Hi-
My vocal tracks in SONAR are made up of several wavs. As much as I try to keep all settings the same whenever I record, there are subtle differences in volume and ambient sounds, etc. And no, I can't sing it all the way through in one take<g>! I'd like to be able to use Har-Bal to smooth out the wavs that make up the tracks. I created a filter for the best-sounding wav and used that as a reference for another. I'm not sure where to go from there. I tried Matching Loudness, but didn't really get the results I expected. I can use regular Normalizing, but I'd like to try this technique too. I should be able to do this, much the way you'd match loudness across the songs on a CD, no?
I'm making some progress with Har-Bal (thanks again for that filter and mp3, Earle!), but I'd appreciate any help with the above.
Thanks much in advance!
-Susan
Smoothing out vocals?
Hi Susan,
I have a question for you. Is it the overall loudness or the tonality that is different between takes? I suspect it is as much the latter as it is the former so simply using match loudness may not work very well at all.
Also, since the match loudness feature works off the average power it is very unlikely to work well on a single track containing a portion of a vocal part. That's because the average power level is going depend a lot on how much silence is in each part (proportionate to singing) and this is likely to vary a lot between parts. You would be better off trying to match the peak power figure of merit but you'll have to do this manually. Do so by taking the difference between the peak power figure of merits of the Reference and the track your working on and use this as the gain setting.
Hope this helps,
Paavo.
I have a question for you. Is it the overall loudness or the tonality that is different between takes? I suspect it is as much the latter as it is the former so simply using match loudness may not work very well at all.
Also, since the match loudness feature works off the average power it is very unlikely to work well on a single track containing a portion of a vocal part. That's because the average power level is going depend a lot on how much silence is in each part (proportionate to singing) and this is likely to vary a lot between parts. You would be better off trying to match the peak power figure of merit but you'll have to do this manually. Do so by taking the difference between the peak power figure of merits of the Reference and the track your working on and use this as the gain setting.
Hope this helps,
Paavo.
Hi Paavo-
I knew Match Loudness wouldn't be the silver bullet for the track -- I was mainly just wondering if it was worth exploring as an alternative to normalizing. Thanks for explaining why it wouldn't be the best choice for a single track. I'd more or less come to that conclusion after experimenting a bit more, but it's always good to know the "why", and your explanation makes perfect sense!
I'll try your suggestion.
Thanks again-
Susan
I knew Match Loudness wouldn't be the silver bullet for the track -- I was mainly just wondering if it was worth exploring as an alternative to normalizing. Thanks for explaining why it wouldn't be the best choice for a single track. I'd more or less come to that conclusion after experimenting a bit more, but it's always good to know the "why", and your explanation makes perfect sense!
I'll try your suggestion.
Thanks again-
Susan
Use the "match loudness" feature in Har-Bal as a reference for your limiter.Susan G wrote:Hi Paavo-
I knew Match Loudness wouldn't be the silver bullet for the track -- I was mainly just wondering if it was worth exploring as an alternative to normalizing. Thanks for explaining why it wouldn't be the best choice for a single track. I'd more or less come to that conclusion after experimenting a bit more, but it's always good to know the "why", and your explanation makes perfect sense!
I'll try your suggestion.
Thanks again-
Susan
When you use this feature just write down the gain number.
Pull the fader back to 0.0, record the newly EQ'd file and open the limiter.
Set the outceiling at -0.1 and the threshold at the number indicated by har-bal. If there was an increase in Har-Bal of 4.3,
Set the threshold for -4.3.
It works every time and the volume level is consistent throughout the entire album.
This is an amazing feature in Har-Bal. I know of no other tool that does this.
Make sure you use a reference CD that matches this genre of music in terms of loudness.
Earle
Here's what I do (sometimes):
Depending on how many 'takes' you have, place each one on a seperate track.
I'm assumming that you may have recorded on different days or with different mic distance. This could give you the difference in 'tone' and volume.
Just place the different takes on their own track. Now, decide on which take will be your 'target' tone. Next place an EQ on the subsequent tracks and attempt to match.
Also, you'll be able to adjust volume levels of each track to even things out.
Once you are satisfied you can then bus these tracks over to a new track and bounce to make a 'composite'. You might even want to add a compressor plug-in - although I wouldn't 'record' the compression so that you can adjust during your final mixing. The same with other effects (i.e., reverb, delay), don't use them on the individual 'takes', place them on your 'composite'.
What you are doing now is too difficult to achieve on a single track. Use as much automation as you can (i.e., levels, auto-tune, etc.) on the individual 'takes'.
This is tedious but will give you the best end-result.
And you can always re-do and tweak, tweak, tweak without messing up the original takes.
How I do this is that I rarely punch-in. I get complete takes - maybe 3-4. Send the singer away and then begin to build a composite track. Occassionally I do need to bring the vocalist back in. This is why I carefully and with detail note mic distance, EQ, compression, etcetra from anything that contributes by way of 'analog'.
Hope this makes sense...
Depending on how many 'takes' you have, place each one on a seperate track.
I'm assumming that you may have recorded on different days or with different mic distance. This could give you the difference in 'tone' and volume.
Just place the different takes on their own track. Now, decide on which take will be your 'target' tone. Next place an EQ on the subsequent tracks and attempt to match.
Also, you'll be able to adjust volume levels of each track to even things out.
Once you are satisfied you can then bus these tracks over to a new track and bounce to make a 'composite'. You might even want to add a compressor plug-in - although I wouldn't 'record' the compression so that you can adjust during your final mixing. The same with other effects (i.e., reverb, delay), don't use them on the individual 'takes', place them on your 'composite'.
What you are doing now is too difficult to achieve on a single track. Use as much automation as you can (i.e., levels, auto-tune, etc.) on the individual 'takes'.
This is tedious but will give you the best end-result.
And you can always re-do and tweak, tweak, tweak without messing up the original takes.
How I do this is that I rarely punch-in. I get complete takes - maybe 3-4. Send the singer away and then begin to build a composite track. Occassionally I do need to bring the vocalist back in. This is why I carefully and with detail note mic distance, EQ, compression, etcetra from anything that contributes by way of 'analog'.
Hope this makes sense...
