Autotune on live recordings

R

rob aylestone

Moderator
There's a guy on YouTube who extracts vocals from tracks and then analyses them and his favourite 'hit' is autotune. It seems he's discovered that live recordings are now being remastered including autotuning - so he's singled out Michael Buble and Simon and Garfunkel as examples of this 'terrible' destructive process and how dreadful it is.

He used an example of the orginal Art Garfunkel extracted audio where he did not hit the right note, but the later version had fixed it. I prefer the fixed version personally. He also used a bad example of Buble where clearly the autotune was excessively applied and it was switched off half-way through the live broadcast.

I've not had any examples of singers - especially the better ones where they've not wanted to take advantage. One opera singer, when shown the technique in comping had a very exciting time, highlighting the best phrases from each take - to a degree I could not do, so think very subtle phrasing on certain words. We put together the best out of lots of takes, and then she sat next to me and tweaked all the tuning. She was over the moon with the results, and actually a little shocked how certain words were slightly out of tune.

I'm very happy to condemn over use, or badly done tuning - but what studio artiste would want live work that was not as good?

Extracting vocals can be done now with decent quality in most cases, but then the artistes are getting slammed when the assistance is revealed?

Fair or not fair comment? What do we think here, being recording folk?

The fella with the videos is this one.

 
Autotune is a tool that works. I like it when the midi device can be the control source. Play the vocal line on a keyboard.

Nothing wrong with its use as long as the person can sing in the first place.
 
I like it when used as a gimicky effect, a la Cher's "Believe". Otherwise, I'm kind of a purist. I want professional talent performing as such without assistance. Otherwise, it's no better than a live action "film" which has been overdubbed or is being mimed. Then that's not terrible if it's advertised as such and ticket prices are reduced to $8 as opposed to $168 + for the artist's "live?" performance.
 
I think if I'd paid top whack for tickets, I'd rather hear as close to the released audio as I can. I hate new re-imagined guitar solos, and different melody lines - so while I don't want Cher, I'd rather hear perfect pitch - and with in-ears, being perfectly flat is often quite possible even with really good singers.
 
Most, if not all, of the classic live albums from my youth were full of studio overdubs to fix live flues. This has been common practice for 50 years. Auto tune is just faster and easier than having the singer come in to the studio and sing it several more times. With budgets plummeting, it becomes the only way to get it done.
 
OK, I asked because I've been fixing everything in the daw, after recording it. Didn't know if it would be better to fix it on the way in.
 
Fixing it in the daw is probably better. If you make a mistake on the way in, you will end up fixing it in the daw anyway.
 
I prefer the sound of real humans if they're good performers. Making it too perfect kills it for me. Fixing one iffy note here or there is not so bad, but just leaving pitch correction on for the whole performance takes the life out of it. Like recording on the grid or using a lot of looped samples, it makes it less interesting.
 
bouldersoundguy said:
I prefer the sound of real humans if they're good performers. Making it too perfect kills it for me. it. Like recording on the grid or using a lot of looped samples, it makes it less interesting.
Click to expand...
See now I cannot stand latency errors. Quantization is a must. Sequencers must align. They happen when key load is complex. How would you fix that without 'the grid'? If turn on quantizing say 1/8 you still need the grid to 'snap to'.
 
Live performance is more exciting if there's a real risk of failure. It's like watching a trapeze performance. Knowing they might fall is what makes it interesting. There's a reason watching a video of a trapeze performance isn't as exciting.
 
LazerBeakShiek said:
See now I cannot stand latency errors. Quantization is a must. Sequencers must align. They happen when key load is complex. How would you fix that without 'the grid'? If turn on quantizing say 1/8 you still need the grid to 'snap to'.
Click to expand...
Being dead on the beat isn't always musically correct or preferable. Performers, especially live, push and pull the tempo, which conveys emotional information. Same with pitch. Singers say things by how they approach a note or how they bend it. Taking all that away destroys certain kinds of musical expression.
 
Dead on the beat works for 80s synth pop but is totally wrong for big band, and some singers ALWAYS sing behind the beat as a style thing. If I can't hear pitch correction, then I'd rather have it than hear mistakes.
 
I guess the difference is fixing mistakes vs. Taking away expression.
I've seen some of this guys videos and it really does such the life out of a song when he pitch corrects old Queen and Journey songs, or puts Van Halen to a grid.

A lot of Van Halen was actually played to a click, but lining every hit up to the grid does ruin the feel.
 
I'm generally ok with autotuning. However I am less keen on auto-generated harmonies, whose pitch and timing are too precise to be enjoyable.

Having spend some time in my younger years in what was then British Somaliland, I have an affinity for Somali and Ethiopian music, and their modern music uses autoune heavily as an effect, which I found strange at first, but have gotten to like.

 
I like hearing a singer with expression in their voice. Having everything pitch and beat perfect gets robotic to me. I've done some midi drum tracks, and while they are adequate as a sub for a real drummer, I would prefer to have someone actually playing, and a good singer doing vocals. Sliding into a note sounds real to me.

I'm not a good vocalist, so I end up doing lots of retakes to get something passable. It might just be redoing things a verse at a time (I really don't do short punch-ins) and patching the verses together. I'm sure when I was doing vocals 30 years ago at dances, they were atrocious, but nobody was recording them. After they drifted through the building, they were gone forever. There's a reason I was not a professional singer. It would be like me playing on the PGA tour with a 16 handicap!

If you're going to make your living as a professional singer, you really should be able to sing without every note being corrected. I'll take an occasional slightly sharp or flat note, good vibrato will usually cover it anyway, just like on a violin.

Oh yeah, I'm too cheap to spend money on Melodyne anyway!
 
This has been very helpful. The reason I asked about autotune on the way in, is a fellow I know who has a studio suggested doing this. I couldn't see the reason for doing it. Still don't know, but I will try and see how it works.

I currently use Melodyne to fix my vocals here and there after the track is recorded and put together. I don't put everything exactly in tune because it sounds better a little off in places to me. I also don't quantize, and since I'm moving the grid around to line up with the music (I have a tough time playing to a click), the playback varies slightly throughout the song. While I'm not a terrible singer and my public performance's are OK, I'm not a professional. Like Rich I do multiple takes and for me it's usually the whole song. I sing through each take a little differently until I find what I like and then I concentrate on that version.
 
I was just watching Ella Fitzgerald and Sammy Davis Jr on an old episode of Ed Sullivan. Autotune would have been working overtime had they used it. The way Ella slides into notes would have been killed. They did a scat at the end and Sammy did a very audible dip on a note at the end. These are two superb singers who use pitch variations to add to the music.

I guess the point is to not use correction on everything.

BTW, I was watching a Moonshiners Christmas show (one of my guilty pleasures on Discovery channel), and they did a commercial of the guys each singing a Christmas carole. At the end, two of the guys said something like "Sugarland Shine - the more you sip, the better we sound". I guess it's like Liquid Autotune.
 
I agree that I think it's a tool to use, like most other in the toolbox. Personally, I frequently have a take where the phrasing, inflections, vocal fry etc are all there and sound nice but it's a little flat, so I'll nudge it up. Like any of the tools, too much of one thing can be bad. I don't think there's a way to objectively define what too much is, just like you can't define what too much compression is... sometimes 1-2db does it, sometimes 8-10db sounds best... But at some point it starts to seem ingenuine.
 
You must log in or register to reply here.
Back
Top