Mixing Question - Levels

[mojobone]

I don't often pause to acknowledge, but thanks for your kind words. Isaac Asimov was my boyhood hero; he was lousy at dialogue, but had an incredible gift for making science understandable. Had I a gift for mathematics, rather than verbiage/verbosity, I'd have probably become an astronaut.

[ernstinen]

I think we dealt with the original question pretty well; the upshot is, it harms nothing to turn your signals down when mixing in a 24-bit digital system. At 16 bits, less detail is captured in low-level signals, so best practices involved keeping the levels hot. It's no longer necessary, and even then, was only necessary during capture, not mixing.

O.K., "Necessary during capture, not mixing" makes complete sense. I hope we're all agreed on that one point! Thanks, Mojo.

Next: Bob Moog's take on analog vs. digital in frequency response (or NOT heheh! ).

Peace and love,

Ern

[mojobone]

Next: Bob Moog's take on analog vs. digital in frequency response

Not for nuthin', but the question has been raised; "Can digital ever beat analog?" I think it can, and perhaps it already has, but since our ears are analog, we'd never know, for certain.

[gtrmann]

I think we dealt with the original question pretty well; the upshot is, it harms nothing to turn your signals down when mixing in a 24-bit digital system. At 16 bits, less detail is captured in low-level signals, so best practices involved keeping the levels hot. It's no longer necessary, and even then, was only necessary during capture, not mixing.

O.K., "Necessary during capture, not mixing" makes complete sense. I hope we're all agreed on that one point! Thanks, Mojo.

Next: Bob Moog's take on analog vs. digital in frequency response (or NOT heheh! ).

Peace and love,

Ern

I think it would be fun to talk about analog vs current digital.... I am a electronics tech. I work on music gear for a living. I have for almost 30 years. I have some serious opinions about this subject, I love and embrace both formats.......... I still own a fully functional Tascam 1" 16 track machine.....

Take care Mr. Ernstinen

[mazz]

Hi everybody,
I am sure I am going to catch a lot of flack for this, but.... the way I see it is....

The original question was about how if you mixed many tracks together that were all peaking at around 0dBFS, the final mix would be over 0dBFS and distorted. The person had to bring the levels of his tracks down in order to keep the final two mix under 0dBFS......

I am not sure what that has do do with the sampling rate, the amount of bits used per sample, or the 0dBFS standard of -18DBu.....

No wonder why we are not getting more forwards...... we are all off target...........

I am here to try and make friends....so if I go to the rally this year, I will have some people to hang out with.....

Ok ....you can shoot me now....!!!!!!

I was one of the main offenders of veering off course, and yet I still think that understanding the "under the hood" stuff that affects ones decisions to turn it down or not is relevant to the original post. I admit it was a bit of a tangent to the original and not a direct answer, but I could see that folks were throwing terminology around without understanding it fully and as Mojo says, accuracy is key in discussing audio engineering.

Back to point:

I think that overloading the summing bus in a DAW is worse than underloading it when mixing and that having a consistent reference level, whatever that ends up being, gives a good place to start when mixing, even if hitting that level means turning things down.

I'm not offended by being taken to task for veering, I take full responsibility for letting my ego run the show in this case! My apologies.

On to more important things! Like writing music!

Cheers!

Mazz

[jdhogg]

mojo
re "The term "bit rate" comes from data compression, and is meaningless in PCM digital sampling, but used to be interchangeable with sample rate, before we had such a thing as data compression, hence the confusion."

I dont remember "bit rate" being interchangeable with sample rate as they are not the same. Bit rate is a comms spec which was later used to describe mp3 encoding quality.

[ernstinen]

I love and embrace both formats.......... I still own a fully functional Tascam 1" 16 track machine.....

Ha! I recorded some of my best sounding stuff on a Tascam reel-to-reel 1/2" 8 track. I love digital, too. --- But, I don't get this "analog is -18db from digital" thing yet. I swear one recording that I did with a real piano and some analog synths sounds a lot louder and warmer than anything I've done since, and that was at least 15 years ago. Sure, the tracks were compressed (some sequenced) and there was tape compression going through a hot analog board, but it "sounds" at least +5db louder (even on CD) than most digital recordings.

An analog vs. digital conversation would be fun for *me*, but in the past the give-and-take on this subject almost started a flame war --- People are passionate about this stuff! So maybe someone would like to start a new thread!? It won't be me!

BTW, after moving recently, I've rediscovered some old vinyl LPs. Bought a new Grado cartridge, and WOW! Vinyl sounds Soooo good (but don't quote me on that Haha!).

My Best,

Ern

[gtrmann]

I love and embrace both formats.......... I still own a fully functional Tascam 1" 16 track machine.....

Ha! I recorded some of my best sounding stuff on a Tascam reel-to-reel 1/2" 8 track. I love digital, too. --- But, I don't get this "analog is -18db from digital" thing yet. I swear one recording that I did with a real piano and some analog synths sounds a lot louder and warmer than anything I've done since, and that was at least 15 years ago. Sure, the tracks were compressed (some sequenced) and there was tape compression going through a hot analog board, but it "sounds" at least +5db louder (even on CD) than most digital recordings.

