Converting Cassette Album to Digital

[Dick61]

I'm a Newbie hobbyist. Back in 1988 my buddies and I recorded a 12 song cassette of songs we had played together for years. I'm trying to digitize that cassette, as I may have the only surviving copy of it. I successfully used MyPin to convert the songs to mp3 but need suggestions on how to clean up those songs. I'd like recommendations on the best plugins to use in my Mixcraft Pro 9 DAW to reduce noise on the tape -- mainly hissing, some clipping. I've been trying with Audacity and that works pretty well except it reduces gain dramatically. Am demo-ing a noise reduction plugin from Wave but it doesn't seem to do much.

Any and all recommendations will be appreciated.

Thanks

Dick61

 

[Mickster]

Hi Dick61....and welcome.

You're on the right track (no pun intended) but my suggestion would be to automate the "cleanup" effects / functions needed for every part of the tape / track / wav. Applying an effect like noise reduction on every part of the track / wav...and then another effect to the track etc...etc...compiles too much and often gives the result you are getting.....low gain and a dull finish. Automating each effect and it's strength by applying it only to where it's most needed in the wav and how much it's needed can significantly maintain the original sound.

You would need a DAW that can do that and there are MANY. And yes....it will take some time as you carefully go through each track and adjust the automation where needed. There are some cleanup effects that sort of emulate automation...and indeed....some do carefully "apply where needed" as best they can......but there's no substitution for the human brain and a good DAW.

2 cents worth of...been there....done that....and it worked........

Mick

 

[Dick61]

Hi Dick61....and welcome.

You're on the right track (no pun intended) but my suggestion would be to automate the "cleanup" effects / functions needed for every part of the tape / track / wav. Applying an effect like noise reduction on every part of the track / wav...and then another effect to the track etc...etc...compiles too much and often gives the result you are getting.....low gain and a dull finish. Automating each effect and it's strength by applying it only to where it's most needed in the wav and how much it's needed can significantly maintain the original sound.

You would need a DAW that can do that and there are MANY. And yes....it will take some time as you carefully go through each track and adjust the automation where needed. There are some cleanup effects that sort of emulate automation...and indeed....some do carefully "apply where needed" as best they can......but there's no substitution for the human brain and a good DAW.

2 cents worth of...been there....done that....and it worked........

Mick

Thanks, Mick. I was likely less than clear -- the "tracks" i referr to are 12 complete songs on the 1988 cassette. We did the recording long ago, probably on some fairly simply machine. I suspect we recorded everything "live" and did not do separate tracks for vocals, guitar, keys, etc., but don't actually recall. If we did do tracking, mixing, etc. those elements are long, long gone. So I am trying to digitize 12 complete songs, and do not have "tracks" (or parts elements) of each song!

 

[Mickster]

Dick61......yup....that's what I'm referring to. By "track" I mean each separate song...not the individual original tracks you recorded long ago. Each song is now a wav file (track). You can apply the cleanup effect(s) to each song (wav file) as needed......a small spot...long area....more cleanup here...less there....any type of cleanup effect you want.

Your original mp3's are (have been) converted to wav files. Sorry if I'm being confusing but the method I'm outlining here is common.

Ask away and we can help.

2 cents worth of......yup.....don't give up.

Mick

 

[TalismanRich]

I'm confused with the statement that it reduces gain. Once recorded you can put the gain where ever you want. Find the loudest peak, set that at about -1 to -0.5dB and you should be good to go. Don't try to make it sound as loud as a commercial CD, that's probably strongly compressed vs your cassette.

As for getting rid of hiss, you have to be somewhat careful, as it can have unwanted side effects. I've converted a few cassettes and albums to digital, and my solution was to cut the song to just before the very beginning of the first wave, and do about 1/2 second fade up. At the end, I find the fade out and do a longer fade out at the end (1-2 seconds). That way I have all the original data but without the "hiss" between tracks. If it was good enough on the cassette, then it should be fine on the CD or MP3s. (isn't that part of the analog warmth? )

If you do use noise reduction, like ReaFIR the one included with Reaper, you can set it for mild noise reduction and maybe avoid a lot of the artifacts.

The cassettes were the easiest. I have done some albums that had clicks and pops from the surface defects. There was a click/pop filter that I found that was ok, if you didn't apply it too drastically. For very bad clicks, I often had to edit it manually to remove the click, by redrawing the waveform.

