Page 1 of 1

Newbie - and tutorials don't work!

Posted: Tue Nov 27, 2018 3:04 pm
by CarSinger
Hi all,

New lifetime licensee here, and my first task is to digitize some cassettes I have- These are mostly "accompaniment tracks" or performance tracks, i.e. the original song without the vocals for performing live. I scored a pretty decent Pioneer cassette deck at a church sale and have its line output into the Sound card line input , and can play the audio on the computer speakers, watch the spectrum/levels in Goldwave, etc.

But is there anything special about doing this? The first two links on the tutorial page are broken/misdirected. I read somewhere that you need to consider the type of noise reduction used (Dolby, etc), but If I'm using line out, don't you kind of get all you're going to get signal-wise? I read another set of recent posts here that said to aim for about -12db max, so do I need to make a pass through each rack before recording?

IRL I'm an mechancial engineer with an extensive background in vibration, so I understand about spectra, frequency, dB, etc, but haven't done much sophisticated audio recording. And I didn't want to jump right in asking how-to like this, but like I said, the tutorial links don't work. So if someone could kindly re-direct me or help with a quick overview, I'd appreciate it.

Thanks!

Re: Newbie - and tutorials don't work!

Posted: Tue Nov 27, 2018 4:29 pm
by DougDbug
The user manual or built-in help is probably a better place to start than 3rd-party tutorials. What's not working?
Go to Options -> Control Properties -> Record to select line-in.

Open a new file and click Record..
and have its line output into the Sound card line input
Right. You need a desktop/tower computer with a "regular soundcard" with line inputs. If you have a laptop with only mic-in and headphone-out you need an external USB interface with line-in.
I read another set of recent posts here that said to aim for about -12db max, so do I need to make a pass through each rack before recording?
The main thing is that you don't "try" to go over 0dB. 0dB is the "digital maximum" for analog-to-digital converters, digital-to-analog converters, regular WAV files, CDs, etc. If you try to go over 0dB you'll get clipping (distorted flat-topped waves).

Nothing bad happens when you get close to 0dB but you need to leave some headroom for unexpected peaks. -12dB is OK and it's a good recommendation for "live" recording. But, cassettes are more predictable so you can shoot-for between -6 and -3dB.

You can run Tools -> Amplitude Statistics to check the peaks after recording. If the peak is 0dB you can assume it's clipped and you may want to re-record.

After recording and after any other processing, you can run Volume -> Maximize Volume* to bring the peaks up (or down) to exactly 0dB. GoldWave itself can go over 0dB without clipping. So if you boost the bass (or something do else) and you end-up pushing the peaks over 0dB, you can run Maximize to bring the peaks down to 0dB before saving.





* In audio terminology adjusting the level for 0dB peaks is called "Normalization", but GoldWave uses the term "Maximize".

Re: Newbie - and tutorials don't work!

Posted: Tue Nov 27, 2018 5:34 pm
by CarSinger
Much thanks, DougDbug!

It's just that easy then?! Not bad! I probably should have checked the Help & manual more closely - but these tracks are real "keepers" and I didn't want to end up with something less than the best quality DAC I could get, and not even know it, out of sheer ignorance!

That's interesting that they set the DAC maximum to 0dB... I'm a little more used to the opposite side of the scale when doing noise & acoustic measurements. But as those aren't very musical, and the test equipment does all the DAC work, I don't have to be concerned with it.

Normalizing (maximizing) will be very helpful, thanks for that. Less fiddling with the volume knob!

Thanks again!
Best,

Tom

Re: Newbie - and tutorials don't work!

Posted: Tue Nov 27, 2018 7:45 pm
by DougDbug
That's interesting that they set the DAC maximum to 0dB... I'm a little more used to the opposite side of the scale when doing noise & acoustic measurements.
Right... The 0dB SPL reference (for acoustic loudness) is the approximately the quietest sound that can be heard and SPL readings are positive..

The digital reference (for integer formats) is 0dBFS (0dB full scale) which is the highest you can "count" with a given number of bits. With 16-bit signed integers you can go from −32,768 to +32,767 (decimal). If your digital peaks hit those values that's 0dB. Everything is scaled so a 0dB 24-bit file is not louder than a 0dB 8-bit file. In floating-point representation 0dB is 1.0.

There is no fixed calibration between digital and acoustic levels, except in movie theaters, because it depends on your volume control, your amplifier, speakers, and how far you are from the speakers, etc. And even in movie theaters it depends on where you are sitting. But, there is a direct correlation and if you reduce the digital level by -6dB the acoustic level will also be reduced by -6dB.

If you are old enough to remember analog tape, 0dB was also the approximate or recommended maximum but you could go over by a few (or several) dB and you'd get some distortion & compression before you'd hit the absolute limit. The result was "soft clipping" where the peaks were rounded-off instead of squared-off. And with analog tape you generally wanted to record "hot", going occasionally "into the red" to overcome tape noise. With digital there is no tape noise* so you can record much lower and you never want to go into the red because of that nasty hard-clipping.



* There is something called digital quantization noise but at 16-bits it's more than 90dB down so not a problem. At 8-bits you can hear the quantization noise, which "rides on top of the audio". It's most noticeable with quiet sounds, but unlike analog tape noise there is no quantization noise if there is no signal.