Your HRTF argument has to change to have logical discourse.
Every recording is EQed even if it is straight off the mic.
Every single one, and that's just the tip of the iceberg. Add in the total # of mics in use and the room effects In a technical sense there is an HRTF coloration. In a real sense it is not an issue b/c people adjust the recording until it sounds good. They even engineer the initial flaw(s) to be a technical flaw that sounds good. Forget about undoing the recording. Not just b/c it is impossible even if you knew how the original sounded, but b/c you don't know what the original event sounded like. Every recording I have listened to has been mixed and/or manipulated. You too.
Is this helping?
Dan
Dan, your argument that this effect is always fixed in the mix would be stronger if there was any written evidence in the various teaching resources on recording that this effect was acknowledged, given its a constant and inherent artifact with stereo. Do engineers eq it to sound natural in the natural process of mastering? Maybe but you would think this would be "standard" practice if it were so well understood. Your argument is analogous to saying that vented cabinets were alll well tuned in the era before TS params, because they listened to their speakers when they made them.
So, I'm all ears, I know you have numerous resources on mastering, so please share references that this is acknowledged and consciously dealt with.
You argument doesn't really apply in this case though since we're comparing the difference between a stereo mix down to a mono center channel vs 2 speaker stereo phantom reproduction of the mono center image. In this case, the perceived tonal difference is independent of the recording.
Dave
So, I'm all ears, I know you have numerous resources on mastering, so please share references that this is acknowledged and consciously dealt with.
You argument doesn't really apply in this case though since we're comparing the difference between a stereo mix down to a mono center channel vs 2 speaker stereo phantom reproduction of the mono center image. In this case, the perceived tonal difference is independent of the recording.
Dave
I have read so many times that a good monitor starts with a flat on axis response .......... I also read from a respected source who was comparing how media for film has a standard reference for mixers and music audio does not , that flat up to 2khz , then a 3db slope down from there is prefered .
The thing is that the ear seems to like the pink noise slope . I have done FFT anaylsis on a cross section of pop songs and save a few exceptions ( ususally newer loudness war artrocities), the wave file ( before it ever hits the speakers ) has the slope in it .
The thing is that the ear seems to like the pink noise slope . I have done FFT anaylsis on a cross section of pop songs and save a few exceptions ( ususally newer loudness war artrocities), the wave file ( before it ever hits the speakers ) has the slope in it .
IOW, you have no recording where this is an issue? That's what I'm taking from this. My 'argument' is crazy strong. You are the one who needs to defend your point or at least have some support of it. It would be easy to do if you had it. There is no 'natural' process to mastering and there would need to be standards for recording, song writing, and mixing before there could be one for mastering. There are of course media standards.
That's actually not my argument. Microphones are not flat and recordings are manipulated to sound "good" or hopefully even better in stereo. You are never capable of hearing what a voice or instrument sounded like stereo or mono straight off a track. Some are closer than others. Read "Behind the Glass" for several famous producers take on this. They all agree with me--or rather I agree with them. In binaural, you can get very close however! Frighteningly close in fact, but that is not in the discussion. The thing is no one is capturing this perfectly natural sound. I don't even know of a recording that is going for a perfectly natural sound. If they were, this curve could not be applied to any track and expected to work. Some recording are going for more natural than others, but all are enhanced or manipulated in a fairly unpredictable fashion in attempt to make them sound better. There are no 'standards' and that is why you can't just apply some curve to your output and it will work to undo the recording process. It would be far more complicated than that if there was some way to undo it! Maybe you'll like it better with your HRTF curve applied, maybe you'll think "that sounds more natural", it just doesn't make it true.
My argument is nothing like your analogy.
I'm not sure why you have such trouble with the concept and surprised by the analogy. Don't over think it b/c it's simple. Believe me. If I can get it, you can. You can challenge Dr. Geddes, Dave, and all the other intellects on the board. I really cannot. IOW, I think you are very bright and it puzzles me to think of why you are having trouble understanding how complicated the recordings you buy truly are.Mixing Audio by Roey Izhaki goes into this HRTF business and the notch FWIW. It's not alone, but that's off the top of my head. Certainly not all mixing engineers are up on this, but since it is not a real problem for stereo recordings to begin with, it doesn't matter. I'd venture to say most of them have no clue about it from my online chats. Fortunately it doesn't actually matter in the manner audiophiles fret over.
