I don't know if it's flawed & thought a crest value from 10 to 20 dB was normal.
A crest factor of 16dB would mean an RMS lower than -16dB for most parts of the song.
I'd have the ME check it out.

Possibly your monitoring doesn't give you a lot of low end from the 808 kick, which would nevertheless contribute a lot to the peak. Also, bringing up the other tracks doesn't have to make things more muddy, as long as you do more/better EQ sculpting/panning/arrangement clean up so that there's a distinct place for everything.
 
That said, if you're happy with the sound and you're hitting the mark for the new norm, there's probably not much to worry about ;)

16 dB is not excessive.

One of the things I realised after I wrote my post was I was also talking about transients and things compared to rms values much more so than obvious volume level changes. And I guess you are too. Your crest factor is high simply because you have allowed the transients to exist and you have not pulled them down. One of the reasons your mix is sounding good.
 
(What I meant by a film mix have a volume range of 25 db was about effects suddenly being way louder than dialogue, not transients being high compared to rms values)
 
The whole idea of dropping the average level now down to some much lower value eg -23 LUFS is that the (tall) transients can now exist and live beside that lowish average level which is all great in fact. These give snap and punch to the mix. Transients are not necessarily related to things like the 808 kick  (although they would if you had that nice clicky sound to the start of the sound)  but the leading edge of sounds that are percussive.
 
The Orban meter (free) is very good too because it features 5 meters all up and lots of different things can be seen including a reconstructed peak meter which shows if you have any intersample peaks.
 
If you hand your mastering engineer a mix that is predominately sitting at around -23 LUFS then what average level are you going to get him to raise the mix to? Because in reality he should leave it alone and not change the average level of the mix at all see. EQ and compression are still very valid and nice processes to apply but limiting now can be just set down to stop any overs and things. The limiter does not have to raise the overall level in fact. Make sure you know what the ME is going to do average level wise for you. Once you start raising that too much you will loose the sound of your transient mix and then the mix will not be conforming to standards.

bitflipper
16 dB is not excessive.

I read a few articles on the internet that the crest factor should not be more than 12db, plus I see a lot of people stating their pre-master RMS is about -18 db and peaks at -8 to -6 db

so the new standard wants -23 LUFS, isn't that LOW as hell for a mastered music file?   Isn't that -23 RMS?   Sooo confused :(

Your original post was asking about an unmastered mix. 16 dB is not excessive for an unmastered file.
 
If the discussion has now shifted to mastering, 16 dB may or may not be excessive for a final master, depending on the application and genre. Are you mastering for broadcast specifically? R128 is a broadcast standard.

Sorry to create confusion. There are standards for broadcasting for sure and Dave is right with that and -23 LUFS is it for that situation.
 
But the question remains are there standards for CD's?  Bob Katz says the Loudness war is over but it may only be over in certain areas and perhaps not relating to CD's.
 
One thing to think about is iTunes (according to Bob Katz that is) raises the level using its 'Sound Check' feature to a value and -16 has been mentioned for that.  Not sure where Replay Gain is setting theirs.
 
So the question is still how loud are you going to make your master in the end. Be aware that the average rms value of a master is not exactly the same as the LUFS reading. It is possible for the rms value to be higher and the LUFS reading to be a bit lower than rms especially of the music quietens down for any length of time. If the music is constantly reaching a certain rms value for the whole track then the rms and LUFS readings will be closer or almost the same.
 
You need to find out how loud the ME is intending to get the track to reach (on its loudest bits) at least rms wise. (and for how long) -12 to -14 are pretty reasonable values that still sound loudish but still have transients that will make the track punchy.
 
 

I have done some more research into things. I have also asked some questions to Ian Shepherd who is the mastering guru. Read some very good info about him and his mastering concepts here:
 
http://productionadvice.co.uk/loudness-war-secret/
 
Something I recommend you do and others too is to join this organisation:
 
http://www.dynamicrange.de/
 
For a $30 donation and it is well worth it in my opinion you become an active member and you will entitled to download a free TT Dynamic Range meter. There is also tons of info about mastering and mixing there especially relating to keeping mixes alive and dynamic and not too loud.
 
