Audacity Forum

Hi I’m Gin,
I’m a complete beginner, I’ve watched a few audacity tutorials to help me improve some 40 year old amateur audio singing recordings which were originally on cassette. Specifically I took the knowledge from this YouTube video: https://www.youtube.com/watch?v=7_vvfMOqHiw

So this is what I’ve been doing. I used, in order:

Noise Reduction, (If you have a particularly noisy piece of audio you can change the value in the noise reduction DB box to between 12 to 18) then hit okay)
then Compressor, then Normalise, then Equalisation, Select curve, Bass boost,
then Equalisation, Select curve, Treble boost,
then Bass and treble, (Adjust the sliders as desired, you can preview if needed)
then Limiter, limit to -4 dB, and finally, Amplify

I find that after applying all of these settings and then playing my track through my widescreen TV speakers, I have a kind of crackly distorting buzz at the high end of the audio. (I have sent a short WAV clip with this post for your perusal) Now when I just play the audio through my computer, well then the buzz is not there and the same with headphones, it’s only through my widescreen TVs speakers that the distortion happens. I’m afraid my audio will not bear being played through big speakers which sucks! it’s a classic case of a little bit of knowledge is a dangerous thing. Please can someone help me to find a solution to this problem, maybe giving me an idea which of those settings it was which caused this problem in the first place. Ongoing, I am digitizing more old cassette tapes and attempting to improve the sound quality. I don’t want to make things worse during restoration so I’d like to figure out a non-distorting way of improving my recordings!

Kind Regards Gin

Could it be to do with something that was due to the original recording?

ginlane wrote: ↑

Wed Jan 15, 2020 8:04 pm

... I have a kind of crackly distorting buzz at the high end of the audio ...

There are digital-artefacts way up at 14kHz, (that's on the limit of most people's hearing) ...

https ://manual.audacityteam.org/man/spectrogram_view.html


If you remove everything above 13kHz with the equalizer that will get rid of them.

Hi Trebor,
Thank you for getting back to me and thank you for your advice.
Please could you give me a little help on how I go about getting rid of all these artefacts that you say are above 13 kHz using the equaliser tool. To be honest I have absolutely no idea how to do this
kind regards Gin

ginlane wrote: ↑

Thu Jan 16, 2020 6:25 pm

Please could you give me a little help on how I go about getting rid of all these artefacts that you say are above 13 kHz using the equaliser tool.

https://manual.audacityteam.org/man/filter_curve.html

It's possible the >13kHz artefacts may not be the problem you're hearing.
These bat-like noises are on the upper-limit of hearing for most people.
.

ginlane wrote: ↑

Wed Jan 15, 2020 8:04 pm

... playing my track through my widescreen TV speakers, I have a kind of crackly distorting buzz ...

Buzz noises can be due to clipping which can be avoided if you leave 1dB-2dB headroom, (rather than 0dB).

Sorry about this long post... I wasn't planning on writing a book today...

I'd like to hear the original. I think most of the "issues" are acoustic (like the room & mic position). Most of that you can't fix. Did you listen to the original on your TV? Have you made it better or worse?

This kind of "acoustic" music is difficult to record because of the background noise against the quiet parts. And of course it was on cassette (tape noise) and probably with a "cheap" cassette recorder with a built-in mic picking-up motor noise from the cassette machine, and probably with a poor mic position.

You can probably make some improvement but there is only so-much that you can do. And with certain effects/processing, "The cure can be worse than the disease" So you have to listen carefully and decide if you're making an improvement. That's especially true with noise reduction. Noise reduction works best when you have a very-low level, constant, background noise.... It works best when you don't really need it.

And, it's often helpful to give your ears (and brain) a break and re-listen the next day. Sometimes you add a little treble boost (or a little reverb) and it sounds good so you add a little more, and then you get used to it and you add a little more... and then it doesn't sound so great the next day.

Headphones are usually more "revealing" than speakers. Professional mixing & mastering engineers use studio monitors for most of their work. They may use headphones when listening for "little details"

Then, they will usually check the result on a variety of other "consumer" devices". That can be even MORE important for amateurs because most of us don't have "perfect" studio monitors in a "perfect" room and we don't have the same skills & experience. You can end-up with a recording that sounds great on your particular headphones but doesn't sound so great on your TV, home stereo, car, etc.

They usually also have a known-good reference recording in the same genre to compare and to "keep their ears calibrated".

