On the limits of "Hi-Fi" intent

[bwaslo]

Earl, shouldn't the tapering off in the highs that you advocate also be "corrected in the mix"? Why this but not center image timbre correction?

 

[gedlee]

My takeaway is that recordings should be “objectively accurate”, but that includes appropriate compensation for microphone response and placement. Speakers, on the other hand, should be “voiced” to suit the listening room (almost certainly NOT “objectively flat”) so as to produce a “subjectively accurate” replication of the original performance when presented with an “objectively accurate” recording. “Studio produced” recordings should take their que from the movie industry and be mixed to sound the same over a range of properly set up reproduction environments.

Since virtually all of my listening is on studio work, and I completely agree that music on movies usually sound better than other sources, I can completely agree with your last line above. But then doesn't that contradict the lines above it? You can't have both as I see it. And yes, it is becoming clear to me that people that are looking for that "realistic" venue sound might well have to tweak their system to get the best match between the recording and the room. But that is not my first aim. It is precise detail of the studio recording that I am after and I think that this is best done through standardization and quantification of sound sources and rooms, and playback source and rooms. There is little room there for subjective opinions.

And that’s where I take (some) exception to Earl’s position (as I understand it). My listening room is not voiced “flat”. It is not “objectively accurate”. It is, however, voiced to sound “right” with “objectively accurate” recordings . . .

I am not sure what a "voiced" listening room is, that's a new concept to me, so I don't know if I agree or disagree. To me the room does need to be "designed" for optimum results, the speakers are not fitted to the room, the "system" is designed as a whole. I understand that this is not always possible, but to me it most certainly is the goal. If the most highly optimized system that you can achieve - room and all - is not the goal then, of course our solutions will differ.

"voiced to sound “right”" is kind of the same thing as saying "If it sounds good to me then it is "right."" isn't it? Yea, not my approach.

 

[Lynn Olson]

Just my own experience: I have made numerous “field” recordings of the small orchestra I used to work with, typically two channel ORTF placed about 6 ft. above and 6-10 ft. behind the conductor. I also have a number of “house mic” recordings of the same orchestra in different halls, some spaced pair somewhere between the proscenium and first row, or in one case a Decca Tree just above the conductor. These placements in all cases are over-bright, and impart a harsher sound to the strings than what is heard back in the hall . . . very much that early “DG sound”, but also (with a relatively moderate high end roll off) reasonably accurate to the sound in “front row” seating (as is the “stereo perspective”). Those recordings, while “objectively accurate”, are when played in my listening room (16x24, with relatively “normal” living room acoustics”) way too “bright”, and I correct with a “house curve” to get to something approximating the “hall sound” of the original performance venue. “Objectively flat” from beginning to end just sounds wrong.

“Spatial perspective” is also “wrong” with these recordings, giving more “separation” and easier “localization” than one actually hears out in the hall. I find that full-range dipoles provide some degree of correction for this in my listening room, and produce a more “lifelike” acoustic image, especially with the ORTF recordings. It’s not “perfect”, but it’s pretty good.

What one finds in “commercial” recordings is a mixed bag. In general what I hear is that the less “processing” is done in production the better the recording sounds in my listening room . . . rather obviously because my system is “tuned” for “objectively flat” source material. A reasonable amount of correction for microphone placement (tempering those shrill strings) is best done at the source. Any “correction” for the acoustics of the “average listening room” is bound to be wrong, and to mess with the room corrections that I apply. “Studio mix” recordings are all over the ballpark, but the well done ones can be excellent. Movies are (almost) uniformly good (there’s the benefit of “standards”), since they assume that “house correction” will be done “at the house” (where it should be).

My takeaway is that recordings should be “objectively accurate”, but that includes appropriate compensation for microphone response and placement. Speakers, on the other hand, should be “voiced” to suit the listening room (almost certainly NOT “objectively flat”) so as to produce a “subjectively accurate” replication of the original performance when presented with an “objectively accurate” recording. “Studio produced” recordings should take their cue from the movie industry and be mixed to sound the same over a range of properly set up reproduction environments.

