Frequency analysis of commercial vs home recordings

JSter

New member
Hi - first post here... I've been recording/ mixing/ mastering my own stuff on a PC (cubase and wavelab) for several years. I'm pretty happy with the results. However - one of the techniques I use when mastering is to capture a frequency curve of a commercial song (similar to the material I'm mastering - anything from Counting Crows to Chilli Peppers to Dixie Chicks) and use it as a guide to mastering my own stuff. One thing I notice is that the commercially mastered stuff is a lot 'flatter' - i.e there aren't any spikes in the graph. My stuff, especially between say 300Hz - 2.5Khz will show several peaks at various frequencies. I'm not entirely sure why this is - I know it isn't a compression or limiting issue since I've tried squashing the life out of a particular track but this doesn't appear to make a difference. The only thing I can think of is that I've got a frequency war going on and I need to go back to the mix and find out what's competing with what. Haven't had much success with this yet. Any ideas folks?
 
if you are using the same mics/pres for every track, or even a just few of them, certain frequencys will be emphasized, giving you those peaks. basically the "character" of the mic you have, which usually allows a track to "sit" better in a mix, will just build up to make for a one-dimensional mix. i dont know if this is even your problem at all, but if it is, try mixing up your mic/pre setups for different tracks more often.

-surf
 
JSter said:
Any ideas folks?
My very first suggestion is to stop comparing frequency curves. You might as well compare automobile quality by the shape of the silhouettes of their respective bodies.

Regardless of what you see on the screen as a correlation between smoothness and professional quality, bumps in a frequency curve are not necessarily a bad thing. Nor is smoothness necessaily a good thing (you want to listen to pink noise all day? That's about as smooth of a curve you can get.)

While a great car may have a sleek silhouette, putting a sleek silhouette on a Yugo will not make it any better. It's confusing cause and effect, form and function. The fact that commercial CDs have smoother curves on them is the end result of a lot of factors in play. Hammering a curve into a similar shape will not magically make your mix sound as good, because that's re-creating the effect without paying attention to the causes.

So what's causing your bumps? It could be gear coloration like surfmaster says. It could be acoustic coloration in two ways: it could be the acoustics of the room in which your recording or it cound be the response of your monitoring chain and the acoustics of your control room (if they are the same room, then there could be a double whammy.) It could be your ears.

Or it could just be that those bumps represent the actual proper response and relative levels of instruments and vocals that you're mixing; i.e. that those bumps are *supposed* to be there.

The fact is that the vast majority of the time there's no way to tell from looking at an RTA display alone whether those are good or bad bumps, whether the curve should be smooth or chunky. You gotta fall back on your ears to do that.

If the mix sounds good to you and most others, then screw what the RTA is telling you. If the mix sounds bad, then fix it according to what your ears are telling you, not by what the RTA has to say. If you can't yet trust your ears, then that's what you have to work on more than anything else, because an RTA is not going to tell you what you need to know the way you think it will.

G.
 
I would greatly appreciate some general feedback on these tracks. Should be a good laugh for you more experienced guys. About training the ear. What analysing frequency curves for particular tracks does allow me to do is to look at a peak on the curve and then cut and boost EQ at that frequency to hear what that peak actually SOUNDS like. Sometimes I won't notice a buildup at a particular frequency until I SEE it and it's not until I cut that frequency that I realise it was a bit overpowering. My ears just aren't that attuned yet. Any hints on ear frequency training appreciated! Oh, and some constructive criticism on my stuff - again, always welcome. Cheers!

Between The Water And The World (Part 2)
Big Wide Open World
Out Of The Blue
 
OK, I am 5 seconds into the first link, and right off it is painfully bright. Mostly I think it's that guitar that needs a setup.

Once everything enters, the bass is muddy and clouds up the mix. It should be punchier and less tubby. The vocal is still crunchy in the 4kHz range, but above that, cymbals, guitar pick attacks, etc., are MIA.

On the whole I think it's a good tune and most of the way there. You may be limited by your monitoring environment at this point.
 
Thanks for the critique mshilarious - so on the first track you listened to - you thought the acoustic guitar is too bright? I'm dissapointed about the bass being too tubby and muddy, because I worked really hard for that not to happen!

When you say the vocal is crunchy at 4kHz - is that a bad thing? Or something to work on? Also, are you saying there isn't enough going on above 4Khz? So, hats, cymbals are too low?

I use a pair of Alesis M1 Active Mark 2's well away from the wall, and I monitor quite low (comfortable listening level) with the monitors fairly close to me. I always cross reference with other commercial material.

I've just discovered this forum, I think it's really great that there are people out there willing to help and critique others. I've only got my wife and the other musicians I work with (who have no idea!).

Cheers!
 
JSter said:
Thanks for the critique mshilarious - so on the first track you listened to - you thought the acoustic guitar is too bright?

The guitar won't be fixed with EQ, I really don't think it's a good sound, like there is fret noise or something. Have a tech look at that guitar.

When you say the vocal is crunchy at 4kHz - is that a bad thing?

Yes, it means the sibilant part of the vocal (consonants) is distorted and too hot. A de-esser will help.

Also, are you saying there isn't enough going on above 4Khz? So, hats, cymbals are too low?

No, I wouldn't say they are too low, but they are missing a part of their spectrum. There could be a number of causes; an EQ change at this point could help but I'd look into why it isn't there for future reference.

I use a pair of Alesis M1 Active Mark 2's well away from the wall, and I monitor quite low (comfortable listening level) with the monitors fairly close to me. I always cross reference with other commercial material.

Well that could be a problem. I am not a big fan of extreme nearfield monitoring; adding low listening levels to the equation seems to be another potential for mischief. I have my monitors six feet away, calibrated to 85dBSPL at my listening position, although I do a lot of listening at 79dBSPL too. Much quieter than that, I don't know what I am listening to, honestly. Turn your mix up, back off, and have a listen.
 
Back
Top