Coloration for Classical Orchestra?

[3rd&4thT]

Originally Posted by khai

This recording by Richter is from a box set. If my memory serves me right, all preludes and fugues were recorded in the same room. But they all sound very different. Some sound more intimate (closer to the piano) with less reverberation than others.

These recordings date from 1972/3, and were made in a castle in Salzburg. The original recordings were roasted by the critics for what they described as "gymnasium" or "swimming pool" acoustics. The CD transfers are on RCA Gold Seal.

Originally Posted by khai

The same prelude recorded by Glenn Gould in the 60's sounds totally different. Very little reverberation comparatively, a closer miking with far more details in the high frequencies. I prefer Gould's recordings in general but not always. Gould didn't like reverberation much. He had his mics placed closer to the strings than most classical players as he liked to hear the attack more. I prefer Gould's piano sound in general for Well tempered clavier, but often more reverb and distance sounds more nicer like on this Richter example.

Producer Andy Kazdin was on a desperate quest to get the piano microphones away from Gould's mouth. Columbia's usual piano pickup of four mics across the soundstage would have been unlistenable with Gould's loud baritone obbligato.

Originally Posted by khai

Another question is recording organs in cathedrals or chapels. In the early 60's Gould was criticized harshly for miking his organ (on the "Art of the fugue") with mics right against the opening of the pipes. I think Gould's choice was right. I've always had a problem with church organ recordings. I hear the room far too much. It's a soup.

That's very much a question of personal taste. I think anyone composing for the organ, from Bach on down, is aware that church organs resonate in space. Small church or large, you hear more reflected sound than direct. I dislike closely-mic'd stereo organ recordings because the music hits you in the face and you get really clear clanking and whooshing mechanical noises. Sometimes soup is good for you.

After hearing Bach played on a church organ from the catwalks inside the ranks of pipes, I think the most purely exciting organ recordings would be made in surround, with the different pipe choirs on all sides. Not realistic maybe, but you'd hear all the inner voices clearly without spotlighting or congestion from the limitations of stereo. However the critics would have kittens just at the thought of it, and I'm not expecting anyone to try it anytime soon.

3rd&4thT

[3rd&4thT]

Originally Posted by andy_simpson

However, the attached mp3 was made with a pro level signal chain (upgrade & recording report to follow shortly) and is quite likely to pass a blind test in the room (even the mp3 version).

You should be quite proud.

3rd&4thT

[andy_simpson]

Originally Posted by 3rd&4thT

You should be quite proud.

3rd&4thT

Yes, I'm proud, but I'm open to constructive criticism or discussion.

Did you listen?

Andy

[3rd&4thT]

Originally Posted by andy_simpson

Yes, I'm proud, but I'm open to constructive criticism or discussion.

Did you listen?

Andy

With pleasure. I love the way you've captured the room, I love the stability of the image (nobody on a skate board here), and I've never heard more realistic lower brass, which is usually a problem.

However this thread has wandered all over the map, and since my questions for you are not about coloration, I think I'd prefer to PM you for further discussion instead of wandering off again here.

Best regards,
3rd&4thT

[andy_simpson]

Originally Posted by 3rd&4thT

With pleasure. I love the way you've captured the room, I love the stability of the image (nobody on a skate board here), and I've never heard more realistic lower brass, which is usually a problem.

However this thread has wandered all over the map, and since my questions for you are not about coloration...

Thanks for listening - this at least can frame my opinions/work in a practical perspective.

In order to get the thread back on track somewhat let's get back to the vintage mics, which are for good reason still popular.

In further consideration of the role that pick-up pattern has to play - how it shapes the auditory masking 'scene' - I have begun to wonder if this does not play a role in the popularity of the LDC (especially, for example, the u47/u67).

It is quite well known that the typical LDC becomes more directional at HF and this, combined with the interacting in-band self-resonance, can perhaps give a reason as to why these microphones are famous for 'depth'.

Also, in terms of auditory masking, we can imagine how even the mono recordings of long ago could potentially have more 'depth' than a modern stereo recording made with microphones which offer a less ideal auditory masking scene.

Andy

[3rd&4thT]

Originally Posted by andy_simpson

It is quite well known that the typical LDC becomes more directional at HF and this, combined with the interacting in-band self-resonance, can perhaps give a reason as to why these microphones are famous for 'depth'.

Also, in terms of auditory masking, we can imagine how even the mono recordings of long ago could potentially have more 'depth' than a modern stereo recording made with microphones which offer a less ideal auditory masking scene.

Andy

I'm sure that's right. You get the best of both patterns, the detail and presence of the cardioid and the hall resonance and warmth of the omni. I think the Decca Tree works at its best with the mic behavior you describe.

3rd&4thT

[aracu]

The 60's Stravinsky mp3 is interesting. It sounds a bit like a pop recording
rather than a real space, accentuating the peculiar anxious vibe of the music.
The placement of guitar and harp at widely opposite sides sounds artificial, symetrical and effective. The musicianship and engineering would sound good with new or old technology.

[Corran]

I really like the Stravinsky recording, but at the same time it doesn't have any real "hall" feel to it, which is okay, but it seems a bit dry. Frankly though it sounds more like really good digital than anything distinctly analog.

I think the basic question this thread started on - is coloration okay for orchestra? - boils down to, what mics and preamps you use and how they respond. I believe that none of your kit should "change" the sound of the orchestra (say, mid-forward). But if a mic or preamp adds something "special" to a recording, well, go for it. You'll know it when you hear it, right? But at the same time it shouldn't jump out at you immediately...

[3rd&4thT]

Originally Posted by aracu

The 60's Stravinsky mp3 is interesting. It sounds a bit like a pop recording
rather than a real space, accentuating the peculiar anxious vibe of the music.
The placement of guitar and harp at widely opposite sides sounds artificial, symetrical and effective.

That was always the rap on Columbia's classical recordings of the period. One of their major producers acknowledged recently that this kind of sound was house policy, and today he would have worked with a lot fewer mics.

BTW, there is film of this recording session included in a Canadian television documentary about Stravinsky that surfaces occasionally on TV, in the US on the Ovation Network. It's available on DVD on the VAI label.