And that’s where I take (some) exception to Earl’s position (as I understand it). My listening room is not voiced “flat”. It is not “objectively accurate”. It is, however, voiced to sound “right” with “objectively accurate” recordings . . .

Well stated. Dewardh has the luxury of making his own recordings with a group he knows well, in a performing space that's also known well. So the task for the playback system is to give the most plausible re-creation of the original performance. This is very similar to the BBC-monitor design brief, where the sound in the control room has to match what's happening on stage a few feet away.

Good loudspeaker designs have converged on smooth response over the listening arc, with a slight tilt towards the bass region. This is not that hard to do with modern drivers and modern crossover designs. It was hard in the Seventies, with sketchy drivers and computer crossover simulations only available to companies like KEF and B&W. Much easier now, and drivers are better too, so there's really no excuse for loudspeakers with erratic response curves.

Unfortunately, many high-end magazine reviewers like a wacko response curve with "slam" in the bass and sizzle in the treble, so there are lots of high-end speakers that cater to this ... based on what I heard at the last RMAF, probably even the majority. For better or worse, the reviews tend to drive the market, since all it takes is one or two negative reviews to bankrupt a small manufacturer.

 

[gedlee]

Earl, shouldn't the tapering off in the highs that you advocate also be "corrected in the mix"? Why this but not center image timbre correction?

This question was asked before. If the mix is done under a house curve then the playback should be also.

My take on why there even is a house curve has to do with the different tonal nature of a mic close to the source and the same sources sound at some much more common listening distance. Because of air absorption the two are not the same, the distant sound is duller - part of what makes it sound distant. Too "bright" and the sound seems like it is in your face because that what it actually would sound like if it were in your face.

 

[gedlee]

Good loudspeaker designs have converged on smooth response over the listening arc, with a slight tilt towards the bass region. This is not that hard to do with modern drivers and modern crossover designs.

Unfortunately, many high-end magazine reviewers like a wacko response curve with "slam" in the bass and sizzle in the treble, so there are lots of high-end speakers that cater to this ... based on what I heard at the last RMAF, probably even the majority.

Hi Lynn - don't you think that there is a reason for this? I do. I think that it not only makes perfect sense, but sounds best as well. And doesn't that just follow from what I am saying, that accurate leads to preference. Why hasn't it been this way? That's easy to see as well - its boring (not the sound, the idea)and has no marketing hooks to hang a jingle on. You can't carve out a niche if everything converged on one design. But I think that convergence is inevitable none the less.

As one marketer put it to me: "Its the sizzle that sells the steak!"

 

[billshurv]

Since virtually all of my listening is on studio work, and I completely agree that music on movies usually sound better than other sources,.

This does raise an interesting point. A lot of effort is put into movie sound (music, effects, speech) and placing it in a sound field. It is contrived, but there are at least some rules (THX etc) that are followed compared with 2-channel stereo. So it is easier to build something that will give the correct reproduction. For an orchestral recording there are no rules.

 

[gedlee]

Movie sound stages operate under very strict rules. For CDs there are no rules or standards things are all over the map. In this scenario there probably is no "one way". But the fault lies with the lack of standardization. Standardize the audio recording world and many of these issues would go away and I think that we would converge on a common understanding and approach - a scientific one.

That said there are good and bad implementers, i.e. producers. To me it makes no sense to tailor things to the bad ones. The good ones do a better job.

 

[billshurv]

Well stated. Dewardh has the luxury of making his own recordings with a group he knows well, in a performing space that's also known well. So the task for the playback system is to give the most plausible re-creation of the original performance. This is very similar to the BBC-monitor design brief, where the sound in the control room has to match what's happening on stage a few feet away.