 

[Mickster]

Yes....as TalismanRich states...you can manually edit many of the issues the way he outlines.

 

[Dick61]

Dick61......yup....that's what I'm referring to. By "track" I mean each separate song...not the individual original tracks you recorded long ago. Each song is now a wav file (track). You can apply the cleanup effect(s) to each song (wav file) as needed......a small spot...long area....more cleanup here...less there....any type of cleanup effect you want.

Your original mp3's are (have been) converted to wav files. Sorry if I'm being confusing but the method I'm outlining here is common.

Ask away and we can help.

2 cents worth of......yup.....don't give up.

Mick

We are now tracking…am cleaning up the 12 songs one by one, via Audacity, then exporting them twice, once as WAV and once as MP3. Any recommendations on noise reduction other than with Audacity?

 

[Mickster]

Pardon my question please. I mean no insult....but are you familiar with manual editing of the wav....or the term "automation" as it applies to a DAW? If not....no problem at all....but it helps us to give you good advice.

I assume what you're doing is putting the song in Audacity......then choosing a cleanup effect like noise reduction and letting it run all the way through to the end of the song.

We're outlining a way to apply cleanup to only the parts / sections of your song where needed along with a few "tricks" of editing where you make "cuts" in your track and applying "automation" where needed.

Sorry if my 2 cents is confusing.

Mick

 

[bouldersoundguy]

MyPin

Ugh. That device is going to start the whole process with a collection of disadvantages. An mp3 file is degraded right from the start. Poor tracking can substantially decrease signal to noise ratio. The crappy transport is probably adding speed deviations (wow and flutter). I bet a lot of what you're doing is just trying to compensate for the poor quality transfer. I bet the most effective way to fix some (or most, or all) of the issues you're having with the audio is to not cause them in the first place with a cheap transfer system.

 

[rob aylestone]

If you are producing .wavs, it was a mistake to digitise to .mp3, as .wav files are only as good as the mp3 - the small quality loss going to mp3 won't be made better.

It does seem to me that Mickster is picking up on what I did? You've not really identified the problem, but are trying to cure it, kind of blind. With cassettes the real issue is hiss, and being ancient enough to have been doing cassette distribution, some problems were managed quite well in the mix. Cassettes hiss - low tape speed, narrow tracks, inconsistent bias requirements, often switchable replay EQ -and of course, Dolby. Back in the cassette days, the hi-fi enthusiasts often recorded with Dolby B on, which boosted HF, and then turned Dolby off for replay which made it very bright (and hissy) and then used their 32 band EQ devices to tame the hiss, but leave important content. So a song with lots of drum kit cymbal work and hi-hat stickwork, and 12 string guitars and high femalke vocals would have the EQ settings to reveal the HF content, then slice it at the lowest point they could get away with. On a more rocky track with drums, guitars and a make singer, the cutoff could be lower, making a less hissy recording. Mix these together on an album and you'd have more or less hiss, depending on the track content.

Distortion, you are probably going to have to live with. Clicks can be nibbled out, and on some software even redrawn with a mouse to soften nasty spikes - but only some DAWs have a hunt, seek and repair distortion preset, and while it might work on an acoustic guitar, piano and voice track - add deliberate guitar distortion and it doesn't know how to tell wanted distortion from unwanted distortion.

I found a cassette of my first band from 76, and it is dire. Thin recording, hiss, and distortion. Trouble is it is a very 'open' recording you can hear everything in a rather clinical mix. I have had no luck improving it at all. Generic reverb fails because it blurs things that should not be 'reverby', and the hiss removal - by manual and presets makes it really dull.

Back in the late 70s this is a local band recording, and in about 1998, I had a short lived digital recording system A Soundscape, with two Mixtreme cards. It tried to rival Cedar that the BBC were using, but it never caught on. The cassette was digitised on a Tascam 112 machine that I still have that actually works, and is virtually noise free - but if you have something that shows you frequency content - there is very little above 7K - because the hiss was very evident. I'd have liked to have tried my old band's cassette on it, but I lost it years back in a clear out. Motto - never throw any tape away! Despite having virtually no HF content above 10K at all, it's a pretty good recording. The band are still working, but I don't see this album anywhere - I wonder if they've actually lost the tracks, or they're too noisy? I might actually contact them and give the the recording - for a 70s recording it still holds up I think.
r

 

[Dick61]

If you are producing .wavs, it was a mistake to digitise to .mp3, as .wav files are only as good as the mp3 - the small quality loss going to mp3 won't be made better.