My software does have several HRTF presets for various angles you can apply to anything on the recording anytime or the whole thing all the time. So plenty of people in the business are aware that HRTFs exist and that applying them to a recording in various situations can be useful--switching playback formats for instance. I'm sure I mentioned this before when you originally brought it up here on DIYA.I have one resource on mastering--Bob Katz's book. No one is consciously dealing with it while mixing for stereo b/c it is not an genuine 'problem' to begin with. It's just not that simple. Mixing is my primary concern at the moment. I like to write/record songs. When I get better at mixing and have at least an album worth of decent recordings, I'll look more serious at mastering. By the time a recording gets to mastering it has been heavily manipulated. This is the case in every commercial recording I know of. The reason for this is that the sound is easily improved straight off the track. IOW, you can make it sound better than the track--and in a variety of ways. It is also possible to make it worse of course. In mastering, it will be manipulated further-but the mastering process could not fix this supposed 'problem' for stereo playback. The degree and manners of which varies track by track and recording by recording. It's actually common knowledge and people on the HiFi/audiophile boards complain about it all the time. Less so in recording forums. No one is consciously dealing with the 'problem' b/c the goal is for the recording to sound good and often unfortunately "loud". They don't have access to a perfect signal(whatever that is) to begin with to cause that proposed 'problem'. I'd venture to say most HiFi enthusiasts do not understand the recording process and the stages/issues involved from the posts in this thread and all others where this comes up. If this HRTF thing was a real issue and it was still a problem by the time it went to mastering, it would be too late for mastering to fix it in a way that would make sense. They could essentially only do what you propose doing to your stereo speakers and that will cause more problems than it fixes if this 'ideal' situation could exist unless there was only one track. Then it would create the same amount that is fixes--one or none. In every modern recording I know of, the mix would have to be trashed and the mixer would have to fix it if it were to be fixed 'ideally'.
Glad you brought this one up b/c I already did on the last page. If HRTF compensation needs to be applied, it would be to the newly introduced center channel that wasn't intended in the mix, and in the opposite fashion. Not to your stereo speakers like your theory. Read my post on page 4 I where I bring this up:
It could(and probably should) keep it more tonally correct to the recording. That would make sense. I don't know if it is being applied or not in Dolby or whatever processing is in use. If there is a time and place to use it during in home playback, it would be when this process is in use. My stereo recordings I listen to in stereo. My mono recordings have been manipulated in stereo to sound good in stereo so I listen to them the same way--in stereo and they sound good.me said:
Does that clear it up? Stereo recordings and playback have enough problems. I propose we try not to invent more. It was a cool and high minded thought when you came up with it. It just wasn't really the case. I'll gladly eat my words in public if a logical refutation occurs.
cheers,
Dan
Last edited:
I have read so many times that a good monitor starts with a flat on axis response .......... I also read from a respected source who was comparing how media for film has a standard reference for mixers and music audio does not , that flat up to 2khz , then a 3db slope down from there is prefered .
The thing is that the ear seems to like the pink noise slope . I have done FFT anaylsis on a cross section of pop songs and save a few exceptions ( ususally newer loudness war artrocities), the wave file ( before it ever hits the speakers ) has the slope in it .
I think you are confusing anechoic measurements to in room measurements.
How did you conduct this analysis and can you show it?
Dan
Music recording is nearly a free for all. Although it has its yin and yang aspects, it's hardly zen. To me, only the mixing and mastering environments should standardized. Then you could have a set in stone goal for your own playback. I'd even argue that you should get a headphone and car mix with every recording you buy. Right now I'm trying to make recordings that work well with headphones and stereo, but you really can't serve 2 masters absolutely. Compromises are definitely made.
You may like this one as well and it is free:
http://human-factors.arc.nasa.gov/publibrary/Begault_2000_3d_Sound_Multimedia.pdf
Dan
You may like this one as well and it is free:
http://human-factors.arc.nasa.gov/publibrary/Begault_2000_3d_Sound_Multimedia.pdf
Dan
I have thought this for a while, do the decent mix and then eff it up for the car friendly radio mix after!
Offer both (same price) or as you say even supply both!
You da man Dan!