3rd&4thT

[aracu]

"I think the basic question this thread started on - is coloration okay for orchestra? - boils down to, what mics and preamps you use and how they respond." -Corran

When designers strive for mics, pres etc. to reproduce reality as closely as possible, it can certainly result in very high quality and usefull gear. In a way though it's impossible to not add any "colour" to a recording. The goal is
to help make the musical idea and performance work as electronic media.
All recorded acoustic music becomes electro-acoustic music. Each application
is unique and can break established rules if necessary, if it is to be part of a creative process.

"Artificial" and "real" each have worlds of meaning and application.
Neither is necessarily positive or negative.

[andy_simpson]

Originally Posted by Corran

I really like the Stravinsky recording, but at the same time it doesn't have any real "hall" feel to it, which is okay, but it seems a bit dry. Frankly though it sounds more like really good digital than anything distinctly analog.

The question is not whether it sounds 'analogue', rather whether it sounds like the orchestra.

This is why it is important to monitor in the room. 'Analogue sounding' would not be a compliment when comparing to the real orchestra!

I think the basic question this thread started on - is coloration okay for orchestra? - boils down to, what mics and preamps you use and how they respond. I believe that none of your kit should "change" the sound of the orchestra (say, mid-forward). But if a mic or preamp adds something "special" to a recording, well, go for it. You'll know it when you hear it, right? But at the same time it shouldn't jump out at you immediately...

The word 'hype' came into popular use to describe those pieces of equipment that add 'something special' at low listening levels, which is exposed as 'hype' at higher or realistic listening levels. !equal loudness contours !
As an orchestral player you well know that the sound of a real orchestra is special!

If your recording can improve on that - in the room - you are doing very well!

Andy

[loranoyd]

Hi, everyone.

I've been reading this thread with great interest as an amateur recordist who works with classical music exclusively. I have been using digital equipment from the beginning, and I have never worked with nor owned tape or tube equipment, so I find the posts by those of you who have used these various tools either exclusively or in combination to be extremely enlightening. I have certainly encountered (and own) recordings both "ancient" and modern, which are the best/worst I've ever heard. There seems to be no guarantee when it comes to "all digital" or "all analogue/tube" signal path.

Without intending to derail this thread again, I am really curious about the signal path and processing Andy Simpson uses in the excerpt from the third and fourth movements of the Tchaikovsky Symphony no. 4 he posted earlier in this series. I ask because I happened to have recorded the same work in May 2007 for the orchestra I play with in Lansdowne, Pa. Several people have commented on the excerpt in the most positive manner, so it is with some hesitation and trepidation that I write about my own experience here. If the excerpt was intended to demonstrate how closely a recording can match the actual sound of the instruments "live," then I am really confused. To my ears, the strings sounded as if they were in another recording session: distant and indistinct. The lower brass and percussion, on the other hand, sounded "hyped:" too brassy (!!) in a way that seemed to emphasize their upper frequencies at the expense of a fuller, more complete instrument. The balance overall sounds way off, as if I'm hearing the orchestra from a position somewhere behind the brass and percussion, instead of from a position somewhere in a hall with the strings in front of me. Can you tell that I'm a string player? So perhaps this is my bias. I've spent the past twenty years living just outside of Philadelphia, and I've had the pleasure of working with some of the older musicians from the Philadelphia Orchestra, who are steeped in the culture of the Philadelphia string sound.

In any event, I'm very curious to learn from Andy how his recording came to be: signal chain and mixing/mastering, if he wouldn't mind talking about it. I'm also curious to discover why my experience of the excerpt is so very different from those of the others on this list. I'm willing to entertain the notion that my ears are hopelessly corrupted. I tried to account for weaknesses in my reproduction chain by listening to the excerpt not only using good Sennheiser headphones, but I've also played it on decent computer speakers and finally routed the sound out to my home stereo (Adcom preamp and amp into Snell tower speakers). I hear the same issues I've described above. Please enlighten me.

And in the spirit of "putting my money where my mouth is," I humbly offer an excerpt from my orchestra's performance of the Tchaikovsky symphony. I've taken the same passage as Andy's. A brief word in my own defense: the final recording is a "rescue" job. I had completely given up all hope of being able to offer this recording to orchestra members because the venue SUCKED!!!! It was not our usual concert hall, but rather a middle school auditorium more suited to assemblies than concerts of classical music. The space was deader than dead, and the HVAC for the hall was suspended above the orchestra's head just off in the wings stage right. The salvage job was made possible by the purchase of Altiverb. I guess this means I'm in favor of "coloration" in classical music recordings. Anyway, my signal chain is as follows: a pair of Beyerdynamic ribbon mics in Blumlein for the basic stereo image; one Neumann KM184 mounted on the same bar as the Blumlein pair to extend my reach into the woodwind section; two Neumann TLM103s as flanking mics. All mics located about 15 feet from the stage, all in a line, about 10 feet high. The Beyers were fed into two channels of John Hardy M1 preamp. The TLM 103s went to two channels of FMR audio's RNP. The KM184 was fed into one of the onboard pres of a MOTU 896HD, which was also the recorder. I used Digital Performer running on an "ancient" Mac PowerBook G4. Apologies for the long note. I am looking forward to continuing the discussion.

Regards,

Lloyd

[3rd&4thT]

Lloyd,

I like both recordings for entirely different reasons. They are what they are. Neither is perfect, but both have virtues.

To start off, both were made in dry, crappy halls. To help make the comparison more direct, do you have a version of this recording you can post without the Altiverb? Reverb covers a multitude of sins, and Andy used none.

And can you tell us if the dynamics were altered in any way? Was there compression or limiting? Was there any rebalancing during or after the recording? It's far easier to get a full-bodied string tone if you can subdue the sforzandi coming from elsewhere on the platform.

To illustrate my point, Antal Dorati recorded Stravinsky's Firebird in stereo twice. The first time was in London for Mercury, with three mics and no gain riding. Strings are present and clear in the Mercury manner, but not predominant.