It does seem to me that Mickster is picking up on what I did? You've not really identified the problem, but are trying to cure it, kind of blind. With cassettes the real issue is hiss, and being ancient enough to have been doing cassette distribution, some problems were managed quite well in the mix. Cassettes hiss - low tape speed, narrow tracks, inconsistent bias requirements, often switchable replay EQ -and of course, Dolby. Back in the cassette days, the hi-fi enthusiasts often recorded with Dolby B on, which boosted HF, and then turned Dolby off for replay which made it very bright (and hissy) and then used their 32 band EQ devices to tame the hiss, but leave important content. So a song with lots of drum kit cymbal work and hi-hat stickwork, and 12 string guitars and high femalke vocals would have the EQ settings to reveal the HF content, then slice it at the lowest point they could get away with. On a more rocky track with drums, guitars and a make singer, the cutoff could be lower, making a less hissy recording. Mix these together on an album and you'd have more or less hiss, depending on the track content.

Distortion, you are probably going to have to live with. Clicks can be nibbled out, and on some software even redrawn with a mouse to soften nasty spikes - but only some DAWs have a hunt, seek and repair distortion preset, and while it might work on an acoustic guitar, piano and voice track - add deliberate guitar distortion and it doesn't know how to tell wanted distortion from unwanted distortion.

I found a cassette of my first band from 76, and it is dire. Thin recording, hiss, and distortion. Trouble is it is a very 'open' recording you can hear everything in a rather clinical mix. I have had no luck improving it at all. Generic reverb fails because it blurs things that should not be 'reverby', and the hiss removal - by manual and presets makes it really dull.

Back in the late 70s this is a local band recording, and in about 1998, I had a short lived digital recording system A Soundscape, with two Mixtreme cards. It tried to rival Cedar that the BBC were using, but it never caught on. The cassette was digitised on a Tascam 112 machine that I still have that actually works, and is virtually noise free - but if you have something that shows you frequency content - there is very little above 7K - because the hiss was very evident. I'd have liked to have tried my old band's cassette on it, but I lost it years back in a clear out. Motto - never throw any tape away! Despite having virtually no HF content above 10K at all, it's a pretty good recording. The band are still working, but I don't see this album anywhere - I wonder if they've actually lost the tracks, or they're too noisy? I might actually contact them and give the the recording - for a 70s recording it still holds up I think.
r

 

[Dick61]

Thanks for all the suggestion -- all of them have been helpful. Am making progress -- decided to drag all 12 songs on the cassette into separate track of a "project" in Audacity. Was cruising along with noise reduction when at one "track" I received a message to the effect that "sample rates have to be identical" in order to identify the "noisy section" that I want to fix. I have zero idea (a) what the sample rate is in Audacity and particularly (b) what the sample rate is of tracks recorded in 1988. Any suggestions??

Thanks

 

[bouldersoundguy]

Cassettes, like all analog, don't have sample rate. The digital files have whatever sample rate was used to capture the analog audio. I'm not sure why it would be different for different files.

 

[Dick61]

Cassettes, like all analog, don't have sample rate. The digital files have whatever sample rate was used to capture the analog audio. I'm not sure why it would be different for different files.

I'm not at all technical, which is a hindrance -- but in the "box" at the left of each track, some show 44100 hz, 32 bit float and others show 48000 hz, 32-bit float. Have no idea how this happens. And another request -- when I get all tracks transferred into my Audacity project, should I "export" the entire effort or try to export one 'song (track" at a time?

 

[bouldersoundguy]

I would export one at a time, unless there's a batch export option that lets you export all of them to separate files using the same settings. Normally you'll want your finished files to be 44.1 kHz and 16 bit wave. You can use those to generate mp3 files later for sharing, but first create wave files for archiving.

And seriously, that cassette machine is junk. It may seem convenient, but it starts the whole process with a substantial reduction in quality that can never be recovered.

