Its so hot I can't touch it !

  • Thread starter Thread starter Stan Williams
  • Start date Start date
S

Stan Williams

New member
How hot should the signal be at mixdown? I mean i'm always trying to get the hottest signal i can without distortion, when i'm recording. But at mixdown i'm trying to do the same thing, is that what were trying to do, get the hottest damn signal we can get? Or is there another level i should be going for? I've even made some cassette copies that were almost as hot as some of my cd's. I have just been thinking thats what i'm suppose to do, put the loudest signal i can out to the mixdown deck, without distorting it. I'm using a slight amount of compression .
What do you think?
 
depends on what your gonna do with it. if it's a final product, sure get it as hot as you can. but if your gonna have it mastered, you want to back off some to give the mastering house some headroom to play with.
 
Do you like how it sounds?? if so then OK..:D
Sound always rules..
 
"but if your gonna have it mastered, you want to back off some to give the mastering house some headroom to play with.

Huh? Why? They will surely output the hottest signal THEY can get, and if you put headroom on there, you'll just end up with more noise.

It seems to me that you ALWAYS should record with as high levels as possible... Isn't that right?
 
In the analog world, highest possible level without distortion (to maximize S/N ratio) is the rule. It is not so in the digital world -- you can still work with that ethic, but it's somewhat unnecessary. (For example, at 16-bit resolution, if your meters are hitting -6, then you've reached the 16-bit mark - there's no advantage pushing the levels higher, since you're not gaining anything in terms of signal quality or bitcount.)

If you're having your final mixes mastered, then it IS better to let the mastering houses do all the adjustments -- if it's a quality shop, they have the absolute end in high-level gear, with mastering engineers who really know how to do it properly. A TC Finalizer in the hands of an amateur is no match for a good mastering engineer with their own gear!

Bruce Valeriani
Blue Bear Sound
 
Originally posted by bvaleria
In the analog world, highest possible level without distortion (to maximize S/N ratio) is the rule. It is not so in the digital world -- you can still work with that ethic, but it's somewhat unnecessary. (For example, at 16-bit resolution, if your meters are hitting -6, then you've reached the 16-bit mark - there's no advantage pushing the levels higher, since you're not gaining anything in terms of signal quality or bitcount.)


Well, in fact, the difference between just touching the 16th bit and maxing out is really almost a full bit, so the "at hot as possible" is still quite valid even in the digital world... However, since digital recorders doesn't introduce any noise in themselves, and digital distortion is nasty, it makes sense to sacrifice bit-depth for headroom in the digital world, while in tape recording you really want to hit things with a lot of level to get rid of noise.
 
Sheesh Reg... you take me very literally... OK... I MEANT "average levels at -6, not simply peaking at..." And for higher resolution - you don't even HAVE to worry about hot levels (since you have more resolution to play with at lower levels!)

Bruce
 
Huh? Why? They will surely output the hottest signal THEY can get, and if you put headroom on there, you'll just end up with more noise.

They'll output the hottest signal THEY can get, if that's what the song calls for. You can't just squeeze the dynamic life out of everything. Yes, you want the levels peaking as hot as you can without going over while tracking. But if your S/N is good on all your source tracks, you should have plenty of "play" in the levels at the mixdown stage. How "hot" is vague, too. Peak levels vs. RMS makes a HUGE difference on "how loud" it seems. There is no music if you have a wave file full of all ones-- but the bit resolution is "perfect". It's about dynamics, too.
 
I think the ol' S/N ratio is a bit overtalked these days anyway. When your noise floor is sitting at -50 to -70db there's no reason to worry about squeezing into that -1 to 0 db range when recording. In fact, I've been backing off and just keeping my signal somewhere in the yellowish area. I have not noticed any noise problems and best of all, I find myself retracking less because of clipping, which becomes a serious issue in the digital realm. I suppose it would be different if I was tracking on a cheapo tape machine, but ametth's advice is still sound.

I never thought about it before because I'm barely an amature, but it does make sense to mix a little tiny bit low if your going to have your work mastered...or just mix SENSIBLY...there's no reason to be hitting the red if you don't have a noise problem. Giving the mastering house some room to things on the high side of the signal is probably a good idea. If you DO have such a serious noise problem that 1db is a huge issue...maybe you should invest that $1000 mastering budget into some decent gear!

I think we homerec'ers take the mastering bit for granted because we kinda do it all ourselves. When I say, "I'm finished with my mix" what I'm really saying is, "this project is mixed and 'mastered' and ready to go!"

And regebro...not hitting the red doesn't mean you're sacrificing bit depth. The difference between red and yellow would be ONE, not an entire high order bit! Let's take a simple 8 bit sample that's right at the top of the yellow:

1111 1110 = 254

and the maximum signal would of course be:

1111 1111 = 255

See, 254, or yellow, is one step down from 255, red. So not hitting the red is quite obviously not a big deal.