Now I have quite a few measurement systems in place. Firstly K system metering and now I have added LUFS metering into my arsenal. But interestingly the Dynamic Range meter is actually something you still need and is not covered by the Orban LUFS meter. (It may be in other LUFS meters though) BTW I really like the Orban meter and I can highly recommend it. It tells you heaps about your mix. The LRA reading on the Orban meter is not the same as dynamic range BTW. It is a measure of loudness variations. Still handy to know but not a dynamic range reading though.
 
Now your premastered mix will have a high reading for dynamic range. Mastered mixes however will be quite different. Ian says that 6 is terrible, 8 is pretty boderline but values higher than 8 are best eg 9 or 10 for example. So what I am saying above in my previous post is it is OK to make a mix louder than your premastered mix for sure. The mastering engineer if he knows what he is doing should be able to advise on these things regarding a mix:
 
What sort of average rms levels the mix is going to reach. (expressed as a dB figure) For example I work with K system metering and I premaster to K-14 dB. So that means my premastered mix is just hitting -14 dB rms most of the time. After mastering it might be -10 instead. Note I have been going for -7 average for clients that insist on very loud masters and now I am going to STOP IT!)
 
What the overall LUFS reading will be. OK the LUFS reading will be close to the rms reading if the music is hitting the rms mark most of the time. But if there are variations in the mix ie it gets quieter here and there then the LUFS reading will end up a little lower. eg for a dynamic mastered -10 rms mix for example the LUFS reading might be -12 or so.
 
What the dynamic range reading will be.
This is actually the important one. It should be a value somewhat higher than 8 if possible. 10 sounds even better if you can get it. But this is a tradeoff between a very loud rms or LUFS mix and dynamic range.
 
If your mastering engineer does not understand or can give you this info above then he is not worth using and you should find someone who does know this.
 
If you suspect a mix is not very dynamic after mastering and sounds squashed (as is with most idiot ME's !!!) you should run the dynamic range test over it and really see what they have done to destroy your mix. I have checked a few mixes that I have mastered reluctantly loud for stupid clients and they fail the DR test rather badly!
 
The Dynamic Range meter comes in two forms. One is a normal DLL file which gives you live monitoring in the form of a plugin of course. The other is a stand alone app that just loads the track and analyses it. This reading is slightly more accurate apparently. But even so it matches the live plugin usually.
 
So all I can say is you are fine to take your final mix up in level rms wise but not at the expense of the DR meter showing values of 7 and under. If you can keep that reading at 8 or above then you will be sweet with a loud punchy and transient and dynamic mix. All possible, you just have to balance all those things out.
 
BTW Steely Dan's 'Everything Must go' has an average rms reading of K-12 (-14 LUFS) but the DR meter also shows a value of 10. No wonder it sounds so bloody fantastic. It is a perfect balance of a loud master but punchy and transient as hell! It kicks ass!
 
You need all these tools I am talking about to master properly. (that is if you are going to do your own mastering but even if you are not you need to be right onto how any ME engineer has mastered your stuff. Have they done it PROPERLY.) Without them you are shooting in the dark as they say. BTW on Ian's site someone asks the question is K system metering good for nice LUFS levels and great Dynamic Range values and he says YES YES YES!!!! I told you so.
 

Jeff Evans
 
Something I recommend you do and others too is to join this organisation:
 
http://www.dynamicrange.de/
 
For a $30 donation and it is well worth it in my opinion you become an active member and you will entitled to download a free TT Dynamic Range meter. There is also tons of info about mastering and mixing there especially relating to keeping mixes alive and dynamic and not too loud.
 

 
This is old & outdated. (great idea, I joined in 2009)
Brainworx made the DR meter originally ( download for free) & when this so called org went commercial & decided to charge this bogus $30.00 fee to members they pulled out.
Not recommended, I'd call it a scam at best.
 
p.s. last time I looked you could still get this free meter along with the offline TT-DR  at some websites.
The last update was a few years ago. V1.4 IIRC