was not going to cloud the explanation by including RMS as well, as that will tend to confuse many non-sound nerds.  But, as you noted, using RMS normalization runs that risk of still clipping, unless you run an RMS peak check first, and make sure to set the level so that it won't cross that 0db line.

But, I am a fan of compression done well, as the end product sounds so much better, and is truly far more representative of what we hear in the room than a recording limited by the artificial 0db barrier that our ears don't have.

[Gutbucket]

No doubt! Most all of our live recordings benefit from some dynamic range compression (if done well, which can be tricky, sometimes best left alone).  EWizard's LP transfers will be best left with as much dynamics as the LP contains, considering that the dynamics of those recordings has already been rather highly manipulated, sometimes for the sound of it, yet always to fit it within in the rather limited dynamic range available on LP, which isn't nearly as much as what is available with a 16 bit WAV file.