Which BBC design brief? Looking at the LS5/8 for example for which the original BBC documentation is online http://downloads.bbc.co.uk/rd/pubs/reports/1979-22.pdf There are 3 key requirements

1.The ability to generate high sound
pressure levels (at least 110 dB(A) at lm
distance).
2.Low coloration,* particularly at mid-frequencies to enable the monitor to be used for balancing serious music.
3.The ability to produce a sharply defined stereo image


I am still confused over one point referenced on the classical recordings. Microphones 10 feet in the air will not 'hear' what someone in a seat will. So surely all classical recordings will be too bright unless equalised in production. If you have to EQ at home to get the '5th row' experience (rather than a preferred seat somewhere else) then there is a fault in the recording process?

 

[billshurv]

Movie sound stages operate under very strict rules. For CDs there are no rules or standards things are all over the map.

If you take that sentence into account then I think everyone is close to violent agreement.

 

[Cask05]

This does raise an interesting point. A lot of effort is put into movie sound (music, effects, speech) and placing it in a sound field. It is contrived, but there are at least some rules (THX etc) that are followed compared with 2-channel stereo. So it is easier to build something that will give the correct reproduction. For an orchestral recording there are no rules.

This argument is beginning to look like the reality with which I see and deal with daily.

I believe that a few responses here on what is in their music is much too trusting of the practices used and motivating factors. I really recommend doing your own analysis and reconstruction of perhaps a couple of hundred stereo tracks, especially orchestral recordings. I believe that what you find will shake your faith more than a little in the heritage-RIAA music products. This will also blow holes in many firmly held beliefs that I see here on the objectiveness of the current state of the art in stereo reproduction.

The THX crowd (cinema) has apparently avoided that mine field entirely by focusing on the cinema experience starting with a much higher sound reproduction standard than the lowest-common-denominator stereo sound systems such as earbuds on an iPod playing lossy music downloads.

Chris

 

[billshurv]

T I really recommend doing your own analysis and reconstruction of perhaps a couple of hundred stereo tracks, especially orchestral recordings. I believe that what you find will shake your faith more than a little in the heritage-RIAA music products.

I tried if you remember. I did not come to the same conclusions as you and in fact it seems that modern recordings are doing a pretty good job. going back in time to the so called golden age there are some nasty compromises I agree.

 

[Cask05]

I tried if you remember. I did not come to the same conclusions as you and in fact it seems that modern recordings are doing a pretty good job.

Since you courteously brought this up:

Does your system faithfully reproduce at least 20 Hz - 20 kHz at something approaching a THX HT standard levels with low distortion (mostly modulation distortion sidebands during music transients)?

The differences in concert-level playback of orchestral material--even stereo only--is really big, so much so that I can't simply agree that you might be right for courtesy's sake.

In order to hear the differences in the orchestral recordings that you showed me, I believe that your system that you described appears to be insufficient in that regard. I could be wrong in interpreting what you said in that thread.

YMMV.

Chris

 

[billshurv]

Since you courteously brought this up:

Does your system faithfully reproduce at least 20 Hz - 20 kHz at something approaching a THX HT standard levels with low distortion (mostly modulation distortion sidebands during music transients)?

Not yet, working on it. Things like life, kids and money make it slow progress.

The differences in concert-level playback of orchestral material--even stereo only--is really big, so much so that I can't simply agree that you might be right for courtesy's sake.

Not sure that has anything to do with remastering the FR?

In order to hear the differences in the orchestral recordings that you showed me, I believe that your system that you described appears to be insufficient in that regard. I could be wrong in interpreting what you said in that thread.

YMMV.

Chris

It is currently insufficient for FR and max level (although I can only afford cheap seats at concerts so 105dB would do me nicely). I accept that and at some stage in the next decade will complete the rebuild to get me in the ballpark. BUT I'm still not convinced that remastering classical recordings to meet a particular FR curve is correct without knowing the particular peformance of the hall it was recorded in. I may be completely wrong and there was a 30 year industry wide conspiracy and would be interested to find out more. But without measuring the venues I am not sure how to resolve this.

