Hi,

Unfortunately I got mp3 files at 44.1kHz sample rate and 96k bit rate. I want to do some editing on them, little noise removal, some EQ. Would it make any difference if I converted them to wav first so there is no further degradation when I do work on them? I do know that I'm not making a better quality file by converting it wav. That's not my intend. I'm just hoping that whatever I get in mp3 doesn't keep degrading when I edit the file. I'm not sure if there is any quality loss when converting from mp3 to wav. I'm hoping that once I do work on it in wav format I'm only going to degrade it further when I save it to mp3 again. Is this true? Or is this a waste of time?

In general every cycle of recompression to a lossy file format results in losses, how significant the losses are will depend on the compression processes and sample rates/depths used. Even working on digital files in a uncompressed format will result in some cumulative losses due to truncation/rounding errors in the A/D process

To minimize losses do all editing/sweetening working with a lossless format files with a bit depth that is hopefully greater than you final distribution format. Example, if you input files are 16-bit, I would use 24-bit intermediate files for all editing, than convert to the distribution format as the very last step.

If your editing happens all in one app, and just saving one time again to an .mp3, then mp3->mp3 is best. If you intend to use different apps for signal processing, encountering multiple saves, then an intermediate .wav is a good idea.

I'm not sure if there is any quality loss when converting from mp3 to wav.?
There should not be any further quality loss transcoding an MP3 to PCM (or 'Saving as in the PCM format), typically WAVE <.wav> .It will not increase the quality though, though many folks mistakenly think it will. The analogy I use to explain it is, "putting one gallon of milk in a 5 gallon container... but it's is still a gallon".
Lossy file types (MP3, AAC, AC-3, ect.) should avoided, except for encoding the final end-user file.

OTOH, some NLEs and DAWs (like Vegas Pro), automatically build floating point 32 bit PCM proxy files when the project's timeline is created, so transcoding is no really necessary, the original file is never overwritten.
So it depends on one's software. When in doubt, transcode lossy audio files to a PCM format.

The downside of the audio compression already there from the mp3 files is that it's going to be like working with the audio equivalent of mud when it comes to fixing issues.

Tell the client you will do your best, but all bets are off if they supply compressed audio to begin with.

Andrew

There should not be any further quality loss transcoding an MP3 to PCM (or 'Saving as in the PCM format), typically WAVE <.wav> .It will not increase the quality though, though many folks mistakenly think it will. The analogy I use to explain it is, "putting one gallon of milk in a 5 gallon container... but it's is still a gallon".
Lossy file types (MP3, AAC, AC-3, ect.) should avoided, except for encoding the final end-user file.

OTOH, some NLEs and DAWs (like Vegas Pro), automatically build floating point 32 bit PCM proxy files when the project's timeline is created, so transcoding is no really necessary, the original file is never overwritten.
So it depends on one's software. When in doubt, transcode lossy audio files to a PCM format.

Thanks. Yes, I understand, as I stated in my original post that there is no increase in quality by converting from mp3 to wav. If there is no data, there is no data, you can't create something from nothing.

The downside of the audio compression already there from the mp3 files is that it's going to be like working with the audio equivalent of mud when it comes to fixing issues.

Tell the client you will do your best, but all bets are off if they supply compressed audio to begin with.

Andrew

Yes, I'm aware of that and so is the client. Thanks

In general every cycle of recompression to a lossy file format results in losses, how significant the losses are will depend on the compression processes and sample rates/depths used. Even working on digital files in a uncompressed format will result in some cumulative losses due to truncation/rounding errors in the A/D process

To minimize losses do all editing/sweetening working with a lossless format files with a bit depth that is hopefully greater than you final distribution format. Example, if you input files are 16-bit, I would use 24-bit intermediate files for all editing, than convert to the distribution format as the very last step.

Yes, I'm aware of all of this except that I have no choice but to work with mp3 at the moment, hence my question.

If your editing happens all in one app, and just saving one time again to an .mp3, then mp3->mp3 is best. If you intend to use different apps for signal processing, encountering multiple saves, then an intermediate .wav is a good idea.
This makes sense. Thank you!

If it's being opened in an audio editor, it's probably being decompressed anyway. Converting as a separate step is redundant. Deliver a WAV file and tell the client not to compress it unless absolutely necessary. Create the project at the sample rate at which you intend to deliver the WAV. It might be good to find out what the intended use is so you can deliver the WAV with the correct bit depth and sample rate.

If it's being opened in an audio editor, it's probably being decompressed anyway. Converting as a separate step is redundant. Deliver a WAV file and tell the client not to compress it unless absolutely necessary. Create the project at the sample rate at which you intend to deliver the WAV. It might be good to find out what the intended use is so you can deliver the WAV with the correct bit depth and sample rate.
Intended use is podcast, so they need mp3.

Intended use is podcast, so they need mp3.

Then find out the exact parameters of the MP3 in order to prevent the need for more conversion, or deliver WAV and let them convert to the right MP3 themselves. Be sure they understand that unsatisfactory audio quality is most likely the result of their decision to provide a low quality original file.