Small speakers in a flat screen TV or laptop can't reproduce bass. If there is lots of deep bass you can end-up distorting the amplifier or "rattling" the speakers with bass that you can't hear.

Specifically I took the knowledge from this YouTube video: https://www.youtube.com/watch?v=7_vvfMOqHiw

To me his "before" recording sounds better! I'm hearing way too much high-frequency boost and it just sounds "over processed".

So this is what I’ve been doing. I used, in order:

Noise Reduction, (If you have a particularly noisy piece of audio you can change the value in the noise reduction DB box to between 12 to 18) then hit okay)
then Compressor, then Normalise, then Equalisation, Select curve, Bass boost,
then Equalisation, Select curve, Treble boost,
then Bass and treble, (Adjust the sliders as desired, you can preview if needed)
then Limiter, limit to -4 dB, and finally, Amplify

I don't see anything "wrong" with that except all of the equalization (including bass & treble adjustment) could be done all at once.

There are certain steps/guidelines you can follow, depending what you're doing, but you can't just follow a recipe. Diagnosis first, then treatment! And, usually there are compromises.

Noise Reduction

Like I said, the cure can be worse than the disease. You can get artifacts (side effects). These are usually little noises, not as loud as the original noise but correlated with the sound and with a "character" that can be more annoying than the original noise.

then Compressor

Dynamic compression (not to be confused with file compression like MP3) evens-out the volume by making loud parts quieter and/or quiet parts louder. In general, it's used to bring-up the overall "loudness" by pushing-down the loud parts, then adding "make-up gain" to bring everything up. And/or it can be used as an "effect" by adjusting the attack & release times (and other settings).

Since you are bringing-up the overall loudness, compression brings-up the background noise (or noise reduction artifacts).

There are several settings (and all kinds of 3rd-party compressor plug-ins) so it can be tricky to configure.

Compression is one of the most common effects. But I wouldn't say "necessary". Virtually all commercial recordings are compressed (and limited). The exception would be some classical & jazz that have very-quiet & very-loud parts where they want to preserve the original dynamics. (IMO - Most modern releases are over compressed so they can be "louder than everybody else" and that constant-volume makes them boring and it makes me turn-down the volume... But that's what sells...)

If you over-do compress (or use the wrong settings) it can sound like distortion.

then Normalise

Normalization is simply a volume adjustment that targets a specified peak level. Typically, that's the last step (because the other processes/effects can change the volume). Or, you can Normalize before limiting or compression so you have a known starting-point.

Audacity's Amplify effect can also be used to normalize and in-fact, it will default to whatever gain (or attenuation) is needed for "maximized" 0dB peaks. (The Normalize effect has a couple more options.)

then Equalisation, Select curve, Bass boost,

OK if you need to boost the bass. Of course, bass boost will also boost any AC hum or other low-frequency noise.

then Equalisation, Select curve, Treble boost,

OK if you need to boost the treble, but that can be done at the same time as the bass. Boosting the treble will boost any tape hiss or noise reduction artifacts.

then Bass and treble, (Adjust the sliders as desired, you can preview if needed)

You've already done that with the equalizer. You want more?

then Limiter, limit to -4 dB

Limiting is a kind of fast-compression, again generally used to make the recording "louder". It "pushes down" the peaks, then make-up gain (or Normalization or Amplify) can be applied to bring-up the overall loudness.

The amount of limiting depends on what you're starting with (diagnosis), and of course, what you're trying to accomplish. There are not as many settings/options as compression so it's easier to use and Audacity's limiter is very good. But if you over-do it, you can get distortion and again if you are bring-up loudness you are bringing-up the noise.

BTW - clipping ("overload" distortion) is also a (bad) kind of dynamic compression.

and finally, Amplify

If you want to Normalize (AKA "Maximize"), that's a good idea as the last step If you use the default it can bring down the volume to prevent clipping (if necessary). Or, Normalize can do the same thing. Or you can run Amplify just to check the peak levels, and then cancel the effect if you don't want to change anything.

...Audacity uses floating-point so it can go over 0dB internally/temporarily without clipping so you may not notice the distortion. But, regular WAV files, CDs, your digital-to-analog converter (playback), and analog-to-digital converter (recording) are all hard-limited to 0dB so your "final product" should never "try" to go over 0dB. (So a last-step volume check/adjustment is a good idea.)