Have another
Thanks Rob! It is Friday and I'm ready to relax with a couple luke warm ones.
I wish I had more time for recording. There are reasons why professional ones are not one man shows like mine. I now try to fudge my recordings in ways that are not too bad either way and enjoyable every way--like make some parts binaural, some parts stereo, and some parts mono. Then I'll try to make them sound like a cohesive group playing in a real, but enhanced, room together. They seem to be fairly well reviewed by the folks on recording forums, but I'm no pro by any stretch. Some of those guys can hear the proverbial gnat farts even through a Bose system in a cave I swear. I leave my playing/singing and even wording mistakes on them as well for the most part. It takes a lot of time to make better recordings! After a few , they sound better anyway. Glad we got that emoticon.
I wish there was a standard from headphone response. It's certainly an obstacle for HRTF compensation on the master track.
Oh well,
Dan
I wish I had more time for recording. There are reasons why professional ones are not one man shows like mine. I now try to fudge my recordings in ways that are not too bad either way and enjoyable every way--like make some parts binaural, some parts stereo, and some parts mono. Then I'll try to make them sound like a cohesive group playing in a real, but enhanced, room together. They seem to be fairly well reviewed by the folks on recording forums, but I'm no pro by any stretch. Some of those guys can hear the proverbial gnat farts even through a Bose system in a cave I swear. I leave my playing/singing and even wording
mistakes on them as well for the most part. It takes a lot of time to make better recordings! After a few , they sound better anyway. Glad we got that emoticon.I wish there was a standard from headphone response. It's certainly an obstacle for HRTF compensation on the master track.
Oh well,
Dan
Not sure I follow you there ; Anechoic measurements for testing speaker build and performance only , never to mix in . Dead rooms are horrible to mix in . Ethan W. (Real Traps) has designed his bass traps to not over dampen the mids and highs . If you want full spectrum absorbion , you have to specify it and buy a seperate product .Most mixers are doing it near field and the ratio of direct signal
to room signal is going to mean that flat response from the monitor would make it possible for them to make ruler flat mixes if they wanted too .
recordings that had a ruler flat spectral responses would be fatuge central .
Bob Hodas - Optimizing The Studio Environment
( see"The "Flat Response Myth")
Irregardless of different mix environments , Mastering Engineers would'nt last long if they E.Q.'d the product to be flat all the way to 20khz either. It's not the playback speakers that cause the slope . It's built into the recordings before playback .
There are plenty of free FFT programs that will do an average of a wav file for you .... Head on over to KVR and track one down .Test some of your music collection and tell me how many don't have some attenuation or downward slope in the high end of the spectrum . I have only seen a small percentage of my CD collections songs that are flat way out there . ( although I will stippulate that I'm Old school so much of my stuff was origionally done with analog which may have an influence)
The funny thing is as an old guy I'm supposed to lose high end response in my hearing , but I don't like allot of the recordings that were made during the transition between analog and digital because I find them to be strident .( nails on a chalk board and what not !)
There are many variations in mix environments and equipment , but mix engineers keep doing the same thing ... can't all be because of the speakers they are using ... and of course there are also genre stylistic factors also.
See
"Tech Tip: EQ Curves and Musical Style"
Tech Tips: Tech Tip: EQ Curves and Musical Style | Musician's Friend
Last edited:
I agree that there does seem to be a lack of standardation in mixing and mastering ( especially in the last few decades !)
I guess this thread"'Flat' is not correct for a stereo system ?"
http://www.diyaudio.com/forums/multi-way/172806-flat-not-correct-stereo-system.html
covered a wee bit of this ground before huh !!!!!
I guess this thread"'Flat' is not correct for a stereo system ?"
http://www.diyaudio.com/forums/multi-way/172806-flat-not-correct-stereo-system.html
covered a wee bit of this ground before huh !!!!!
Thanks for the compliments (I think?) but first let me clear up a Donald Rumsfeldism in your response, in that "you don't know what I don't know".
FWIW, I posted about the audio circle of confusion (of course it wasn't called that then) almost 20 yrs ago on rec.audio, before it splintered. I was refuting the concept of an "absolute sound" and back then made the argument that any true "absolute" required the playback environment to match the recording one, and worse, that since we don't know in playback what the recording was targeting as a playback environment, it left a lot of latitude for tweaking playback response. But we've been through this together in the past. FWIW, back then, people thought this was nuts, but time marches on.