 

[Cask05]

Thanks for your measured response. I don't like to have those type of discussions. It doesn't make me feel any better about saying them.

I have to say that one only needs well-calibrated ears from playing in full orchestras and wind symphonies for a number of years to know when the stereo orchestral music tracks as distributed don't sound anything like the real thing. The only thing that sounds realistic are the orchestral tracks played back on THX or other multichannel compliant discs that feature separate LFE channels. such that the mastering guys don't just apply their stereo bass-hobbling EQ to seriously cut the LFs. Additionally, I've found that all frequencies between about 400 Hz and 100 Hz are also similarly attenuated via not-so-gentle slope high pass filter.

It's those frequencies, played cleanly and in balance with all the other frequency bands, that most strongly recreate the listening experience that I've described. In many instances, HF attenuation above 400-800 Hz is applied to stereo orchestral tracks.

The music winds up sounding "distant" and without any visceral feeling in those instances--which includes most of the stereo classical recordings that I've dealt with. This is the problem that I was describing above.

I find it difficult to read about the assumptions of absoluteness of the source material quality in reference to the subject of this thread and integrate that with my experiences.

YMMV.

Chris

 

[jcx]

do we really want "in the orchestra perspective" - you should be wearing hearing protection there

 

[billshurv]

But surely as a player you get a completely different perspective to the audience?

EDIT: JCX beat me to it.

 

[dewardh]

I am not sure what a "voiced" listening room is, that's a new concept to me,

I suspect it's not "new to you" except as the phrase I use to describe it. What I mean is addressing the combination of loudspeaker frequency response and polar and the room absorption and reflection curves and the artificial and wrong overlay of "small room acoustics" on the recorded material. It's all a big interactive mess, and cannot be effectively addressed at the speaker alone.

Small room acoustics, especially early reflections but including also overall high frequency reverberation, simply don't exist in a concert hall. When the boundary reflections in the listening room account for much (sometimes most) of the sound you hear it is of course necessary to take them into some sort of account. Proper integration of reflected sound can restore a "spaciousness" that's lost with two-speaker "stereo", but that comes at a price as the acoustic signature of the room changes the tonal balance of the sound. The back and forth process of adjusting speaker response (and polar pattern if you DIY), absorbers and diffusers, speaker and listener placement . . . all of that is what I refer to as "voicing the listening room". The point of it all (for me) is to produce a listening environment that allows "objectively accurate" recordings to sound right.

 

[dewardh]

It is currently insufficient for FR and max level (although I can only afford cheap seats at concerts so 105dB would do me nicely).

Although it obviously can be different at amplified venues I have NEVER seen 105dB at an audience seating position in an orchestra concert hall. There are few places on stage that reach that level either (maybe right in front of the trombones). Like John Eargle I've never seen 105dB at an ORTF or "house" microphone in such a hall either. Likewise you will never see over 105dB in a properly set up movie theater (and still they commonly sound "too loud").

It is not at all difficult to produce realistic concert hall levels in a "home" environment with relatively modest speakers and amplifiers. It's different, of course, if you want to burst your eardrums with rock and roll . . .

 

[billshurv]

There is evidence I found in a thread on this site that shows this has been measured in a concert hall in a good seat on peaks. I can only afford cheap seats or promenading so have no reference. I also forget to sneak in my SPL meter.

I would be interested if you have a reference to what JE thinks is an acceptable max SPL for orchestral peaks.

 

[Pano]

It's all fine and good to talk about standards. The music recording world should have them. Standards work pretty well in the cinema world.*

But then what? I have to throw out all my old recordings because they don't fit the standard? I can't listen to anything recorded in the 20th century?

There are some current artists that I really like, but I have a difficult time listening to their recordings because of the loudness wars. The music is nice, but the recordings are too painful to hear. The use of modern methods does not always yield great results.


*speaking of which - anyone ever notice how darn good old American TV shows sound? I don't mean Hi-Fi, I mean the unerringly good intelligibility of the dialog. Glen-Glenn sound did phenomenal work.