Hi DVDdoug,
Thank you for that big long post, it must have taken some time I have gone through t as best I can. Actually, I’ve been rather coming to the conclusion, that may be what I’ve got, is the best that I can expect. Yes I’ve listened to the original on the TV and there is a slight high-end crackle, …. Not as much as after I played around with Audacity settings. But I’m coming to the conclusion that the high end crackle is to do with the TV speakers. See when I turn it down lower, the crackle goes away. It’s only when the speaker is quite loud, that the crackle comes. (It’s like a slight high-end distortion) I don’t think it’s to do with tape hum or stuff like that. I could be wrong but it’s my belief that the high end crackle is due to it being recorded in a large hall on an old mediocre cassette recorder, and recording soprano singer’s. I’ve no idea how far away the cassette player was placed from the singers, (at a guess, maybe a meter or 2) or any information like that.
I’ve gone through your settings tutorial and I appreciate your help, I think that learning about all of this is fascinating but it takes up so much time to figure out. Also, I have many other cassette recordings I need to clean up, all of these would have been recorded at different times in different buildings and possibly with different equipment, with different atmospheric conditions. Unfortunately I don’t have the time to deal with each recording individually, and learn what I must do individually to each recording to clean them up. I sort of need a one size fits all approach to this.
Kind regards Gin

ginlane wrote: ↑

Sat Jan 18, 2020 2:49 pm

... I sort of need a one size fits all approach to this ...

Audacity is capable of batch-processing, see ... https://manual.audacityteam.org/man/macros.html

BTW the sample you submitted covers the Full Scale : +/-1.0 ...

If it at some point it crossed the +/-1.0 lines there will be a horrible crackling noise.

+/-0.9 is plenty.

Unfortunately I don’t have the time to deal with each recording individually, and learn what I must do individually to each recording to clean them up. I sort of need a one size fits all approach to this.

The ONLY thing I'd recommend without careful listening is Amplify or Normalize. That's simply a volume/level adjustment that doesn't affect the character/quality of the sound.

Something to consider - Digitize the tapes, normalize, and "archive" as WAV or FLAC files (lossless formats). You can listen to those, or make MP3s or CDs, or whatever. Then if you feel like spending more time later you can, or maybe there are certain special or favorite recordings that you want to spend some time & effort on, and others that you just want to digitize and preserve.

...Speaking of preservation, make a backup (at least one) so you don't have to digitize again, or in case the tape deteriorates or your tape player dies, etc. You can still buy cassette players but they are no longer manufacturing VHS players/recorders and someday that will happen to cassette. Digital is "forever" (as long as you have backups) and you can always convert to another lossless format if WAV or FLAC is no longer supported 100 years from now.

You don't HAVE TO do any editing/processing and it's better to do nothing than to do things blindly that might make it worse!

Audacity is a (multi-purpose) TOOL. Imagine that you are doing home improvement. You can't just walk into a house and start sawing, drilling, hammering and painting, without first deciding what needs to be done and what you want to do. And, it's lot easier to break stuff with a hammer than to build stuff with a hammer, If you have a bunch of older houses you can make a general plan of what needs to be looked-at and what usually needs to be done, so you can have a general outline or a checklist but you can't do exactly the same things in every house.

My rule-of thumb is a minimum of 4X real-time. i.e. If you have a 1 hour recording, you're going to spend at least 4 hours on it. That's an one hour for recording, another hour for an initial listen-through (diagnosis), another hour for editing/processing/effects, and one final listen-through. It can take a LOT longer.

Or, SOMETIMES you don't need to listen-through the whole thing after recording, or you can listen carefully while recording, and if you are applying the same affects/processing to the whole file that step can go faster too. Or, you can go-off and do something else during recording ("calendar time" but not "billable time"). But at the end, you should listen to the whole thing. So, you're probably going to spend more than 2 hours on a 1-hour recording if you want to make any improvements.

You need to LISTEN to the noise reduction to make sure you are making an improvement in overall quality. Since the noise is bad, Noise Reduction can make it sound worse so you may not want to use it, or you may need to tweak the settings.

You also need to LISTEN to EQ. Except it's probably "safe" to filter-out the bass below 40Hz or so. Maybe below 100Hz is OK if there is no "real bass" on the recording. Cutting below 40Hz will have very-little (if any) effect on the sound and it might help if your TV speakers are "rattling". Just for reference, the lowest note on a regular bass guitar is about 41Hz.

I'd leave out the compression & limiting unless you need or want it. It's probably going to do more harm than good. Analog tape has a kind of compression/limiting "built in" if you go "into the red" while recording, but of course you don't know anything about the original recording levels.