So, I think we can agree that those making the recordings are doing so from various environments and this is one reason recordings are not a perfect representation of "natural" without that environment as playback.
I also believe that the full HRTF is not in need of compensation. HRTF is a loaded term with in-built phase response, spectral notches etc, and wide variance over angle and azimuth and significant variance from person to person. For context on my position in this mini-debate, I worked on HRTF in the 90s: measured several subjects HRTFs in an anechoic chamber, and designed DSP algos for conference terminals which used HRTF processing to provide better intelligibility. So, I'm definitely NOT arguing for HRTF crosstalk cancellation in speakers. However, if you've heard a good one, its provides stunning imaging (aside: I was in the Lipshitz and Vanderkooy labs at the time they were working on this (look up Kevin Kotorynski) and heard it first hand)
I think we can also agree that a phantom image played from +/- 30 deg will sound tonally different than a source of the same power spectrum physically located directly ahead. Its this tonal effect that I'm focusing on, not full HRTF compensation.
Where we seem to disagree is that you seem to be arguing that the error in central image tonal balance from this is largely eq'ed out of the recording unconsciously as a natural artifact of the recording process. I think you argue that the natural recording process takes care of it so things sound "good", so not to worry. Does this sum it up? I think I have your position correct based on your comment that center channels should be eq corrected when playing material recorded using 2-ch stereo.
If so, my point is, well, maybe. I believe neither of us can prove it either way. But I can look at the fact that it's not thought about consciously and surmise that maybe it's often (if not universally) an issue. And I think we agree that I don't think your refs for HRTF compensation support the idea that it is thought of consciously since HRTF comp is something related but quite different. HRTF comp is far more aggressive and more about providing localization cues than adjusting tonal balance (actually, IME, full on HRTF comp in speakers introduces gross tonal distortions most of the time).
Also, take the case of with minimally mic'ed classical, such as the very popular Mercury Living Presence recordings. How would or could these have been processed to "fix" this issue? In fact, these sort of recordings are even trickier if the recorded space room contribution is significant in that the reverb will suffer even greater tonal balance error since so much is from up down and hard angles laterally (I think SL goes into this on his site).
I'm not arguing for its universality but maybe a "switch" that allows this eq to be in/out recording dependent wouldn't be such a bad idea.
Last edited:
It does sound(no pun) like you are confusing anechoic loudspeaker measurements for in room responses-or something to that effect. Every studio monitor I measured is flat. No room response should be flat. What comes through the monitor is the mixed music. The spectral content of the music is up to the person doing the mix and master.
Dan
DDF, that post I take less issue with than all your previous ones on the topic. I don't believe in the 'switch' per say, but tone controls or even EQs are fine by me. If such a switch were necessary, the Mastering or mixing facility should have done it. The likely hood of getting a switch to work on a lot of recordings seems VERY low. There are likely some recordings out there that are bad enough to require EQ after you've got your speakers and room in order to be listenable, but they are rare in my experience. I don't own one. I got rid of my EQ and don't feel the need for one now. Removing a limiter would be nice on some, but no way to do that after it's on the recording. I think a lot of the desire to EQ by the end user is to minimize the effects of excessive limiting(loudness war), but I've never got it to make a squashed recording sound good. Classical music must be a pain to reproduce for its diehard advocates. People who love this music seem to be persistently displeased with their recordings and/or playback. Classical sounds good at my place. I never thought about it sounding like the live event though. It certainly doesn't, but it doesn't make me crazy. They do put it in my mind however. I'd love to walk on Mars, but I doubt I have a chance. It's a lot easier to put a small box inside a large one than it is putting a large one inside a small one. I use this principle to add convolution reverbs to my dry recordings.
It was definitely an absolute compliment. I don't see where I've underestimated you at all. I know you know a lot. That's what I respect about you and what's flabbergasting me. It would just as easy for me to say the same--'you don't know what I know' as I give you a view of my nostrils he he. I did see what you didn't understand. Not to be a jerk, but your old position was ignoring the facts of recording and it seems you are still trying to roughly hang on to it. It's just toned down now(no pun).
