Mixdown mastering routines?

  On 12/15/2011 at 11:50 PM, TwiddleBot said:

  On 12/14/2011 at 11:20 PM, elusive4 said:

it's about the time-domain, not the frequency-domain.

 

Going thru your posts, are you referring to eg. comb filtering from reflections that occur below the threshold that humans hear an audible reflection?

 

I always thought of the frequency domain as the standard training one does learning about constructive/destructive interference patterns yadda yadda

comb-filtering is simply a manifestation in the frequency response ... which is a result of the summation (superposition) of the direct and indirect signals (specular reflections) at the listening position (destructive and constructive interference). the frequency response gives you zero information as the incident boundaries, time arrival, and gain of these indirect signals that cause such destruction. it provides you zero information on how to hunt down such indirect signals - it merely tells you you are sitting in an interference pattern. people spend big amounts of currency to buy "flat" monitors, and then place them in a bounded space where the bounded space (room) dictates everything.

the time-domain (impulse response, and generated envelope time curve (real+imaginary)) - will detail to you how specular energy impedes the listening position - gain with respect to time. you can think of the impulse response as a sudden event like a balloon popping, only containing the entire frequency range/content. the direct signal will arrive first, due to the straight vector being the shortest path between the source and listening position. off-axis indirect energies (edge diffraction, specular reflections of boundaries, etc) - will arrive slightly later in time (with respect to the direct signal) due to taking a longer flight path (and speed of sound being equal within the medium in your room). eventually, these sparse reflections will become more complex and 'decay' until the energy is entirely damped. when looking in the time-domain of how energy impedes the listening position, one can then use such information to trace the reflection path, and determine which boundaries are responsible for such incident indirect energy - and apply a range of treatments to attenuate the signal.

but a common misconception is that all reflections are bad, or that all (geometric) first reflection points need to be attenuated. depending on the acoustic model one wishes to achieve (the total specular response of the room), one can use the ETC response tool to tailor the room to achieve the designed response. NER (non-environmental rooms) are essentially a completely damped room such that the direct signal is processed and nothing else (the room is entirely dead from the speaker-listening perspective). while 'accurate', it may not be a room many would like to work in for extended periods of time. LEDE (and RFZ), with the advent of the TEF analyzer such that the time-domain could finally be measured, built on the psycho-acoustics to achieve a specular response that gives the brain ample time to digest the direct signal while then providing a diffused termination and decay such that the room is not entirely dead. there is much, much more to LEDE (and RFZ) than this, and much of what anyone discusses of LEDE is entirely incorrect (LEDE is one of the most misinformed topics i have come across). from the Beranek and Schroeder days of studying concert halls and the time arrivals of what makes such particular halls sound pleasing - the ITG (initial time gap) was modified for small acoustical spaces to the ISD (initial signal delay). the ISD is essentially the time-delay length from when the direct signal reaches your ears, to when the next significant reflection impedes the listening position. the longer this time gap, the longer the reflection had to travel, which increases the perceived size of the room. naturally, large rooms wil have large ISD's due to the inherent nature of the boundaries being so far away. by attenuating the first order reflections in a small room, one can artificially increase the ISD gap and give an increased perceived size of the space. with LEDE, the critical factor in all of this was to make the control room's ISD gap be 2-5ms longer than the tracking/recording room's ISD. this way, when the engineer is in the control room, the reflections from the recording room (recorded on the mic) will arrive earlier than the reflections of the room that the music is being reproduced in (control room). what this does, is let the engineer hear the totality of the recorded room before the control room has a change to 'mask' it's sound on top. think of listening to a recording in a large room such as a concert hall. now play it back in a small space such as your bathroom. the reflections in the bathroom dictate to your brain how 'large or small' the perceived size of the room is - and will mask the room/reverb that is in the recording from the concert hall. you can close your eyes, and your brain will not be tricked at all into thinking it is in the concert hall - the arrival time of the indirect energy dictates to the ear-brain that it is indeed in a very small space. by increasing the ISD of the listening/control room, you will be able to hear the totality of the recording room - and it will be much easier to immerse yourself into the reproduction and simulate that you are in the room the music was recorded in.

as such, once the length of the ISD (eg, 20ms) has been definied, which essentially dictates the acoustical perceived size of the room, then one would set about objectively measuring from the listening position how specular energy is impeding the listening position. the ISD is essentially an effectively anechoic time period. not fully anechoic, as then you would haev a completely damped room, but the goal is to make the room effectively anechoic for the length of the ISD - such that the only signal the ear-brain receives is the direct signal. any energy high enough in gain and arriving within this time period is said to be 'early'. the goal is to attenuate any early arriving energy to below the human detection threshold (eg, -20dB to -30dB dependent upon time arrival). now, this does not mean that the reflections are not heard, but , they are so far down in gain that the brain no longer processes them and they no longer compete with the direct signal. this is why it is referred to as being 'effectively' anechoic. attenuation of these signals can be done in a magnitude of ways, but for cost effectiveness, absorption with porous rigid fiberglass or equivilant mineral wool is usually preferred.