 

[Dick61]

I would export one at a time, unless there's a batch export option that lets you export all of them to separate files using the same settings. Normally you'll want your finished files to be 44.1 kHz and 16 bit wave. You can use those to generate mp3 files later for sharing, but first create wave files for archiving.

And seriously, that cassette machine is junk. It may seem convenient, but it starts the whole process with a substantial reduction in quality that can never be recovered.

I appreciate that the MYPin is not exactly audiophile quality, but all I'm trying to accomplish is creating a digitized version of what two pals and I recorded way back in 1988...it's a souvenir, not headed for Spotify! At this point, I think I'll just export the WAV file for each track, then drag them into my DAW (Mixcraft 9), do whatever additional cleaning up I can do, and send the results off to my pal. I DO enjoy what I am learning thru this process, and appreciate all the helpful suggestions.

 

[Dick61]

I appreciate that the MYPin is not exactly audiophile quality, but all I'm trying to accomplish is creating a digitized version of what two pals and I recorded way back in 1988...it's a souvenir, not headed for Spotify! At this point, I think I'll just export the WAV file for each track, then drag them into my DAW (Mixcraft 9), do whatever additional cleaning up I can do, and send the results off to my pal. I DO enjoy what I am learning thru this process, and appreciate all the helpful suggestions.

Wish I understood what it means when I get a message that "all noise profile data must have the same sample rate". I am soloing a track which is identified as "Stereo, 44100 hz, 32 bit float". I'm sampling nose on an song on an analog cassette deck. Can any one explain the message in terms that a non-techie like me might understand? Just trying to drag myself up a learning curve!

 

[bouldersoundguy]

You've created a noise profile from a file that has a sample rate of 48 kHz. Evidently, the software you're using can't use that profile on a file with a different sample rate. Either you digitized the one song at a different sample rate or you've opened it up in a window with a different sample rate than the others. It might be possible to go back a step and make it 48 kHz to match, but that depends on where the mismatch happened. How does that MyPin thing work? Does it capture to its own application or does it rely on Audacity?

When converting from analog to digital, the converter takes a sample of the voltage of the incoming waveform at regular intervals. That interval is the sample rate. The rate has to be high enough to accurately represent the highest frequencies of the incoming audio. 44.1 kHz is the standard for CD audio, 48 kHz is the standard for DAT and most video formats. Higher rates are fairly common in the recording world, but those two are typical for delivery formats.

 

[Dick61]

You've created a noise profile from a file that has a sample rate of 48 kHz. Evidently, the software you're using can't use that profile on a file with a different sample rate. Either you digitized the one song at a different sample rate or you've opened it up in a window with a different sample rate than the others. It might be possible to go back a step and make it 48 kHz to match, but that depends on where the mismatch happened. How does that MyPin thing work? Does it capture to its own application or does it rely on Audacity?

When converting from analog to digital, the converter takes a sample of the voltage of the incoming waveform at regular intervals. That interval is the sample rate. The rate has to be high enough to accurately represent the highest frequencies of the incoming audio. 44.1 kHz is the standard for CD audio, 48 kHz is the standard for DAT and most video formats. Higher rates are fairly common in the recording world, but those two are typical for delivery formats.

The manual for the MYPIN is as skimpy at the device itself! The converter type is provided - mp3. Other than that, it just shows a Bit Rate of 128kbps. I'll contact their support to seek more information. Thanks for the response.

 

[Dick61]

You've created a noise profile from a file that has a sample rate of 48 kHz. Evidently, the software you're using can't use that profile on a file with a different sample rate. Either you digitized the one song at a different sample rate or you've opened it up in a window with a different sample rate than the others. It might be possible to go back a step and make it 48 kHz to match, but that depends on where the mismatch happened. How does that MyPin thing work? Does it capture to its own application or does it rely on Audacity?

When converting from analog to digital, the converter takes a sample of the voltage of the incoming waveform at regular intervals. That interval is the sample rate. The rate has to be high enough to accurately represent the highest frequencies of the incoming audio. 44.1 kHz is the standard for CD audio, 48 kHz is the standard for DAT and most video formats. Higher rates are fairly common in the recording world, but those two are typical for delivery formats.

The MYPIN "converts a song on the cassette into mp3 format which goes to a USB stick inserted in the MYPIN when you start converting. It has worked pretty well, but I was just trying to convert a last song from the cassette and only silence was produced on the USB stick! Oh well....