Later, Dorati did it again in Detroit for Decca, with two dozen mics and nonstop gain riding. The strings are rich, lush and never get buried no matter what. You'd hardly know it was the same conductor. (See attachments below)

I have had an interesting off-line conversation with Andy. I don't consider his version as perfect, but I will tell you that I find his closer to the sound of a small orchestra with not enough strings (unconventional imaging perhaps but in no way unnatural), and your string section more conventionally beefy because of your "up close plus reverb" soundstage. His has more of a back-of-the-hall perspective, yours much more from the front rows.

I suspect that the purpose of the two recordings is different, and therefore everything else about them is too.

3rd&4thT

[loranoyd]

Hi, 3rd&4thT

Originally Posted by 3rd&4thT

Lloyd,

I like both recordings for entirely different reasons. They are what they are. Neither is perfect, but both have virtues.

To start off, both were made in dry, crappy halls. To help make the comparison more direct, do you have a version of this recording you can post without the Altiverb? Reverb covers a multitude of sins, and Andy used none.

Ah, I didn't realize (or remember) that Andy's recording was made under less than ideal conditions in a less than ideal space. I do have the original files, and I will work to post another excerpt in the manner you describe.

Originally Posted by 3rd&4thT

And can you tell us if the dynamics were altered in any way? Was there compression or limiting? Was there any rebalancing during or after the recording? It's far easier to get a full-bodied string tone if you can subdue the sforzandi coming from elsewhere on the platform.

D'oh, I failed to describe the mastering process, didn't I? OK, then. Once gain levels on the preamps were set (modestly, to prevent any clipping), they were left untouched for the duration of the performance. In post, I typically begin with the idea that the Blumlein pair is the "core" image, the flankers and center mic are "supports," and they are brought up into the mix until my wife (the "ears" of the operation, and the one who monitors the recording on site--I'm principal violist at these concerts) feels the sound is the best it can be. Again, gain structure across the five tracks is set so that there are no "overs." No gain riding. Five channels bounced down to a stereo pair. Then over to a final "mastering" environment. Altiverb on the stereo track as an insert, Digital Performer's MasterLimiter on the master bus, with the peak set to 0.5 dB and the threshold set to 0.5. There is some light (and I mean light) gain riding by hand to bring the quietest passages up somewhat. My "clients" listen in their cars and on their iPods, and, to be completely honest, I also prefer the recordings with the quietest passages brought up. The recording is my guide here: bring the quiet passages up too far, and those passages sound miserably out of place and artificial. Ride gain too frequently and musical climaxes, for example, are immediately robbed of their power. Mess up the gain riding (in any number of ways) and the musical phrases don't make sense. I completely understand why photographs of the best engineers were often shown with scores on or near the mixing board.

[Interesting comments on Dorati excised]

Originally Posted by 3rd&4thT

I have had an interesting off-line conversation with Andy. I don't consider his version as perfect, but I will tell you that I find his closer to the sound of a small orchestra with not enough strings (unconventional imaging perhaps but in no way unnatural), and your string section more conventionally beefy because of your "up close plus reverb" soundstage. His has more of a back-of-the-hall perspective, yours much more from the front rows.

I suspect that the purpose of the two recordings is different, and therefore everything else about them is too.

3rd&4thT

I believe you are correct here, 3rd&4thT. I would only object to the implication that my string section sounds beefed up due to Altiverb. Of course, you don't have anything for comparison, so I will tell you that the Lansdowne S.O. strings sound beefed up because they ARE beefed up: typically we seat 8 stands of first violins, 8 of second violins, 5 stands of violas, 5 stands of cellos, and 5 bassists: over 50 string players. I do not have a program of that concert at hand. I suspect he added even more strings for this all-Tchaikovsky concert. Altiverb did nothing in this regard. It was used very lightly (I guess I should have included that info in the mixing/mastering description) In fact, I used only 30% of the wet signal in the final mix--just enough to put the "bloom" back into the sound. Most importantly of all: the conductor kept everyone in check. The recording shows primarily his work: he tames the brass and keeps them in line during the opening moments of the fourth movement; he keeps the strings down so delicate wind melodies can come out, etc. My point is that I don't believe I do much, if anything, in this regard, with the recorded files. I can't, really, since no microphone is so isolated that it has only one kind of sound (wind, brass, percussion, string). And, forgive me for being so bold here, but I do find it both unnatural and unbalanced to hear a recording in which the brass swamp the strings. If the conductor has a small string section, then he or she must work harder to maintain the balance within the score. I own many recordings in which the conductor employed a leaner string sound in repertoire we conventionally hear with a large string section. In these recordings, the winds and brass are still balanced against the strings. I did not hear a balance in the other recording. It is certainly unconventional, and its purpose, indeed might be different than mine. I have no way of knowing this, as Andy has not said on this thread what the intention was of his recording. I believed that the intent was to demonstrate to us that a recording can sound as accurate as the live sound. I find it hard to judge accuracy when the overall sound seems to me so unusual. No disrespect to anyone intended. I appreciate you taking the time to write, and I will have a listen to the Dorati examples, and post an "unaltered" version of the Tchaikovsky soon. Not today, though. Today is housecleaning and laundry day at the O.K. Corral. I hate it when the ordinary pressures of life encroach on good gearslutting!!

Regards,

Lloyd

[andy_simpson]

Originally Posted by loranoyd

Without intending to derail this thread again, I am really curious about the signal path and processing Andy Simpson uses in the excerpt from the third and fourth movements of the Tchaikovsky Symphony no. 4 he posted earlier in this series. I ask because I happened to have recorded the same work in May 2007 for the orchestra I play with in Lansdowne, Pa. Several people have commented on the excerpt in the most positive manner, so it is with some hesitation and trepidation that I write about my own experience here. If the excerpt was intended to demonstrate how closely a recording can match the actual sound of the instruments "live," then I am really confused. To my ears, the strings sounded as if they were in another recording session: distant and indistinct. The lower brass and percussion, on the other hand, sounded "hyped:" too brassy (!!) in a way that seemed to emphasize their upper frequencies at the expense of a fuller, more complete instrument. The balance overall sounds way off, as if I'm hearing the orchestra from a position somewhere behind the brass and percussion, instead of from a position somewhere in a hall with the strings in front of me. Can you tell that I'm a string player? So perhaps this is my bias. I've spent the past twenty years living just outside of Philadelphia, and I've had the pleasure of working with some of the older musicians from the Philadelphia Orchestra, who are steeped in the culture of the Philadelphia string sound.