now, eventually, energy must finally impede the listening position otherwise you truly would have an anechoic room (which is no fun!). so, due to psycho-acoustics, there is also a very specific way to 'terminate' this 'effectively anechoic' time period - and that is called the 'termination'. the termination is the first significant energy to impede the listening position, signifying the end of the ISD-gap....sharply delineating the end of the ISD-gap, and signifying the start of the decay. originally, a loud (high gain) specular reflection was used for the termination, but as the models evolved, and with the advent of Schroeder diffusers, a diffused termination off of the rear wall was preferred. the gain of the termination (gain with respect to the direct signal) will determine the livliness of the room - and the rate of decay will dictate the spaciousness of the room. the precendence effect (corollary) allows this termination to psycho-acoustically force the brain to 'lock-on' to the direct signal, and thus, remove the directional components from the remaining specular energy bouncing around within the room. however, to do this, the gain should not be below -12dB of the origianl signal, and is very difficult to do with a diffused return (which, naterually , will be much lower in gain).

ideally, large reflectors (large with respect to wavelength) are used to redirect the 'early reflections' away from the listening position and towards the rear wall such that the precious finite energy can be maintained and 'managed' such that it can be presented back to the listening position and a later time (eg, the termination). this is one of the core designs with regards to the RFZ room model.

since many of us are constrained to such small spaces, the early arriving (high gain) specular energy only causes an ongoing magnitude of problems - intelligibility (multiple indirect energies arriving within the fusion/haas interval zone - reducing clarity), localization (directivity of the direct signal is skewed due to early arriving energy from the sidewalls), frequency response comb-filtering (due to summation of high gain indirect signals with the direct signal, with time-offset such that phase differences produce constructive and destructive itnerference), and perceived size of the room (earlier the first significant reflection arrives, the smaller the perceived size of the room). broadband absorption is by far the easiest, cheapest, and highly effective form of treatment to attenuate such early specular indirect energies. for the remaining energy, highly complex surfaces that can diffuse spatially and temporally are preferred - to break up the 'hot' gain specular reflections bouncing around like laser beams into fully diffused returns in many directions (like a lamp shade) - this also greatly narrows the polar lobing. the advent of dr. manfred schroeder using quadratic residue and primitive root number theory to build mathmatically predictable 'diffusers' was one of the great discoveries that led to control room model designs.

again, the time-domain is in reference to the specular region (above ~ 300hz). the modal region issues are attacked slightly differently. the wavelengths are longer and energy content is much higher vs frequencies in the specular region, which makes for a bit of a challenge. save that conversation for another time.

look at the photos i posted on the rpevious page - especially the ETC response. you have to visualize how specular energy (that behaves geometrically - like billiards bouncing around a pool table (angle of incidence = angle of reflection) are arriving at the listening position with respect to time of the direct signal. managing these reflections will do more for quality than any gear or speaker upgrade will get you. real, objectionably measured issues. if you're further interested in how to measure for such issues and how to proceed, let me know. ok?

  On 12/16/2011 at 12:10 AM, TwiddleBot said:

Ah, yes, in our recording engineering classes way back when we talked about interference patterns and reflections but not so much this. Hence my previous comments. I'm still trying to understand though why both time and frequency wouldn't be important. EG proper placement of a subwoofer in a room is very much based on the frequency domain, no?

modal region (where sound behaves like waves) is different than specular region (where sound behaves and is modeled like light (geometry)).

frequency response is important in the modal region (along with the time-domain which gives you 3 dimensions (freq response + time decay)) - this is referred to as the waterfall plot (google image search).

again, this is still really dumbed down and basic - but the foundations are important. how energy arrives (in the time-domain) with respect to the direct signal is the root of many acoustical models.

  On 12/15/2011 at 11:43 PM, missingsense said:

hey elusive chill man, I'm looking to get my room in better shape for audio, do I need some sort of SPL meter to get calibrated results or can I somehow calibrate a (quality) mic with just the mic, preamp and my room? I just want to get some basic but practical results here, my room could be pretty good for mixing I reckon, but I need to fix a bunch of stuff for sure.. I hear you on the phase distortion smearing, that sounds very plausible actually to what I'm struggling with when I'm mixing in my room, and not on the headphones. didn't read much of what you wrote beyond what I'm interested in for the moment, but you should probably stick to the basics if you want to explain anything here, cause right now you are probably the worst teacher ever, alright? and I believe you're not getting paid or anything.. it's completely beyond me why you want to be so miserable about something you are well into. if you don't have the patience/skill to draw people in then just leave it haha. man I wouldn't want to be a teacher I would probably get tired of that really quickly actual..... tl;dw

a good, low-cost omni mic such as dayton emm-6 or berheingr ecm8000, a pre-amp, and software (room eq wizard is free - fuzz measure pro for the mac guys, arta, etc).