The only time you run into bit depth problems is if you REALLY REALLY REALLY track low. Watch this and be amazed!

0111 1111 = 127
1111 1111 = 255

YIPES! THAT is sacrificing bit depth because you're not all that room above 127, or HALF your headroom!

When you add a bit to the high side you are DOUBLING your effective headroom. When you subtract a bit from the high side your are halving your effective headroom. But we count from the LOW SIDE, so your logic is backwards. Don't let the bits fool you, the same thing applies in our standard base 10 numbering system. Take a simple 4 digit number:

8999 = 8999
9999 = 9999

When you subtract one from the most significant number (e.g. the number to the left) you are cutting the number by 1/10th. But again, we don't usually do that unless we're trying to cheat somebody out of some money:

9998 = 9998
9999 = 9999

See, we count from right to left, not left to right. The same applies with all numbering systems.
(....ok ok don't slam with with the big endian little endian crap...I know that "left to right" is arbitrary but you know what I'm getting at.)

The point if you missed it: the difference between red and yellow in a 16 bit signal (theoretically), is ONE, not 32,768!!!!!!! (2^16 - 2^15 = 32768) The reason I say theoretcially is because red and yellow are abstract concepts. I could make a VU meter where red started at 65500 instead of 65535 for instance.

Slackmaster 2000
 
up and down

So are you guys saying that if you have no problem with the s/n, and you are going to have it mastered, don'tpush it too hard. And if you are going to master it yourself, push it up a little? If you took 10 major lable cd's off the shelf and compared them, would they all be the same level.[ simular type of music ]
 
My mic pre/digital converter has a gizmo that maps the input signal in the last few dB of headroom onto the output signal in such a way that gives you about 3 dB more room to play with. If you hit it hard, you can hear some artifacts, so I've been aiming at getting all my peaks around -1 dB to minimize this. Doesn't introduce any noticeable noise giving up the space between -1 dB and 0 dB.
 
Slackmaster: Whats "red" differs from software to software (which is why I didn't mention it). I said that the difference between -6db and 0db is one high order bit, which it is.

I'll try to rephrase what I said before to hopefully make my self clear:
There is no reason to not make your levels at hot as possible. The question rather is "what is as hot as possible"? Since digital distortion is nasty, and digital has low noise, the "safe bet" is too keep the top level a couple of db's down from 0 when recording, because you might twack that guitar a bit harder than when you set your levels.
But if you have full control on what your top level is, like when mixing, go for max level. The mastering argument just doesn't cut it. If the song level is higher than what it will end up on the CD, the mastering studio will lower the volume, no problem, but if it is to low, they will crank it up, and then you will have lost some resolution in the process.
 
Regebro, I agree completely.

I was mistaken in that I didn't consider that the db scale is logarithmic. Bummer for me! :)

Slackmaster 2000
 
Mix to what sounds good!

I agree with most of the stuff said in this thread. Inversely, I disagree with much of it in some situations. Sigh.......audio production.....

Getting your mixes as close as possible to the finale sound you want is the most desirable thing you can do. If you can make the mix happen, then that is less for the mastering process to take care of. In effect, the process of mastering pastes a much different sound on the audio then what you can do at mix time. Mastering is about subtlety, not gross fixing!!!

Mastering can be a lot of different things. It can be just arranging the order of songs and applying a slight amount of tonal and/or volume change (same difference really...). Any cut of boost in eq that exceeds around 2db means that the mix had some problems. Any volume change that is more the maybe 3db also means the mix had problems.

I will not go on and on about this. All's that I can say is that you should have your mix as close to possible to the volume and eq you want BEFORE mastering. Mastering is just supposed to add a slight "polish" to the sound. Correct a few discrepencies from song to song. That is all.

Mastering CAN be about really changing what is there too. But this is not without both good and bad results to the original audio. If the mixes are way too quiet and/or the tonality of the mix is off then mastering can certainly correct these things, IF, the engineer understands what he is doing, and what the benefits/drawbacks are. Hopefully, in mixes that need major changes, the benefits outweight the drawbacks. But the skill of the mastering engineer has a lot to do with this.

Don't forget that mastering is every bit as much of a creative contribution to the finale product as songwriting, performing, tracking, and mixing. What the mastering engineer does or does not do to the audio is subjective, and can add benefits and/or drawbacks to the finale product.

Bottom line. Make your mixes as close to what you want. Let mastering take them to the next level. If your mixing skills are weak, you will rely upon the mastering process to take the product to a much higher level. If your mixing skills are good to excellent, the mastering process will just give it a slight nudge to the better (hopefully).