In any event, I'm very curious to learn from Andy how his recording came to be: signal chain and mixing/mastering, if he wouldn't mind talking about it. I'm also curious to discover why my experience of the excerpt is so very different from those of the others on this list. I'm willing to entertain the notion that my ears are hopelessly corrupted. I tried to account for weaknesses in my reproduction chain by listening to the excerpt not only using good Sennheiser headphones, but I've also played it on decent computer speakers and finally routed the sound out to my home stereo (Adcom preamp and amp into Snell tower speakers). I hear the same issues I've described above. Please enlighten me.

And in the spirit of "putting my money where my mouth is," I humbly offer an excerpt from my orchestra's performance of the Tchaikovsky symphony. I've taken the same passage as Andy's. A brief word in my own defense: the final recording is a "rescue" job. I had completely given up all hope of being able to offer this recording to orchestra members because the venue SUCKED!!!! It was not our usual concert hall, but rather a middle school auditorium more suited to assemblies than concerts of classical music. The space was deader than dead, and the HVAC for the hall was suspended above the orchestra's head just off in the wings stage right. The salvage job was made possible by the purchase of Altiverb. I guess this means I'm in favor of "coloration" in classical music recordings. Anyway, my signal chain is as follows: a pair of Beyerdynamic ribbon mics in Blumlein for the basic stereo image; one Neumann KM184 mounted on the same bar as the Blumlein pair to extend my reach into the woodwind section; two Neumann TLM103s as flanking mics. All mics located about 15 feet from the stage, all in a line, about 10 feet high. The Beyers were fed into two channels of John Hardy M1 preamp. The TLM 103s went to two channels of FMR audio's RNP. The KM184 was fed into one of the onboard pres of a MOTU 896HD, which was also the recorder. I used Digital Performer running on an "ancient" Mac PowerBook G4. Apologies for the long note. I am looking forward to continuing the discussion.

Regards,

Lloyd

Hi Lloyd,

It's great to see somebody else putting something up here for a real world comparison and with the same section of the same piece this is very interesting.

I downloaded your clip & put aside a good half hour for listening & comparison.

I did listening firstly via my big mackie SRM450 workshop monitors and then via my DT100 headphones (amp was behringer 'powerplay').

As usual, since I have enough headroom in my monitors, I am able to bring up the level to that of a concert performance, which I did.

The first thing that strikes me is that there is so much compression/gain riding that I cannot balance one recording against the other without a -24dB or so on your recording (which can be more or less measured by looking at the RMS of either clip).

Taking the PP entry section, I tried to match RMS loudness, which is where the -24dB came from.

After reducing gain by 24dB for 'equal' RMS levels at the outset, the lack of dynamics is quite enormous and makes comparison very hard, excepting the measureable lack of dynamics.

For example, directly before the first explosive entry there is a PPP section. In matching PPP sections such as this we see a huge difference in dynamics.

The compression, aside from removing the entire build-up before the explosion & all the dynamic interplay of the plucked strings & orchestra, aside from removing the actual explosion itself, also removes any vibrancy from the orchestra.

If the conductor had wanted zero dynamics it would have been more effective to play with zero dynamics.

Finally, at anywhere near to realistic levels the noise-reduction (gate) is very audible and sucks the air out of the hall directly before the first explosive entry.

I hate the air-con in performance halls as much as the next man, but I would draw the line at a complete -infinity gate (or at least, if I was trying to be sneaky, I would put it before the reverb in the signal chain, so that the reverb tails masked the effect as much possible!).

Would it stand up next to the orchestra?

I don't think there is any way that a recording missing more than 20dB of dynamics would stand up next to the real orchestra.

Now, was the performance better? The re-inforced strings? The brass under better control? (probably, but that question I leave to others!)

Regarding my recording - I have a few key points:

1. The recording was monitored in the room with the orchestra at matched SPL - it sounded very close to the reality, dynamics & all

2. The sound you describe of your experience of my recording sounds like a classic case of both listening at well below performance SPL (equal loudness problems) and on less than accurate monitoring.

3. If you have been comparing the two recordings, you will need to take account of the missing 20dB of peak dynamics. After that, comparison is a very different kettle of fish.

4. Regarding the 'weak' strings & overpowering brass - this was exactly so in the room, as captured.

My recording was a straight stereo capture. No mixing/reverb/compression/etc or mastering involved.

The signal chain was parallel pair of Model A into BG1 into Mytek stereo96, taken from a good distance.

Perhaps if I were to compress ~20dB out of my recording & post it here, would that be of interest?

I'd certainly be interested in hearing the uncompressed original of your recording - perhaps even just the main pair, for similar comparison?

Please forgive me if it seems like I'm being needlessly rude - I do take this very seriously and hope that my comments will not be taken personally.

After all, your customers may be very happy and that is what counts.

Andy

[Plush]

Monitoring at actual volume does not prove anything. Why are you attaching such importance to it? Your musical clip is not a balanced recording because the strings are swamped by the brass. This simply means that you are too far away and also too high up in the air. (looking at the back of the orchestra)

I can play PA speakers loudly too. It doesn't mean squat, so what is the point?

[loranoyd]

Hi, Andy.

Many thanks for your comments. I do appreciate you taking the time to listen and to provide the comments below. It is clear to me from what I read that I am seriously out of depth in this thread. It will be nearly impossible for me to continue to participate here because:

1. I clearly do not own equipment equipment capable of reproducing sound at near orchestral SPL levels. I do not own a soundproof studio where I could experiment with these higher SPLs (I live in an apartment building with lots of neighbors), and I cannot set up reproducing equipment in a concert hall.

2. I do not own (or if I do own I do not know how to use or understand) the sort of calibrating equipment you do.

3. I do not possess 1/100th of the technical knowledge and expertise of recording you do, and so I feel I cannot carry on a meaningful discussion with you at your level.

I am not being sarcastic here. I am old enough to realize that I know what I know, and that there may be limitations to my ability to understand what is going on in this thread. That being said, I would greatly appreciate it if you could address a few questions I have regarding your observations and conclusions regarding my recording.