the frequency response is really only relevant in the modal region (0-300hz), and only when used with the time-domain to determine the LF modal ringing (LF decay times) - this is referred to as the waterfall plot.

in the specular region, the frequency-response is really not relevant because it merely displays the interference pattern you are in due to the summation (superposition) of the direct and indirect signals which cause comb-filtering. note that comb-filtering does not exist in the real world, it is merely a manifestation in the frequency response. polar lobing, on the other hand - is very much real. we use the Envelope Time Curve (ETC) response measurement for the specular region.

alright thanks I can work with that info. already spent some time in the software.. I have a Rode NT1-A driven by my mixing desk, that will do the job right?

here is a nice 'real world' example of hunting down specular energy via use of the envelope time curve (i use magikman's thread a lot for reference, as he illustrated all of his steps beautifully):

thread;

http://www.gearslutz.com/board/bass-traps-acoustic-panels-foam-etc/559191-magickmans-room-tuning-thread.html#post6098894

relevant ETC investigations (individual posts):

http://www.gearslutz.com/board/6118734-post47.html

and http://www.gearslutz.com/board/6145795-post59.html

  On 12/16/2011 at 1:14 AM, missingsense said:

alright thanks I can work with that info. already spent some time in the software.. I have a Rode NT1-A driven by my mixing desk, that will do the job right?

im not a hardware guy -

That is a very clear and informative response, and more than totally answered my question, thank you. (Aside from a couple terms I've to look up as time permits)! I'll probably direct a friend or two here to read this as we're very interested in such topics, as we do indeed on occasion when time and $ permits experiment with setting up rooms (though typically rooms that are sadly an acoustic nightmare, and with limited materials).

andreas (always a great contributor!)'s thread provides some very clearly illustrated ETC's from his build thread.

http://www.gearslutz.com/board/studio-building-acoustics/481797-lede-room-haas-trigger.html

notice how clean the ISD is, and then in the 2nd ETC the termination (loud gain specular energy that sharply delineates the end of the ISD).

and probalby the most imprtant thing that people need to take away from all of this, is the early arriving energy is bad (arriving within the ISD), but the early-early arriving energy (1-3ms after direct signal, causing image shift) is really the most detrimental. this is generally due to the speakers/monitors placed on a large, flat working surface (desk) or having a nice big fancy mixing board in front of you (which very well may impress family and friends, but is detrimental until made acoustically transparent!) - hence, get the speakers on stands many feet in front of the desk such that no incident specular energy is reflected off of the desk (can also be assisted by slightly angling the desk downwards).

  On 12/16/2011 at 1:21 AM, TwiddleBot said:

That is a very clear and informative response, and more than totally answered my question, thank you. (Aside from a couple terms I've to look up as time permits)! I'll probably direct a friend or two here to read this as we're very interested in such topics, as we do indeed on occasion when time and $ permits experiment with setting up rooms (though typically rooms that are sadly an acoustic nightmare, and with limited materials).

ask questions.

post pictures (layout) of room.

become familiar with the measurement tools.

have fun.

  On 12/16/2011 at 1:21 AM, TwiddleBot said:

That is a very clear and informative response, and more than totally answered my question, thank you. (Aside from a couple terms I've to look up as time permits)! I'll probably direct a friend or two here to read this as we're very interested in such topics, as we do indeed on occasion when time and $ permits experiment with setting up rooms (though typically rooms that are sadly an acoustic nightmare, and with limited materials).

also, just remember the acoustic models are strictly defined for a reason! LEDE is one of the most scientifically developed models - yet also one of the most highly misunderstood.

if anyone tells you LEDE means "live end dead end" - that is the quick litmus test to know that they have zero clue what they're talking about.

also, if you're going to get into this, d'antonio is the authoritative figure on the subject (http://www.rpginc.com/) - the one who basically commercialized Schroeder diffusers such that they could be introduced to the public.

http://books.google.com/books?id=jYBR-aL2vJMC&printsec=frontcover&dq=Acoustic+diffusers+and+absorbers&hl=en&ei=-W4oTaqMIozInAehg5SOAQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CC8Q6AEwAA#v=onepage&q&f=false

http://www.amazon.com/Acoustic-Absorbers-Diffusers-Application-ebook/dp/B001Y35IHW

play nice with me and ill get you a copy.

Here's what a pro mastering room looks like.

Don't let that NOC list get in the open!

  On 12/16/2011 at 5:28 AM, chenGOD said:

Here's what a pro mastering room looks like.

 

yeah but (null) and stuff!

  On 12/16/2011 at 9:24 PM, sneaksta303 said:

  On 12/16/2011 at 5:28 AM, chenGOD said:

Here's what a pro mastering room looks like.

 

(null) but (null) and (null)!

Sorry I couldn't hear you.. too many nulls.