I've been working on a song that felt like it dragged a little bit. So, I bounced all the tracks to create a new track for mastering. Then, I did the "ctrl+click-drag on the end of the clip" time-stretch function and dragged leftward to shorten it a bit, thus speeding up the tempo. Surprisingly (at least to me!), shortening the song from 4:27 to 4:22 put the "feel" where it needed to be. I bounced to clip to do offline rendering for optimum fidelity, then did the mastering.
 
I would have thought that 5 seconds out of a 4:27 song would be insignificant, but that definitely wasn't the case. In fact cutting a few more seconds off it felt really rushed.
 
Seemed interesting enough to merit a post here.
 
 
 

Depending on the groove and overall tempo, a 1BPM difference can really matter!
It's probably more a percentage thing. You shorted a 267 second clip by five seconds, so about two percent difference, like going from 100BPM to 102, or 148 to 151. I would definitely call that significant.

Craig,
 
I agree.
 
I did a cover of Like A Rolling Stone at 95BPM that when I bumped it to 100BPM it had movement that I liked. I kept it simple and only bumped the final mix (at mastering).

Tempo changes are a great way to create tension and also a great way of releasing it.  I got in so many arguments about this as a teenager because 1-3 bpm can be both very subtle & very dramatic. 

Let's face it, playing live (without a click track) is never 120BPM for three to five minutes straight. There will be fluctuations of at least 1BPM.

There is an internal clock in the brain that is extremely sensitive to the timing of music commonly called groove. That is what allows musicians 50 feet apart on a stage to stay in time even though the "latency" introduced by the speed of sound over that distance would be disturbing if it were coming from a computer. They are not actually hearing the sound from the other musicians and coordinating to it. They are all following their own synchronized internal clocks.
 
The designers of current CPR guidelines have taken this effect into account by recommending that chest compressions in cardiac resuscitation should be timed to the beat of the BeeGees "Stayin Alive."  In formal studies it has been shown that people not only remember the music but that they reproduce the tempo when they do so to within a very narrow and reproducible range. God knows what will happen to patients stricken with heart attacks when the last of us who still remember disco are gone.
 
http://download.springer....8fd23c0d7&ext=.pdf

I'm just working on a song that feels right starting at 109bpm. When I finished the basic backing tracks I moved the choruses to 110, dropping back to 109 for the verses and things now feel just right.
 
I always experiment with changing tempos. Some songs I've done sounded great with a 1% or so increase every verse and every chorus, so ramping up a fair bit by then end.  There's no formula, I just experiment and listen to the feel. But it's surprising how such tweaks can really make a difference to the groove.     

The click track has done more to harm recorded music than any other innovation short of 100% quantization and auto-tune. Here's a guy who likes to try and figure out which classic songs were and were not recorded with a click:
http://musicmachinery.com/2009/03/02/in-search-of-the-click-track/. Here's a program that'll let you do the same thing: http://static.echonest.com/bpmexp/bpmexp.html
 
Most of us here probably don't even use a click track, but rather a programmed MIDI drum track. Same thing but even worse because you lose all the timing inflections that a real drummer would provide. Everybody who uses programmed drums is constantly trying to figure out why it never sounds quite right. That's why.
 
SONAR's tempo map can help, though. Where I start is with very small tempo changes between each verse and chorus: up going into the chorus, halfway back down going into the next verse. Each chorus is then slightly faster than the preceding verse, each verse is slightly faster than the previous verse, and each chorus is slightly faster than the previous chorus. The overall song slowly ratchets upward in tempo until the finale is perhaps 3 bpm faster than the intro. 

the same can be said for slowing a piece down as well.
just a few seconds.
 
you can hear a lot more in a song with a slower tempo than a song with a faster tempo
this does not apply to all situations but I think you know what I mean.

slartabartfast
The designers of current CPR guidelines have taken this effect into account by recommending that chest compressions in cardiac resuscitation should be timed to the beat of the BeeGees "Stayin Alive."  In formal studies it has been shown that people not only remember the music but that they reproduce the tempo when they do so to within a very narrow and reproducible range. God knows what will happen to patients stricken with heart attacks when the last of us who still remember disco are gone.

The things I learn from this forum...