Originally Posted by andy_simpson

Hi Lloyd,

It's great to see somebody else putting something up here for a real world comparison and with the same section of the same piece this is very interesting.

I downloaded your clip & put aside a good half hour for listening & comparison.

I did listening firstly via my big mackie SRM450 workshop monitors and then via my DT100 headphones (amp was behringer 'powerplay').

As usual, since I have enough headroom in my monitors, I am able to bring up the level to that of a concert performance, which I did.

The first thing that strikes me is that there is so much compression/gain riding that I cannot balance one recording against the other without a -24dB or so on your recording (which can be more or less measured by looking at the RMS of either clip).

OK, you've lost me right out of the starting gate. I used no compressor in my signal chain, and I rode the faders only a little bit. In Digital Performer, I rode gain on the master fader only up +4 dB at maximum. The faders only allow up to +6 dB on any single fader. I rode that doggy only to bring up the quietest passages to a level that would be acceptable to listeners at home (or cars, etc.). I do not understand at all where the "so much compression/gain" statement is coming from.

Originally Posted by andy_simpson

Taking the PP entry section, I tried to match RMS loudness, which is where the -24dB came from.

After reducing gain by 24dB for 'equal' RMS levels at the outset, the lack of dynamics is quite enormous and makes comparison very hard, excepting the measureable lack of dynamics.

For example, directly before the first explosive entry there is a PPP section. In matching PPP sections such as this we see a huge difference in dynamics.

The compression, aside from removing the entire build-up before the explosion & all the dynamic interplay of the plucked strings & orchestra, aside from removing the actual explosion itself, also removes any vibrancy from the orchestra.

If the conductor had wanted zero dynamics it would have been more effective to play with zero dynamics.

Finally, at anywhere near to realistic levels the noise-reduction (gate) is very audible and sucks the air out of the hall directly before the first explosive entry.

I hate the air-con in performance halls as much as the next man, but I would draw the line at a complete -infinity gate (or at least, if I was trying to be sneaky, I would put it before the reverb in the signal chain, so that the reverb tails masked the effect as much possible!).

OK, if I may: The reason the sound goes away between the third and fourth movements is that my posted excerpt is taken from two different tracks of the CD I made of the performance. I faded out the signal at the end of the third movement, and added two seconds of silence to the end of the track. I think this is the source of the confusion regarding the "complete -infinity gate," which I don't really think I understand anyway. I claim inartful conjoining of two distinct tracks. That aside, don't you think you overstate things a bit by claiming there are "no dynamics" in my recording? Come on, now. There are most certainly easily perceived dynamics in my recording. Statements like yours above do nothing to promote the dialogue here, especially the crack about the conductor wanting no dynamics. Really! Did you honestly contemplate that the conductor of the orchestra asked me to produce a recording devoid of dynamics? Of course not!! So what do I take away from your comments? What can I learn here? You apologize at the end for seeming to be "needlessly rude," but why go there at all? By the way, apologizing for a perception of rudeness is not quite the same thing as apologizing for being rude.

So would someone please explain to me: what about the signal chain could possibly contribute to the effect you describe, that is, a tremendously reduced dynamic range? Digital Performer's limiter plug? Altiverb? Does gain riding a few decibels on my faders translate into that kind of reduction in dynamics?

I also get the distinct feeling that I missed some very important component about this thread: namely, that somehow we can compare the recorded sound of an orchestra to the actual orchestra. Since I'm not convinced that we can (in which case, I suppose I'm in company with those Remotesters who believe the art of recording is to create a sound, and that perfect replication is impossible/impractical), I feel as if I walked into the Smith wedding reception when I really was looking for the Frank Bar Mitzvah! But I am interested in clearing up some of my confusion. Anyone?


Lastly, Andy, I do not know what a Model A or a BG1 is. I would love some enlightenment.

Regards,

Lloyd

[rest of Andy's post trimmed for brevity]

[d_fu]

Originally Posted by Plush

Monitoring at actual volume does not prove anything. Why are you attaching such importance to it? I can play PA speakers loudly too. It doesn't mean squat, so what is the point?

I've been wondering that myself... I wouldn't necessarily consider a recording that needs original SPL levels to sound good to be a good recording. (This is a general comment, not about loranoyd's recording which I've had no time to listen to...)

[resound]

What a circuitous thread.

FWIW, LLoyd, take no offense to Andy's comments. As a casual listener, I much preferred your recording to his. I won't even venture a thought at critiquing the two against one another as it seems so clear that you're each after something different in your end result.

And I really don't understand the argument of reproduction SPLs, either.

[andy_simpson]

Originally Posted by Plush

Monitoring at actual volume does not prove anything. Why are you attaching such importance to it?

I am attaching importance to it for reasons of how SPL affects our hearing - 'the equal loudness contours'.

To restate, if we listen at any different SPL than that of the performance, the sound cannot be 'realistic' because of the equal loudness effect.

The thread is about 'colouration' VS 'accuracy'.

My point is that accuracy can only be judged at matched SPL, where colouration is ugly.

The derivative of my point is that those who prefer colouration are more likely to be listening at lower-than-performance SPL.

Your musical clip is not a balanced recording because the strings are swamped by the brass. This simply means that you are too far away and also too high up in the air. (looking at the back of the orchestra)

As I have said several times, this is almost exactly the balance that was in the hall in any of the audience seats. Too little strings, too much brass.

We checked this at length.

I of course do agree that it is unbalanced, but this is a matter for the conductor. I am only there to reproduce as closely as possible whatever sound the orchestra made.

If I came away with a recording with any other balance, my work is in vain.

I can play PA speakers loudly too. It doesn't mean squat, so what is the point?

You don't have to use 'PA' speakers. A large (efficient) pair of Westlakes, JBL, Strauss, K&H, Western Electric, Altec, etc will do the job.

The K&H 0500s actually sound quite similar to my large mackies (like a very clean version of them, with less dynamic resolution in the mids due to lower efficiency).

The point is: for those who ask for realistic, accurate sound, I am working to that end. However, the reality of realistic sound is that it requires realistic SPL in order that the equal loudness equation is in balance.

