I am digitizing some vinyl albums I have, in this instance it is the Led Zeppelin Zoso album. For the other albums I have done I was able to set the Recording Slider to ensure the Maximum Peak of the input signal level was around -6.0 dB as suggested in the Audacity Tutorials. On Side 1 of the Zoso album, I left the settings from the previous album and didn’t have a peak that exceeded -6.0 dB, therefore Side 1 is good based on the -6.0 dB guidance.
Side 2 was a little different, during the Misty Mountain Hop the input signal exceeded 0 dB in one location even though I used the same settings as Side 1. It also appears that -6.0 dB was exceeded around the 0 dB but I am not sure how to find all the locations >-6.0 dB. I attached the waveform for your viewing pleasure.
I am running Audacity 2.3.0 on Windows 10.
My questions are:
If I leave the digitization the way it is, what limitations in post processing will I run into (e.g. normalization, etc.)? Will there be an impact on the overall quality?
If I go back and adjust the input signal level lower on Side 2 to ensure -6.0 dB is not exceeded then how will Side 1 sound compare to Side 2?
How can I find all locations where the input signal is greater than -6.0 dB?
I have been saving each digitized side in it’s own file and plan on post processing each side individually then separating out the individual “Tracks”. Is this method advised or should the recording be split then post processed?
but I am not sure how to find all the locations >-6.0 dB.
Nothing bad happens if you exceed -6dB, but your analog-to-digital converter will [u]clip[/u] if you “try” to go over 0dB.
The -6dB recommendation is just a guideline to give you headroom so you can avoid clipping but in this case it wasn’t enough. …Pros often record at -12 to -18 dB, but they are recording “live” so the peaks are not as predictable, and they are recording at 24-bits with very-good equipment so they have plenty of dynamic range.
The clipping doesn’t look that bad and if you can’t hear the distortion you may just live with it. Otherwise you can re-record at a lower level.
If I leave the digitization the way it is, what limitations in post processing will I run into (e.g. normalization, etc.)? Will there be an impact on the overall quality?
The damage is done. Normalizing to 0dB won’t do anything since the peaks already hit 0dB. If you normalize to less than 0dB you’ll “hide” the clipping from Audacity but you won’t change the wave shape.
If I go back and adjust the input signal level lower on Side 2 to ensure -6.0 dB is not exceeded then how will Side 1 sound compare to Side 2?
You can lower side 1 by -2dB in post-production to re-match the original recording.
How can I find all locations where the input signal is greater than -6.0 dB
Don’t worry about -6dB. But, you could temporarily Amplify by +6dB, and allow clipping. Then anywhere you “see red”, that was over -6dB before you amplified. Now it’s over 0dB. Then, Amplify (attenuate) by -6dB to bring the levels back down.
I have been saving each digitized side in it’s own file and plan on post processing each side individually then separating out the individual “Tracks”. Is this method advised or should the recording be split then post processed?
It’s up to you. I usually combine the sides and process all at once. If I’m making a CD, I leave it as one big WAV file and use a cue sheet to set the track markers. If I’m making MP3s, I’ll split the tracks after any noise reduction or EQ, etc. I usually do some “trimming” of the silence on the individual tracks before making the final MP3.
I don’t know what input you are using but even if you intend to save the final tracks at the standard sample rate of 44.1Khz I would suggest that, unless you don’t intend to do any post processing whatsoever, you try to record at twice this rate. Personally I break out each track then tidy up any surface noise if its intrusive by zooming in to see the click as an inconsistent wave and then simply delete it. Its easy to recognize clicks in most cases or you can just play a small area to make sure your not looking at a percussive sound which is part of the performance. After you have removed any such noise it becomes more realistic to check the peak levels and I use the amplify tool to see if amplification is desirable ( plus my ears of course). Vinyl recordings often have very varied levels for reasons which are not always musical and can relate to how tracks of certain lengths are squeezed into the available space by using different widths on the vinyl itself. There is also the issue of how the tracking speed slows down towards the centre of the disk, often leading to quieter tracks being used at the end of each side.
You might also want to write tag data to each track with various data such as artist, track no., album name etc. Its all extra work so how much effort you put in is for you to decide but my preference is to handle and save each track as a separate file.
I disagree.
High sample rates increase demands on the system when recording, increase the required disk space, increase the likelihood of intermodulation distortion due to EMI interference, decrease the sound quality in some compressed formats, and in consumer level sound cards will often increase the noise floor level.
The main “benefit” is that for sound cards that don’t limit the bandwidth, inaudible extreme high frequencies may be recorded (useful when recording bats, dolphins, or dog whistles). There are also marginal benefits when close-mic’ing sounds that have exceptionally strong high frequencies (such as cymbals), assuming that you are also using microphones that have strong high-frequency response (most microphones drop off their response very rapidly at around 16 to 20 kHz).
There is a definite benefit with certain types of digital signal processing, particularly distortion effects, though plug-ins that benefit from high sample rates will often up-sample to a higher rate internally, so no need to feed them high sample rates.
I don’t think it will do the original poster any favours by rehashing this argument here. I just report what works for my listening enjoyment. “increase the likelihood of intermodulation distortion due to EMI interference” probably means as little to the OP as it does to me! Although I must admit that I did not even consider the possibility that vinyl output was being captured via microphone, that would be a very crude method of A to D conversion.
