Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[philmagnotta]

Dear Forum:

I read the following article submitted by T.C. Electronic and would like your comments please.
Unfortunately, I was not able to locate the web page to link you there, so hopefuly the pasted article below won't be a problem.
Thanks in advance.

1
0dBFS+ Levels in Digital Mastering
S

[cass anawaty]

Can I get a "bottom line"?  

[dubwise]

Cass Anawaty wrote on Thu, 08 October 2009 16:15

Can I get a "bottom line"?

"Everything's too loud and that's bad."

[dcollins]

Cass Anawaty wrote on Thu, 08 October 2009 13:15

Can I get a "bottom line"?

D/A or sampling rate conversion can create signals greater than 0dBFS.


DC

[philmagnotta]

dcollins wrote on Thu, 08 October 2009 19:59

Cass Anawaty wrote on Thu, 08 October 2009 13:15 Can I get a "bottom line"?   D/A or sampling rate conversion can create signals greater than 0dBFS. DC

Thanks DC,
I do not have experience with these 0dBFS overs, however since it would appear that the 0dBFS indicators would most likely draw attention to their actual status, do these types of outputs somehow get missed or is this one form of max loudness anomaly?

I am new to this forum and from what I can tell, so far, the folks here are very savy regarding, at least, the technical side of things, especially gain/level, etc.

While I am presently engaged as an artist/composer (electric violin), this is my first foray doing it itb.
Now, I do have some knowledge regarding a wide variety of fields in proaudio and I believe that observing digital peak meters in Sonar, eventhough the updating of those meters is somewhat slow, I tend to generally track at about -20 dBFS and in the master out meters, before any compression or other treatments, with all of the cumalitive gain from tracks, I still target for about -9 to -6 dBFS at the master buss.

So, I assume that if there are no overs, then there are none, correct?
However, at the mastering stage, if the end result at the master out is close to 0dBFS, is it possible to have overs that do not display, or is it a matter of the right meters/software?

I will be attempting to master my own material.
I think that some here may find that to be a little pre-mature, but nevertheless, I expect to get good results.

After auditioning many commercial CDs on a fairly high-end system throughout many years, I can state confidently that many of them are not very good and yet those recordings are mastered and handled through the major labels... this is not an area that I am familiar with, that is, how top pop artists proceed engineering/mastering/dup.


 

[TotalSonic]

philmagnotta wrote on Thu, 08 October 2009 20:24

dcollins wrote on Thu, 08 October 2009 19:59 Cass Anawaty wrote on Thu, 08 October 2009 13:15 Can I get a "bottom line"?   D/A or sampling rate conversion can create signals greater than 0dBFS. DC Thanks DC, I do not have experience with these 0dB overs, however since it would appear that the 0dB indicators would most likely draw attention to their actual status, do these types of outputs somehow get missed or is this one form of max loudness anomaly?

There's tons of meters readily available (including a few that are freeware) that allow oversampling for their peak metering so that the potential for intersample peak overs on playback can be easily detected - they're just not part of consumer systems.

Quote:

Is the condition of '0dB overs' being allowed/produced from only unknowledgeable people, as well, some who believe this to be a method for ultimate loudness?

Both.  A lot of mixes (as well as "home masters") these days are made by inexperienced folks unaware of the issue or of proper digital gain staging.  And - a good bit of masters aimed at being "loud" (i.e. current mainstream metal or hip-hop) often use heavy amounts of clipping - which can induce intersample peak overs easily - as a manner of achieving this for their clients while still retaining some semblance of "snap and punch" even at very high average levels - but with the sacrifice being greater distortion.  Sad but true.

Quote:

I was under the impression that CD pressing facilities and or mastering firms, in general, spotted thes issues.

A number of highly regarded mastering studios get paid top dollar in order to create these issues in fact! (see above)

And plants stopped caring about these issues years ago as in these economically difficult times for them they really need to get paid for making discs - not for futile attempts at being the "level police."  

Quote:

Interesting regarding various CD players being unable or faultering when asked to produce these signals.

There are certainly a number of DAC's that can handle intersample peak overs generally very cleanly - but you need to pay a good bit more for these than most folks are willing to for a CD player.

Best regards,
Steve Berson

[philmagnotta]

TotalSonic wrote on Thu, 08 October 2009 21:54

philmagnotta wrote on Thu, 08 October 2009 20:24 dcollins wrote on Thu, 08 October 2009 19:59 Cass Anawaty wrote on Thu, 08 October 2009 13:15 Can I get a "bottom line"?   D/A or sampling rate conversion can create signals greater than 0dBFS. DC Thanks DC, I do not have experience with these 0dB overs, however since it would appear that the 0dB indicators would most likely draw attention to their actual status, do these types of outputs somehow get missed or is this one form of max loudness anomaly? There's tons of meters readily available (including a few that are freeware) that allow oversampling for their peak metering so that the potential for intersample peak overs on playback can be easily detected - they're just not part of consumer systems. Quote: Is the condition of '0dB overs' being allowed/produced from only unknowledgeable people, as well, some who believe this to be a method for ultimate loudness? Both.  A lot of mixes (as well as "home masters") these days are made by inexperienced folks unaware of the issue or of proper digital gain staging.  And - a good bit of masters aimed at being "loud" (i.e. current mainstream metal or hip-hop) often use heavy amounts of clipping - which can induce intersample peak overs easily - as a manner of achieving this for their clients while still retaining some semblance of "snap and punch" even at very high average levels - but with the sacrifice being greater distortion.  Sad but true. Quote: I was under the impression that CD pressing facilities and or mastering firms, in general, spotted thes issues. A number of highly regarded mastering studios get paid top dollar in order to create these issues in fact! (see above) And plants stopped caring about these issues years ago as in these economically difficult times for them they really need to get paid for making discs - not for futile attempts at being the "level police."   Quote: Interesting regarding various CD players being unable or faultering when asked to produce these signals. There are certainly a number of DAC's that can handle intersample peak overs generally very cleanly - but you need to pay a good bit more for these than most folks are willing to for a CD player. Best regards,