I am speaking to those who are saying 'accurate recording equipment is boring'.

In most cases of claimed accuracy it actually IS boring, but when you do hear a really realistic reproduction of an exciting orchestra, it is only as exciting or boring as the event itself.

Andy

[andy_simpson]

Originally Posted by loranoyd

Hi, Andy.

Many thanks for your comments. I do appreciate you taking the time to listen and to provide the comments below. It is clear to me from what I read that I am seriously out of depth in this thread.

This is Gearslutz! - there is no depth qualification required! As a person recording orchestra you are more than qualified to discuss anything with me or anyone else.

1. I clearly do not own equipment equipment capable of reproducing sound at near orchestral SPL levels. I do not own a soundproof studio where I could experiment with these higher SPLs (I live in an apartment building with lots of neighbors), and I cannot set up reproducing equipment in a concert hall.

This is a very real limitation to the pursuit of accurate sound recording & reproduction, though it may be almost universal.

However, if we are aware of it we can at least work with the knowledge of how it might affect our judgement.

2. I do not own (or if I do own I do not know how to use or understand) the sort of calibrating equipment you do.

A loudness meter is helpful, but not accurate enough to use for this kind of thing.

The only way to do it really is by ear, by experience or in direct comparison with the source.

3. I do not possess 1/100th of the technical knowledge and expertise of recording you do, and so I feel I cannot carry on a meaningful discussion with you at your level.

There is no need for deep technical knowledge, though a little reading on the subject of psychoacoustics would be very helpful. A google of 'equal loudness curves' will bring up a wealth of information that will be a good place to start.

The discussion is made meaningful by the experience of people like you, please don't drop out.

OK, you've lost me right out of the starting gate. I used no compressor in my signal chain, and I rode the faders only a little bit. In Digital Performer, I rode gain on the master fader only up +4 dB at maximum. The faders only allow up to +6 dB on any single fader. I rode that doggy only to bring up the quietest passages to a level that would be acceptable to listeners at home (or cars, etc.). I do not understand at all where the "so much compression/gain" statement is coming from.

Perhaps I misunderstood your previous posts. I thought you mentioned a compressor/limiter.

In any case, on examination of the two recordings, from PPP to FFF, it is quite clear that there is a very large disparity in dynamics (some 24dB!) between the two clips.

Is it possible that your orchestra played with 24dB less dynamics?

Or that your microphones recorded with 24dB less dynamics?

Or perhaps that with a little gain-riding, a little reduced dynamics performance (not uncommon in concert) and a little compression from the microphones, that this all added up to a 24dB difference from PPP to FFF?

I'm not sure.

As I said, if I match the PPP (prior to explosive intro) section from both clips, then the following explosive part has a 24dB difference in peak level.

OK, if I may: The reason the sound goes away between the third and fourth movements is that my posted excerpt is taken from two different tracks of the CD I made of the performance. I faded out the signal at the end of the third movement, and added two seconds of silence to the end of the track. I think this is the source of the confusion regarding the "complete -infinity gate," which I don't really think I understand anyway.

I see! That explains the 'slow action gate' sound (a fade). By '-infinity gate' I meant a variable 'ducking' gate, with the gain reduction set to -infinity. Some noise-gates can be set to reduce gain by a set amount, from -x dB to -infinity (silence). A -5dB gate can have a 'subtle' effect of noise-reduction, but a gate which closes to silence is not very subtle (often used in restoration work of very old recordings).

I claim inartful conjoining of two distinct tracks. That aside, don't you think you overstate things a bit by claiming there are "no dynamics" in my recording? Come on, now. There are most certainly easily perceived dynamics in my recording.

I only meant that a difference of 24dB of dynamics is huge!

'Perceived dynamics' is quite a different thing to 'actual dynamics'.

I do agree that there are similar 'perceived dynamics' in either clip.

Statements like yours above do nothing to promote the dialogue here, especially the crack about the conductor wanting no dynamics. Really! Did you honestly contemplate that the conductor of the orchestra asked me to produce a recording devoid of dynamics?

My appologies, perhaps I was overly blunt.

I didn't mean that the conductor would ask for 'no dynamics', rather that if 'less dynamics' was required, I would rather ask the conductor to play with less, than reduce it with processing.

Also, if the conductor had asked 'I want to hear everything in my car at 70mph', this would imply a drastic reduction in dynamics, which would be a legitimate customer request.

Of course not!! So what do I take away from your comments? What can I learn here? You apologize at the end for seeming to be "needlessly rude," but why go there at all? By the way, apologizing for a perception of rudeness is not quite the same thing as apologizing for being rude.

Indeed. I am sorry for being rude.

So would someone please explain to me: what about the signal chain could possibly contribute to the effect you describe, that is, a tremendously reduced dynamic range? Digital Performer's limiter plug? Altiverb? Does gain riding a few decibels on my faders translate into that kind of reduction in dynamics?

I'm still wondering about that. A limiter can certainly do so, as will gain-riding. Even clipping of either mic-amps, converters, or internal DSP can do the same thing.

Depending on microphone technology, there is some potential for 'compression' or 'clipping' to occur. This can happen at the capsule, electronics or mic-amp and if any of these stages are vulnerable, promixity can also play a part.

Generally, my microphones are designed to record with cleaner dynamics than a direct-radiator microphone, but this alone would not account for the differences.

I also get the distinct feeling that I missed some very important component about this thread: namely, that somehow we can compare the recorded sound of an orchestra to the actual orchestra. Since I'm not convinced that we can (in which case, I suppose I'm in company with those Remotesters who believe the art of recording is to create a sound, and that perfect replication is impossible/impractical), I feel as if I walked into the Smith wedding reception when I really was looking for the Frank Bar Mitzvah! But I am interested in clearing up some of my confusion. Anyone?

Oh no, you are in the right place! I am only addressing the idea of 'accurate recording' as being 'boring', where equal loudness effects are not taken into account, which is an integral part of the question: 'colouration vs accuracy'.

Lastly, Andy, I do not know what a Model A or a BG1 is. I would love some enlightenment.

The DAV BG1 is a microphone pre-amp I have recently added to my workshop & location recording tests (at the sound advice of Plush).