Hi,

Unfortunately I got mp3 files at 44.1kHz sample rate and 96k bit rate. I want to do some editing on them, little noise removal, some EQ. Would it make any difference if I converted them to wav first so there is no further degradation when I do work on them? I do know that I'm not making a better quality file by converting it wav. That's not my intend. I'm just hoping that whatever I get in mp3 doesn't keep degrading when I edit the file. I'm not sure if there is any quality loss when converting from mp3 to wav. I'm hoping that once I do work on it in wav format I'm only going to degrade it further when I save it to mp3 again. Is this true? Or is this a waste of time?

For simple quick conversion just select "Audio Converter" then add all your files in this online audio app and then download the converted files. You can select the encoder you want plus set sample and bit rates to your choice,

This app also can do many other things as you will see. Bonus? It's all free. :)

https://vocalremover.org/converter

Chris Young

Good mp3 files are not always terrible. You need to assess what you have, and importantly, what you will then fo yo them again. Every conversion to a compressed file destroys more, so if your file ends up as .wav, and stays as .wav that’s best, but if it’s going to be compressed again, then a plan is needed to retain as much as you can.

Good mp3 files are not always terrible. You need to assess what you have, and importantly, what you will then fo yo them again. Every conversion to a compressed file destroys more, so if your file ends up as .wav, and stays as .wav that’s best, but if it’s going to be compressed again, then a plan is needed to retain as much as you can.
Yeah, that's what I'm asking, how to convert with a minimal loss. Do you have an answer?

For simple quick conversion just select "Audio Converter" then add all your files in this online audio app and then download the converted files. You can select the encoder you want plus set sample and bit rates to your choice,

This app also can do many other things as you will see. Bonus? It's all free. :)

https://vocalremover.org/converter

Chris Young
I know how to convert files. I don't need this tool

Then find out the exact parameters of the MP3 in order to prevent the need for more conversion, or deliver WAV and let them convert to the right MP3 themselves. Be sure they understand that unsatisfactory audio quality is most likely the result of their decision to provide a low quality original file.
I know the parameters, but I still need to convert for final delivery (whether it's me converting or them, the conversion needs to happen). This isn't about me vs them. We all understand the downsides of low quality file and just want to preserve as much as we can.

Kathy - you are worrying about nothing in real terms. I finished a project and delivered it - client happy, and it was a music product, so quality was vital. I'd done 5 previous ones and 2 more afterwards. Three months later, I was using it as a template and noticed the filename in the Premiere clip collection - audio23v2.mp3 - I'd used the mp3 clip instead of the 48K wav I should have used. Even worse - that track is also in iTunes, Spotify, apple etc - and they were all delivered from the same source file - usually a 48K 32 bit .wav, but this time a 48K 320 mp3.

What you need to do is recreate the project using the mp3 and the mp3 converted to .wav and see if you can hear it. I suspect very much that you can't.

If I load an mp3 and a wav into premiere, in my edit system I have to look at the screen to tell which is which. There are differences, but when you export your video, it is re-rendered anyway - so it seems foolish to convert the mp3, then have that converted again when layered into the video.

Just try to ensure the next one is what you want.

Conversion is always going to be at best - a copy, bit for bit, or a data reduced version. The quest is to keep the output file as close as you can - so reduce conversions, but I suspect that the receivers of the file would not notice if the mp3 is a 320 version instead of a lower one. I suggest you simply try the process and then audition the results. I suspect you will have trouble determining which is which. Some audio content degrades badly with bit reduction and compression - others survive fine.

It's the old saying. You can't make a silk purse out of a pig's ear.

Regardless of the quality of the incoming MP3 files the best way to maintain their inherent quality, whatever that may be would be to convert them to full 64-bit float WAV files for your edit / sweetening exercise. For example if the original audio was recorded at 192 kHz frequencies of up to 96 kHz in the original audio would be recorded. Whether those frequencies are present in the existing MP3 files would need to be determined.

While working on a project, it’s not a bad idea to work with a higher audio bit depth. Because the noise floor drops, you essentially have more room before distortion occurs, this is known as headroom. Having this extra buffer space before distortion is a good failsafe while working and provides more flexibility.

You could finish up exporting your final sweetened files as full 64-bit float WAV files but one has to remember not all application can handle these files.

The noise floor, even in a 16-bit system, is incredibly low. Unless you need more than 96 dB of effective dynamic range, 16-bit is viable for the final bounce of a project. As most would know and understand the most common delivery format is 16-bit 44.1 kHz.

Chris Young

Kathy,

It seems to me that a few people are missing your question, and telling you things you already know. Please allow me to contribute my 2¢ worth toward answering your question. I hope I don't sound condescending or repetitious.

I completely agree with Don Palomaki: the degradation happened when the original audio was compressed to MP3. No [significant] further degradation occurs when it's uncompressed to PCM. But further degradation will occur when the finished PCM is compressed into the final MP3 format.

I know of a few programs that can edit (i.e. cut, paste, delete) MP3 files without transcoding (specifically, the final file will be the original file with some number of bytes added or removed). Even changing to a different bitrate involves transcoding.