A global HRTF EQ applied to a loudspeaker for a stereo in order to make it sound more natural is totally dead in the water. End of story. I wish SL wouldn't have had this idea as well recently. He brought some popularity to it and ended up getting too many wheels spinning in the wrong direction. The DIY HiFi community will be paying for this one for a while I bet though I think he's modified the Orion again since then. I could be wrong about that.
I'll go further: there is no absolute need for a mix engineer to know what HRTF is. I think they should--just part of being a pro IMO, but plenty of good ones don't seem to have a clue. Some of the best in fact never mention it at all; on line or in their books. They don't even seem to understand how it might matter for headphones. I've hashed it out with them in the past. Still, they'll make a much better recording than I will. It's an art. It can be akin to a stylized painting or nature photography. Even nature photography tries to get the best angle, the best lighting, and of the most fantastic stuff, etc...
Drunk and tired.
Dan
It was definitely an absolute compliment. I don't see where I've underestimated you at all. I know you know a lot. That's what I respect about you and what's flabbergasting me. It would just as easy for me to say the same--'you don't know what I know' as I give you a view of my nostrils he he. I did see what you didn't understand. Not to be a jerk, but your old position was ignoring the facts of recording and it seems you are still trying to roughly hang on to it. It's just toned down now(no pun).
A global HRTF EQ applied to a loudspeaker for a stereo in order to make it sound more natural is totally dead in the water. End of story. I wish SL wouldn't have had this idea as well recently. He brought some popularity to it and ended up getting too many wheels spinning in the wrong direction. The DIY HiFi community will be paying for this one for a while I bet though I think he's modified the Orion again since then. I could be wrong about that.
I'll go further: there is no absolute need for a mix engineer to know what HRTF is. I think they should--just part of being a pro IMO, but plenty of good ones don't seem to have a clue. Some of the best in fact never mention it at all; on line or in their books. They don't even seem to understand how it might matter for headphones. I've hashed it out with them in the past. Still, they'll make a much better recording than I will. It's an art. It can be akin to a stylized painting or nature photography. Even nature photography tries to get the best angle, the best lighting, and of the most fantastic stuff, etc...
Drunk and tired.
Dan
One comment on the "circle of confusion". The notion is that the monitoring room and loudspeaker create an error that the recording engineer inadvertantly EQs out. If the speaker peaks up at 8kHz then the engineer, not knowing the source of the peak, will EQ it out and put an inverse dip into the recording. The sound he hears of the monitor system is not "going to tape" so any compensation for it becomes an error in recording balance.
I think this is true but only in a broad sense and to a limited degree. First, the errors of the monitor system will be quite intricate and particular. Measure any monitor and room combination and to, say, verbally describe the curve would typically take a lot of words. There would be a number of significant features. Do we think the recording engineer effectively does an inverse EQ of every little feature, or maybe just pulls down the treble a little in response to "3 response bumps at 3.4kHz, 5.9kHz and 12.5kHz".
Secondly, I think that engineers get used to the sound of their monitoring systems and because of that have a sense of what a mix should sound like under those conditions. In my experience they will spend some time in a new environment listening to their older material and competitive material and get a sense of the character of the system. If most of the recordings they listen to "sound a little on the bright side" (or whatever the character happens to be), and assuming they don't go through the excercise of of re-EQing the system, then they tend to be careful not to create a mix that deviates from the average of what the other recordings sounds like.
I'm not saying that the sound of the monitoring system doesn't have an impact on recordings, just that the recording won't be the 100% inverse of it.
David S.
I think this is true but only in a broad sense and to a limited degree. First, the errors of the monitor system will be quite intricate and particular. Measure any monitor and room combination and to, say, verbally describe the curve would typically take a lot of words. There would be a number of significant features. Do we think the recording engineer effectively does an inverse EQ of every little feature, or maybe just pulls down the treble a little in response to "3 response bumps at 3.4kHz, 5.9kHz and 12.5kHz".
Secondly, I think that engineers get used to the sound of their monitoring systems and because of that have a sense of what a mix should sound like under those conditions. In my experience they will spend some time in a new environment listening to their older material and competitive material and get a sense of the character of the system. If most of the recordings they listen to "sound a little on the bright side" (or whatever the character happens to be), and assuming they don't go through the excercise of of re-EQing the system, then they tend to be careful not to create a mix that deviates from the average of what the other recordings sounds like.