According to my own criteria - of comparison to the source - this is an amazing microphone amp (not to mention very cheap).

The Model A is a relatively newly developed microphone type - you can find details of on my site (below).

Again, I do appologise. I wish I had more time to review my own posts to ensure that I am not being rude.

As for transparency in my own work & testing, I make a standing offer of attendance & participation in my testing here in Poland.

Anyone who is interested, please email/pm and we can make the arrangements.

If anybody is interested in hosting such a test, on any scale (solo instrument, ensemble jazz, chamber, etc) please get in touch and perhaps we can do something interesting.

I am planning some outdoor testing with a local jazz band, where I will try some exploratory blind testing scenarios.

This kind of testing can be quite accessible to anybody with basic location facilities and reduces the variables quite significantly with regards to acoustics.

With best regards,

Andy

[David Rick]

Really, I don't think there's any such thing as an "accurate" recording of a musical event. I say this as someone who has made a lot of "documentary" type recordings using single-pair microphone techniques. I say it as someone who places a high value on tonally uncolored microphones and preamps. I say it as someone who much prefers to use the actual room reverberance rather than use something that comes out of a box. I don't take the issue lightly.

The truth is, ears and microphones don't hear the same way, because ears have brains attached to them and microphones don't. Also, a stereo recording can't possibly reproduce the immersive experience of sitting in an actual concert hall. A surround recording may suggest that experience, but it still isn't the real thing. Even if we could do complete sound field reconstruction in a living room, the resulting listener experience would still be much different because of the absence of visual cues.

So we can't really aspire to "Accuracy", but we can aspire to make recordings which fairly represent the musical intentions of the performers (and hopefully those of the composer). That's hard enough, but we can also aspire to convey some small but essential part of the experience of actually being at a performance. Sometimes these two goals are in conflict. In that case, I believe our primary responsibility is to the music.

David L. Rick
Seventh String Recording

[d_fu]

Originally Posted by David Rick

The truth is, ears and microphones don't hear the same way, because ears have brains attached to them and microphones don't.

And the brain has eyes attached to it, too. These tend to interact with the ears...


Daniel

[Remoteness]

Originally Posted by d_fu

And the brain has eyes attached to it, too. These tend to interact with the ears...


Daniel

Yes, but only when our eyes are open.

[andy_simpson]

Originally Posted by David Rick

Really, I don't think there's any such thing as an "accurate" recording of a musical event. I say this as someone who has made a lot of "documentary" type recordings using single-pair microphone techniques. I say it as someone who places a high value on tonally uncolored microphones and preamps. I say it as someone who much prefers to use the actual room reverberance rather than use something that comes out of a box. I don't take the issue lightly.

We can all agree that the term 'accurate' is a deeply flawed one.

Also, we can agree that when it comes to reproducing multi-directional reverberation 'cues' we are at a loss.

However, in the case of reproducing a phantom source that is indistinguishable from the real source - we are closer to this than most people imagine.

In any case, when you say 'tonally uncoloured microphones & pre-amps', the only way to judge this is with the correct 'equal loudness' relationship.

A perfectly uncoloured recording becomes very coloured at any other listening level and what may pass for uncoloured at low listening level may not turn out to be so at the correct level.

The truth is, ears and microphones don't hear the same way, because ears have brains attached to them and microphones don't.

Although this is a popular perspective, I think it is somewhat beside the point. If the microphones heard like ears do, why would we need our own ears?

We don't need microphones to hear like our ears, we only need them to convey the correct information so that our ears can 'hear like ears'.

In other words, microphones don't hear like ears, they hear 'like speakers' and the problem with both microphones & speakers is that they don't have the mechanical resolution that the ears do (though some come 15-20dB closer than others).

Even if we could do complete sound field reconstruction in a living room, the resulting listener experience would still be much different because of the absence of visual cues.

This is a good point. Not just the absence of, but conflicting visual cues can do even more damage.

So we can't really aspire to "Accuracy", but we can aspire to make recordings which fairly represent the musical intentions of the performers (and hopefully those of the composer).

In the simplest technical terms, the musical intentions of the performers are set out in absolute sound pressure over time.

Notice I don't say 'variation in sound pressure over time' but 'absolute sound pressure over time'.

Because our ears are subject to the 'equal loudness contours', both tonality & perception of musical balance are subject to the absolute sound pressure.

Further to this, since the 'equal loudness contours' are averaged over a given number of subjects, we can safely assume that we are not all subject to the exact same equal-loudness effects, which is to say our equal loudness related perception will only match the source in the case of specific absolute loudness being matched.

Where loudness is not matched, since there is variation in equal loudness effects between the subjects, there can be no valid 'correction curve' for more than one person.

It may seem as if I am repeating myself terribly on the subject of the equal loudness, but this aspect of our hearing is responsible for much of the subjective confusion that rules this industry and this serves to undermine much of the real progress made.

Andy

[d_fu]

Originally Posted by andy_simpson

In any case, when you say 'tonally uncoloured microphones & pre-amps', the only way to judge this is with the correct 'equal loudness' relationship.
A perfectly uncoloured recording becomes very coloured at any other listening level and what may pass for uncoloured at low listening level may not turn out to be so at the correct level.

Andy,

I know this is your big personal "issue", but I'm afraid I'm not buying it... My recordings (as well as others I've heard) don't significantly change colour at different levels...

We all know human hearing needs a bit of compensation at low levels (which is why hifi gear has "Loudness" switches and knobs), but the significant frequency ranges that determine a recording's colour don't change so much with volume, and neither does the mix.

Anyhow, how and where does one determine the "correct" SPL? At the listener's position or at the (main) microphones' position, which more often than not will not be identical. How does the presence of spot mics become a factor? And how would one determine the "correct" playback SPL for a given recording? And what's with future CD publications? Would these need to print the recommended playback SPL on the cover? "This CD will sound correct at 98 dB SPL @ 0dBFS. We do not assume responsibility for diminished listening pleasure at lower levels" Each CD would have to include a 0 dBFS test tone and listeners would have to by SPL measurement devices...
How does this apply to semi- or non-acoustic recordings? How would you determine the correct, uncoloured SPL for a Fender Rhodes or an electric bass?