... I just installed a Sonalkis Stereo Meter, would this provide that facility to detect/display for intersample peak overs?
I don't know if my system would be classed as a consumer system, but basically I have a custom (Jim Roseberry at studiocat)  intel i7 cpu/vista x64/gigabyte ga x58/6 gig ram at i666 mhz/Sonar 8 producer/three baracuda sata2 drive at 7200 rpm.
Some nice synths and other plugs.
My Zeta electric violin is the main sole of my compostions.
I running it through an eventide h8000fw and also an AxeFX Ultra guitar processor.
My final output is the  Benchmark DAC1 Bryston 4bst and Von Schweikert VR2.
Its a little more than consumer, but it is what I have.
Obviously, at least on this site, it doesn't come close to most of the equipment used by most other here, but I will be able to produce very good sound.
Any suggestions for mastering plugs/outboard?
Obviously, since I don't plan on entering the mastering field next month and I intend to record electric violin/electronic sourced instrumentation mostly, it will be fine for now.
I hope to achive a high level of musicality and quality of sound.
When I feel that I may be in the ball-park sonically, can I post it here for comments?

[cerberus]

do they reach any sort of conclusion, like:  "don't clip"?

jeff dinces

[Tomas Danko]

The brilliant Mr. Paul Frindle has explained a small and easy experiment one can do to prove these things. Google (or better, do a search here on PSW) it up and try it out, it's very enlightening.

The bottom line, again, is really to shy away from full scale. By recording individual tracks at 24 bit dynamic resolution, you will yield a lot of benefits by staying low. Yellow is the new red, green is the new yellow and all of that...

Downmix several dB away from full scale, as well. And finally, if you need to mp3 encode something, or using a similar lossy compression encoder, atleast don't top out above, say, -0.3 dBfs (the lower the better).

I figure for the most part of it all, the things stated above takes care of the practicalities concerning the information stated in that long TC Electronic paper.

[philmagnotta]

Thank you, everyone.
BTW, the civility on this forum is first rate.

[blueboy]

...Link to PDF...

http://www.tcelectronic.com/media/level_paper_aes109.pdf

[philmagnotta]

blueboy wrote on Fri, 09 October 2009 12:50

...Link to PDF... http://www.tcelectronic.com/media/level_paper_aes109.pdf

Blueboy:
Your link is the one that I originally copy/pasted in full at the top of this post.
Did you mean to link to something else?

[Darius van H]

I hope i don't ever get stuck talking to these guys at a party.

[blueboy]

philmagnotta wrote on Fri, 09 October 2009 12:22

blueboy wrote on Fri, 09 October 2009 12:50 ...Link to PDF... http://www.tcelectronic.com/media/level_paper_aes109.pdf Blueboy: Your link is the one that I originally copy/pasted in full at the top of this post. Did you mean to link to something else?

No...your original post said you didn't have a link, and the PDF is much easier to read.

[yoink]

I've thought a lot about this issue, also thanks to Paul Frindle and his very open discussions about these and other digital gain staging issues.

Part of the problem is often amplified by the metering within DAWs or digital audio systems. For example, Logic  ships with a default "Exponential" metering. In this case, any 24 bit files recorded at a nominal -20dbFS = 0dBvu won't really show up on the meters - how ridiculous is that?

I spent a lot of my time, when I was learning/exploring, bouncing off of 0dbFS because a) I didn't know any better and a lot of the old 16-bit hotter-is-better mentality was still around and b) the interfaces of DAWs (Samplitude at the time - I don't know now) encourage such behaviour. Heck even -6dBFS "appears" far too low, for the novice or uninitiated, on the stock meters of Logic Pro in their defaults.

In the past few years I wonder why so many DAW interfaces continue to foster poor gain staging with equally poor visual feedback. These issues could be remedied so easily.

It puzzles me why internal metering on software mixers can't be set to match the same gain staging we use in hardware. It would help so many people who don't know better; this works out even better with the ridiculous headroom of many 24bit+ internal summing and mixing engines of the modern DAWs.  (Say VU-style meter where there's lots of visual resolution below 0db with 0 equating to -20dbFS, and then a very limited visual resolution above 0db, and exponential set of numbers up to +18 or something.)

The other problem is not so much with the metering but the unfortunate emphasis DAWs place, by the very nature of their screen-interface, on the visual feedback instead of the auditory feedback. Essentially, all this stuff forces everyone to pay too much attention to the faux blinkenlights and eye candy that modern software provides.