I'm not saying that the sound of the monitoring system doesn't have an impact on recordings, just that the recording won't be the 100% inverse of it.
David S.
Don't confuse the normal spectral balance of music with flatness of frequency response of a speaker or recording chain though, the distinction is extremely important.I have read so many times that a good monitor starts with a flat on axis response .......... I also read from a respected source who was comparing how media for film has a standard reference for mixers and music audio does not , that flat up to 2khz , then a 3db slope down from there is prefered .
The thing is that the ear seems to like the pink noise slope . I have done FFT anaylsis on a cross section of pop songs and save a few exceptions ( ususally newer loudness war artrocities), the wave file ( before it ever hits the speakers ) has the slope in it .
It's true that the ear prefers the spectral balance of music to slope downwards from about 150Hz upwards, and in particular there is a lot of drop off between midrange and treble on the order of 10dB or so.
This is plainly obvious when looking at music on a real-time analyser as you have done, and I've also done many times. This is also why average tweeter power dissipation is so low (as little as 1/10th) compared to a midrange driver despite covering a similar number of octaves - the signal amplitude at these high frequencies is just a lot lower in "normal" music.
The question is, why does the ear prefer this ? The answer I think is pretty simple - because that's the way things sound in the natural world. High frequency content is naturally lower in amplitude in real instruments and common environmental sounds. In the case of instruments, treble frequencies usually come only from harmonics whose fundamentals fall in the midrange, and the fundamentals are usually stronger than the harmonics.
If you were to take a perfectly flat response microphone and record a live unamplified band playing right in front of you, you would see a very similar downwards slope in the spectrum above 150Hz, and relatively little energy in the treble, apart from a few specific instruments such as hi-hats, muted trumpets etc. In other words its not an artificial phenomena in the recording/mixing process, its just how real sounds are spectrally balanced.
The point is, the preferred spectral balance of the ear for music is unrelated to the need for flatness of response of a recording and playback chain. Because we like this downwards slope does not mean that flat response speakers are incorrect, because the downwards slope already exists in the music, even before mixing and mastering is done, and although the mixing process may alter it somewhat, a downwards trend will still be there.
If the speakers deliberately introduced this slope we would then be doubling the total slope, which is obviously wrong. Speakers should be flat to preserve the pre-existing downwards spectral slope of the music, not create it.
Last edited:
You often hear them speak of using decidedly and purposely inaccurate monitoring environment that allows listening further "into" the recording (heightened detail, or exaggerated frequency ranges). They can't and don't eq these for "natural" sound on their end, but try to mentally account for the tonal distortion of their system in the job.
I guess I'm just not sold on Dan's more absolutist position that, for the recordings that care to do so, they universally unravel all these different effects properly and fully account for them in the recording process. I think its possible (but very difficult) on multi-miced close miced recording, not so on more minimalist classical in natural spaces due to the recording technique "locking in" the tonal distortion inherently.
Dan, I think that's as close as we'll come to agree to disagree on this one. I've highlighted the effects, and your response is unconvincingly (to me) "trust me, they deal with it without even knowing what it is". I think that's sometimes true, maybe often true, but logically can't see how it is true so universally or can even be physically accomplished for minimally miced classical (or jazz as well).
Last edited:
This is true.
The problem doesn't exist in the first place. The mic and position make it unnecessary. You cannot hear the instrument the way it would naturally sound no matter what. HRTF problem then cannot be an issue.
It's not even a problem to begin with. If it's 'often true' then you are admitting your curve cannot work anyway, but since it isn't a problem to begin with it's a moot point.
I'm glad that your position has bent, but it is really broken. Hopefully everyone else gets it. If not I'd gladly answer any sticking points.
still simple(me and the facts),
Dan
Very nice summation of the situation on your part ! Agreed.
I would only add that room effects are obviously a bit more important than you expressed in that summation important ( a tiled bathroom verses a heavily carpeted and draped room ) Not many mixers try to make an end result that sounds like one of those environments ,regardless of genre; so that makes the slope an even more prominent component in most all recordings .
Although I can't remember ever advocating changing the design parameters of monitors away from a flat spectral response , everyone around here keeps telling me I'm confused about something !
Interesting.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.