In the simplest technical terms, the musical intentions of the performers are set out in absolute sound pressure over time.
Notice I don't say 'variation in sound pressure over time' but 'absolute sound pressure over time'.

This would mean that any live concert performance could only be listened to "correctly" at one certain distance from the performers (which?). Everyone else would not be getting the correct uncoloured sound and would thus be eligible for a refund...

I must say I find the concept/idea fundamentally flawed... And yes, you are repeating yourself terribly...


Daniel

[andy_simpson]

Originally Posted by d_fu

We all know human hearing needs a bit of compensation at low levels (which is why hifi gear has "Loudness" switches and knobs), but the significant frequency ranges that determine a recording's colour don't change so much with volume, and neither does the mix.

According to the graphs available any number of places (eg. Equal-loudness contour - Wikipedia, the free encyclopedia) what you are saying is not correct.

'...colour don't change so much with volume' is of course a subjective statement and is also relative to how much difference in volume we are talking about.

1-2dB difference in SPL may not cause a drastic change in spectral balance, but 10dB, 20dB or even 30dB will certainly do so.

Anyhow, how and where does one determine the "correct" SPL?

This can be measured & recorded with callibrated measurement mics, if necessary.

At the listener's position or at the (main) microphones' position, which more often than not will not be identical.

This is an artistic or philosophical question, which we need not confuse ourselves with at this point.

How does the presence of spot mics become a factor?

It surely complicates things but I don't see it as any more of a factor than the other problems associated with spot-mic'ing.

And how would one determine the "correct" playback SPL for a given recording?

Although a valid question, this doesn't address the fundamental validity of the question, so we'll come back to it later.

And what's with future CD publications? Would these need to print the recommended playback SPL on the cover? "This CD will sound correct at 98 dB SPL @ 0dBFS. We do not assume responsibility for diminished listening pleasure at lower levels" Each CD would have to include a 0 dBFS test tone and listeners would have to by SPL measurement devices...

Why not? A little information for the end-user would do no harm - perhaps then the users would realise when to use their 'loudness' buttons.

How does this apply to semi- or non-acoustic recordings? How would you determine the correct, uncoloured SPL for a Fender Rhodes or an electric bass?

As far as this applies to this kind of music recording, the engineer/producer should be aware of these things since they will affect the balance he achieves at a given SPL.

For example, those creating dance music 'club anthems' for playback in a club at high SPL would be well served to be aware of this. Indeed many of them are aware (if not aware of the science behind the effect) and often do engineer on very large scale loudspeakers at high SPL.

This would mean that any live concert performance could only be listened to "correctly" at one certain distance from the performers (which?). Everyone else would not be getting the correct uncoloured sound and would thus be eligible for a refund...

Does it not sound brighter, deeper & more exciting at the podium than the back-row?

You have listed some apparently comical but logical conclusions, which whether you intended it or not are close to the mark.

I must say I find the concept/idea fundamentally flawed... And yes, you are repeating yourself terribly...

I will continue to repeat myself until this sinks in, if you don't mind too much.

Andy

[d_fu]

Originally Posted by andy_simpson

According to the graphs available any number of places (eg. Equal-loudness contour - Wikipedia, the free encyclopedia) what you are saying is not correct.

Then we seem to interpret these graphs differently.

1-2dB difference in SPL may not cause a drastic change in spectral balance, but 10dB, 20dB or even 30dB will certainly do so.

Please tell me what kind of differences you hear (besides the obvious, which may require a Loudness correction). I remain skeptical.

This can be measured & recorded with callibrated measurement mics, if necessary.

Ummm... Yes, obviously.

This is an artistic or philosophical question, which we need not confuse ourselves with at this point.

I think that's the main point. Where do you "correctly" measure an orchestra's SPL?

Does it not sound brighter, deeper & more exciting at the podium than the back-row?

Depends on the hall. It may also sound less well-balanced, if you're sitting in front (and possibly at the feet) of the strings...

I will continue to repeat myself until this sinks in, if you don't mind too much.

I think I can live with that... As long as you tolerate that I will continue to doubt the validity of this concept..

Daniel
Andy

[andy_simpson]

Originally Posted by d_fu

Then we seem to interpret these graphs differently.

Please tell me what kind of differences you hear (besides the obvious, which may require a Loudness correction). I remain skeptical.

I'm not sure what you mean by 'the obvious, which may require a loudness correction'?

There is an article here which might shed a little more light on the matter.

Also graphs here are little easier on the eye.

What I hear is pretty close to how these graphs look upside down.

It is worth noting how things vary with SPL - as we reduce master gain on our orchestra, we find different effects at each level of performance from PPP to FFF.

I think that's the main point. Where do you "correctly" measure an orchestra's SPL?

This is a matter for artistic perspective - where do you wish to put the audience?

I think I can live with that... As long as you tolerate that I will continue to doubt the validity of this concept..

That's fine with me.

Andy

[d_fu]

Originally Posted by andy_simpson

I'm not sure what you mean by 'the obvious, which may require a loudness correction'?

Bass and treble sensitivity at low levels.

There is an article here which might shed a little more light on the matter.

Ineresting reading indeed. However, I'd think that most speakers' nonlinearities are greater than those between different SPLs described here. Also, these curves were determined with sine tones - can they really be used to judge complex (musical) signals, which don't usually feature all frequencies in equal SPL levels...

Also, the ear apparently adapts to different volumes as well as to different speakers. How else would you be able to recognize a voice on any shitty speaker? Or how else would I recognize the typical sound of the ride cymbal in the beginning of "Take Five" in an instant on my car stereo, at low volume?

This is a matter for artistic perspective - where do you wish to put the audience?

Usually, my microphones are not right where the audience is - so where should I measure "correct" SPL? In the audience position, regardless of the mics' position, or where the mics are?

I would definitely consider it a failure if a recording I had made only sounded good at "original" SPL. If it didn't sound good at low levels (minding the neighbours), in mono on the kitchen radio, or in the car, then I'd have to change something. But I don't think I've ever had a recording where I really thought playback SPL was an issue.

Daniel


Daniel