An analog vs. digital conversation would be fun for *me*, but in the past the give-and-take on this subject almost started a flame war --- People are passionate about this stuff! So maybe someone would like to start a new thread!? It won't be me!

BTW, after moving recently, I've rediscovered some old vinyl LPs. Bought a new Grado cartridge, and WOW! Vinyl sounds Soooo good (but don't quote me on that Haha!).


My Best,

Ern

How I read this is........

The -18dBFS spec has to do with analog to digital.... some one decided a level standard when coming out of a mixer or other analog device into a digital device........... The US standard is...... when the analog device is reading 0dB on it's output metering the digital device input meter should read -18dBFS..... I believe the 0dB on the analog device is referenced +4dBu or around 1.3 VRMS ( Pro levels )........ I believe this spec was set up to keep the user from going over 0 dBFS on the music peaks.... is see this as only useful info for tracking or mixing analog to digital......

I believe this standard was set by the broadcast industry..... that is where the music industry gets most of it's standards.....

So what you want to do is put a 1K tone in you mixer / preamp and adjust it's output to read 0dB / +4dBu....... then adjust the input levels to your digital recorder / dat mahine / interface...... to read -18dBFS on it's meters when in the record ready mode.......

I think I saw in a earlier post you own a Soundcraft mixer, OK, were brothers now... I own a Soundcraft 800B-24 mixer.....

Talk to you later .....brother....

Sorry to everybody else for taking this blog way way off topic.......... I will ty to restrain myself in th future....

[mazz]

Since we're going back down the level road:

For music recording, there is no level standard with digital (there were electrical standards with Analog but folks even pushed those to get a better signal to noise ratio). Broadcast has had standards in place for decades, probably partly because it's regulated by the FCC (radio stations could get fined for "overmodulation", for instance) and music recording isn't regulated at all.

I think Katz' level premise also has to do with bringing back this old fashioned thing called "Dyanmic Range", a quaint old concept left over from the pre loudness wars days and also still practiced in orchestral music and other classical ensembles . If I understand what he is saying, pushing the levels up toward 0dbfs causes other issues like distortion that don't come into play if the average levels are more down around -18dbfs. I think his argument has to do with the D/A converters in consumer playback systems not being able to handle the electrical signals as well. Of course, this could just become one of the "sounds" of digital and people will get used to it and use it to their advantage. I think that's already going on in some styles of music.

But you still can't have all your tracks with a super hot average level and mix them all together with the faders up at 0dbU without overloading the bus, you have to turn stuff down to make it all "fit" in the available space, whatever reference the mixer chooses. Hitting a bus compressor super hard with all that level can cause it to distort (depending on the compressor design) too. There's such a thing as "gain management" which those of us that worked on analog gear had to learn in order to keep a good signal to noise ratio all the way down the line. Nowadays, for those of us working with super well recorded digital samples, noise is not really an issue, but gain staging, even in the virtual world, is still a valid concept, yet another reason to talk about this in context of the original post. All signals going into a plugin can't necessarily be super hot as recorded, depending on the plugin and it's function and design.

There's a reason that recording at home is just not as simple as throwing up a mic and pushing record. People spend their lives learning to engineer, just like people spend their lives learning an instrument and practicing songwriting or composing. It's not as simple as the sales hype would make it to be, it still takes study and practice to get it right, it's like any other skill.

I'm enjoying this thread, even if I nudged it off track a time or two.

Mazz

PS: Some of my best sounding music is still the New Age album I did in 1993 on a 2" Otari machine through an Amek Angela board, mixed to DAT tape (DAT not my favorite format). I LOVED the headroom available in the Amek. It is possible to get that kind of headroom in digital by using the level strategy discussed in this thread.

[gtrmann]

Should we start another topic for this............

The difference between the 0 VU meter +4dBu output level of a analog board, and the max output at clipping on a analog board is around 18dB....
+22Bu - +4dBu = 18dB headroom......hey I recognize that number ( -18dBFS )

......my old Soundcraft only gives 17dB....Whine, whine , sob, sob....

Dynamic range basically is the difference between the loudest and softest audio possible....
Digital audio dynamic range is represented in bits... one bit = 6dB

Theoretically
16 bits x 6 dB = 96 dB dynamic range
24 bits x 6 dB = 144 dB dynamic range

24 bit audio will not produce 144dB dynamic range currently because of resistor noise in the A/D converters ..... it gives about 124dB real world dynamic range ( or signal to noise ratio as it was known in analog audio )

124db ( 24 bit real world ) minus 96dB ( 16 bit ) equals 18dB .....that same number again ( -18dBFS )

So if you mix into 24 bits and use -18dBFS as your 0 reference you will get the same dynamic range as 16 bit ( and a analog console headroom), but will have better resolution than 16 bits, because you will use 21 bits.....

( 144dB [ 24 bit theoretically ] - 18dB [ desired headroom ] ) / 6 = 21 bits

Just mix, normalize, truncate to 16 bits, master to suit your tastes....

I guess the object is just to make it sound good anyway.......

As a side note, the specs on my old Tascam MS16 1" 16 track reel to reel is 107dB weighted and 100dB unweighted... which on paper is better than 16 bit digital audio.....go figure

Don't get me started on sampling rates........