I read an interview with Bob Ludwig, possibly the most well known and respected mastering engineer ever! In it, he said that much of the material he gets in from the big time studios requires nothing more then a 1db volume adjustment from song to song, and in rare cases, a slight eq tweek. Of course he is quite prepared, and quite possibly one of the most skilled people to deal with any drastic changes that may need to take place, but, simply put, he will not touch the audio with any processing if it doesn't need it.
That is not only a testiment to his skill, but also to the skill of the engineers that have mixes submitted to him for mastering. The point is that the mixes sound outstanding without any further processing needing to be done. Anything you do in mastering could possibly be good or bad to that. Sometimes it is just best to leave it alone, even if that means letting a little tweek that you could have done go.

The hottest possibly signal is not always the best sound for the material in question. To illustrate:

Compare Tracy Chapmans Fast Car CD to Everclears first release. You will find that the Tracy Chapman disk hardly ever reaches digital 0. Everclears first release resides in the -4 to 0db range through out the whole CD!!! We are talking about at least a 6db difference in average level between these two CD, more then that in many cases!!! But you know what? Few would argue that the Tracy Chapman CD just flat out sounds way better!!! It has sweeter sounds, and is far easier to listen to throughout. The Everclear CD tires my ears after 2 songs. It is fatiquing to listen to.

So which is better? The CD that is lower in volume, but very listenable? Or the CD that is maxed out in volume, and makes me want to turn it off after 2 songs?

Which way would you want your CD to sound?

Do you want people to get tired of listening to your CD for some inexplainable reason, or do you want them to play it over and over again?

Loud is not always good. People are so hung up with trying to get their home recorded project to sound as loud as big time modern products that were produced in the finest facilities in the world, with the premier engineers in the world. Thing is, these products don't always sound all that great really. Sure, they are clear, and they definately are very loud. But many modern recordings are very annoying to listen to after a couple of songs. Whereas, older recordings that are much quieter are far more enjoyable to listen to, and you find yourself listening to them over and over again. It is not just a difference in the songs, it is how the audio was dealt with. Older recordins just seem to have more dynamics and color. Most modern recordings seem to have very narrow dynamic and tonal characterists.

Mix and master your CD to sound good, not to compete in volume with other products. You may be trying to compete with a product that could have sounded much better at a lower volume. Yes, you can jump on that bandwagon and blast the shit out of it, but that may not do a thing for making your music more enjoyable to the average listener.

I will end here. Either you get this or you don't. If you don't, then you will keep asking how to make your product louder.

If good sound is what you are after, forget levels and just do what sounds good.

Good luck.

Ed
 
mastering................you don't have to worry about leaving headroom for something that is going to be mastered. Headroom is only relevant to a given system. He is starting with a new original signal (just like you starting with a mic) he will record it into his system at the level he wants to allow himself all the headroom he needs to do his job.
While on mastering, he won't change your mix if he doesn't have to. His job is to make it "commercially acceptable" ie to make it sound as good as possible on all those millions of system around the world, which is often a compromise on the sound quality. Most of the stuff I have had mastered doesn't sound as good in the control room (where I mixed it) as the original mix does. However it generally sounds better on most of the other systems I check it on than my non mastered version. He has a monitoring system designed to reflect this, we don't.

metering...........you shouldn't be making mix decisions using peak meters (they are designed to track transients which can be extremely high even though the rest of the track sounds very low) they are there to stop you from getting any distortion. I track to around -6 (give or take) no more, you don't want to retrack something a few times just because you clipped something.
When mixing I use VU meters (which are designed to meter average level) which is the way are ears work, then mix to zero VU (standard practise).

Hope this helps
Brenton
 
Playback volume

Hey Sonusman,
You say that some music that was recorded too hot is hard to listen to compared to some at lower levels, is this due to loss of dynamics or tone or? The reason i bring that up is that the person listening to the cd has control of the volume at playback, regardless of what level it was recorded. There is one thing I do keep in mind, if it sounds good, it is good, to me.
Thanks for the info !
 
Try this

Sit in on a band practice sometime with a db meter in hand. Watch that meter as it goes up and down because of the natural dynamics of the band.

Music is SUPPOSED TO HAVE DYNAMICS DAMNIT!!! A well mastered CD will WILL NOT reside from beginning to end between -4 and 0 on the meters!

So yes, the reason many modern recordings are hard to listen to is because their are little natural dynamics going on. The CD in compressed all to hell, and that becomes annoying to the listener, even at lower volumes, because it just doesn't sound natural.

Reading your meters is a matter of experience. I prefer watching Peak Meters rather then RMS metering (commonly called VU meters, but there ARE Peak VU meters too.....). It is what I am used to. I like to see exactly what is going on. My ears will tell me the rest of the story.

Anyway. Mastering is the last place where you audio can achieve significant inprovements, or significant degregation. If you are not really experienced in dealing with compressors, limiters, and eq, then it is a process you should leave to someone with more experience. Just making it loud is not the idea in mastering. The idea is to "enhance" the audio, and to give the CD a consistency and nice flow. You really have to pay attention to accomplish this.

Ed
 
Back
Top