I say the above to make a distinction between editing and audio alteration. I doubt that any program can actually alter the sonics of an MP3 file without first transcoding to PCM/WAV. Such alteration would involve operations on the PCM audio data, so the altered part would not contain any of the original bytes. By contrast, a truly edited MP3 file would contain the exact original bytes, minus any bytes that have been edited out, and plus any bytes that have been pasted in; nothing is transcoded.

I'm curious about Rick Reineke's comment that some editors do not overwrite the original file. Is this true even if the finished (edited, processed) file is saved using the original file name? If you "open" an MP3 file in a program such as Audition (or the predecessor Cool Edit) the "opening" process is really transcoding from the compressed file to uncompressed PCM in RAM, where the audio processing will take place. You can confirm that by starting with any [relatively lengthy] WAV file, Saving As MP3, then closing everything. Re-opening the MP3 file will take more time than re-opening the WAV file; that extra time is the transcoding time.

I don't know the content of the audio, but for example IMHO a mono MP3 at 96kbps is entirely adequate for use in a voice podcast, or even for location news gathering, etc. After all, it's not for playback in a THX theatre.

To expand on what Paul Johnson said: start with some given WAV file, convert to MP3, close both, reopen both, and subtract one from the other in your audio editor. That will let you hear the difference ... you will hear some artifacts that would normally be covered by the desired audio. It's an instructive experiment. But it won't have any effect on your workflow for this project.

Given the above, just open the file in your editor and it will end up in RAM as PCM. No need for separate conversion. Then do what you have to do, and save in whatever format the client needs. And save your final WAV in case client wants further work on this same project. The end.

Peace, out.

PS: What ever happened to that long room with the glass wall and low ceiling?

I'm curious about Rick Reineke's comment that some editors do not overwrite the original file. Is this true even if the finished (edited, processed) file is saved using the original file name? If you "open" an MP3 file in a program such as Audition
When a file is 're-saved' in audio 'editing' software like Sound Forge, it overwrites the data in the same file format, if it is a lossy format like MP3, that is an additional encode generation (like copying a cassette tape to a cassette tape). It can 'Save as' in a lossless non-compressed (data) PCM format though.
OTOH, most 'multi-track' DAWs creates high-res and floating point bit depth PCM proxy files for the timeline, leaving the original files untouched. A new master mix file is eventually rendered.

Sound Forge Pro is not a multi-track application like ProTools, Audition, Reaper, ect, but has a project file mode option, which is similar to multi-track applications so the original file(s) are never overwritten. There are other options in SF to maintain lossy file quality as well. All options have are pros and cons, so it depends on the project, file types and workflow preference.
btw, I am a Magix tech forum moderator and on the beta test team. MAGIX Software GmbH acquired the Sound Forge, Vegas and Acid products from Sony in 2016, which were originally developed and published by Sonic Foundry, USA in the 1990s.

It boils down to one has to know what one's software is doing. In general a SAVE will overwrite the original file (hopefully with warning) while a SAVE AS allows you to save under a different file name (keeping the original). It may also allow you to change the file format.

The 44.1 K sample rate and 96 kbps is not especially close to the maximum MP3 potential of 48 K sample rate and 320 kbps data rate. However, the success will depend on the material being encoded. Speaking voice generally does better than complex music at lower sample and bit rates.

MP3 uses psycho-acoustic coding techniques to drop audio detail that is considered less important to the file playback. A potential issue is that editing efforts such as equalization changes may want to work with some of the detail that was lost in the original MP3 compression. That could sour some of the sweetening prospects.

Any saves should be to a lossless intermediate file (e.g., uncompressed PCM /WAV). of suitable bit depth. The final action would be export to the desired distribution format.

Thanks, Rick. That is pretty much as I expected.

There is one program that I use for non-destructive *editing* of MP3 files, without audio manipulation. It's called Total Recorder Pro. I'll describe a few scenarios.

Open File1.mp3
Select and delete any part(s) to be removed. Repeat as needed. Copy and paste is also possible if you need to move some piece of the audio to a different place in the file.
"SAVE AS NewName.wav" will create a new file (as you'd expect). All data from the original edited file will be transcoded to WAV and saved as a new file. (Output file can be anything other than original File1 format and bitrate.)

Open File1.mp3
Edit as in first example.
"SAVE" will save back to the original file. It will NOT transcode any of the audio. If the edited file is shorter, of course the final length of File1 will be shorter than it was originally. But it's just a byte-for-byte process. Some bytes will be removed, but the remaining bytes are just the exact bytes from the original file.

Open File1.mp3
Edit as in above examples.
"SAVE AS NewName.mp3" will create a new file. The contents of that new file are copied byte-for-byte from File1.mp3 (minus any edits and trims) ... NO audio is transcoded (unless bitrate is changed).

The first example is what you'd expect. The last two examples are the feature I really like: no transcoding. The resulting file is as clean as the original, and the process is much faster than transcoding.

Total Recorder Pro also has a lot of recording functions including direct data capture from any source, before the audio goes through the sound card. It has timed record, as well. I've been using it for at least ten years. The program costs $35.95, which includes free lifetime upgrades to newer version numbers.

In your original post you mentioned eq and denoising. I doubt you can do those without decompressing.