Head-Fi.org › Forums › Equipment Forums › Sound Science › What makes piano sound so hard to reproduce?
New Posts  All Forums:Forum Nav:

What makes piano sound so hard to reproduce? - Page 7

post #91 of 191
Quote:
Originally Posted by Eisenhower View Post

Part of the problem is that people are very familiar with the way pianos sound. Grand pianos are not kept in cases tucked under peoples' beds like guitars, they are usually out in the open for anyone to play (or sneak a few notes in). I mean, who hasn't played notes on a piano before? Also pianos (of the same type) do not have wildly different sound qualities, unlike vocals or synthesizers or electric guitars. This means that there is essentially a standard piano sound that people are quite familiar with. They are also used in just about every genre of music.

 

For other instruments with more variations in the sound, it is more difficult to say it's not reproduced accurately unless you were in the studio and heard the actual guitar / recording. A piano is not a complex instrument, but I think they have a sound that is simple enough for us to remember what they ought to sound like, but complex enough for reproduction to not be trivial like a sine wave.

 

So it isn't that pianos are unusually difficult to reproduce -- it is that it's easier for us to identify flaws in a piano recording because of our familiarity with them.

 

A great example is the funky upright in abbey road studios that is so famous in recordings NI released it as a sample.  I really like the NI grand pianos, its nice they let you load a full version with all overtones and it sounds quite amazing. 

 

Thats one thing I've never understood about high end stage pianos and workstations synths....have can they compete with the sample size of a software sample and still make pretty great piano sounds. 

post #92 of 191
Quote:
Originally Posted by bigshot View Post


Custom made 12 inch five ways from the 70s and a pair of JBL towers up front, a top of the line 12 inch Sunfire subwoofer, Klipsch center channel, and very soon, I'll be hooking up 10 inch custom three ways from the 70s in the rear.

 

So you like bass.

post #93 of 191
Quote:
Originally Posted by Eisenhower View Post

Part of the problem is that people are very familiar with the way pianos sound. Grand pianos are not kept in cases tucked under peoples' beds like guitars, they are usually out in the open for anyone to play (or sneak a few notes in). I mean, who hasn't played notes on a piano before? Also pianos (of the same type) do not have wildly different sound qualities, unlike vocals or synthesizers or electric guitars. This means that there is essentially a standard piano sound that people are quite familiar with. They are also used in just about every genre of music.

 

For other instruments with more variations in the sound, it is more difficult to say it's not reproduced accurately unless you were in the studio and heard the actual guitar / recording. A piano is not a complex instrument, but I think they have a sound that is simple enough for us to remember what they ought to sound like, but complex enough for reproduction to not be trivial like a sine wave.

 

So it isn't that pianos are unusually difficult to reproduce -- it is that it's easier for us to identify flaws in a piano recording because of our familiarity with them.

Piano is second only to the familiarity with human voice to most people. It is a part of the complex problem of both recording and reproducing it correctly It is also very dependant on one's expectations - when once recording a heavily piano featured programme of new compositions, one of the composers objected heavily to my recording, whereas his best friend, also fetured composer that night, said he never had a better recording of his composition. Same hall, same piano, same pianist, same microphones, same recorder.

 

There is a study made for most instruments how do they sound or sound being captured by microphone(s) from every possible angle. Nice as information, but to me of little use or significance, as I always strive for the sound as perceived by the listener in the public. 

post #94 of 191
Quote:
Originally Posted by pabloaugustus View Post

 

Thanks for the links...I'd really like to try out a Mic that can pic up those real highs and do some multi-micing on my tenor pans (aka steel drum)  and see if I can make it sound better.  Obviously I EQ out much of the highs but the fact that they are there may make the total instrument sound more realistic.   Its so hard to record and not have it sound harsh.  This is another very hard to capture instrument.

 

Anyway, we are way off topic just want to say I appreciate those links, and sound above 20k, whether I can hear it or not it affects me and the instrument I play, which is probably quite rare, I've never heard such high pitches or suffered such hearing loss as when gigging with a 40-60 person steel band.  If you live in NYC or London there are great pan scenes, they always welcome beginners and its really fun to play.  (brooklyn has a trinidadian carnival with a panorama competition, its like the super bowl for people from Trinidad & Tobago...the Trinidad Panorama or course, Paris has one now too)

If you, or anyone else, is interested just how similar microphones with 25, 30, 40 and 50 kHz as specified high frequency response sound, contact Earthworks - couple of years back they used to have free CDs available with exactly the same music recorded under exactly the same conditions recorded with similar mics in exactly the same position with ever wider frequency response. These recordings also featured comparisons to some of the most established familiar studio mics from competitive manufacturers. I no longer see these CDs on their site ( perhaps they simply grew tired of people saying above 20 k does not matter to continue sending free CDs around the globe  ).

 

Even on CD, the difference among the mics is immediately audible. It might not hurt to ask - perhaps they still have some on stock.  

post #95 of 191
Quote:
Originally Posted by analogsurviver View Post

If you, or anyone else, is interested just how similar microphones with 25, 30, 40 and 50 kHz as specified high frequency response sound, contact Earthworks - couple of years back they used to have free CDs available with exactly the same music recorded under exactly the same conditions recorded with similar mics in exactly the same position with ever wider frequency response. These recordings also featured comparisons to some of the most established familiar studio mics from competitive manufacturers. I no longer see these CDs on their site ( perhaps they simply grew tired of people saying above 20 k does not matter to continue sending free CDs around the globe  ).

 

Even on CD, the difference among the mics is immediately audible. It might not hurt to ask - perhaps they still have some on stock.  

So what are you trying to say here, exactly? That different mics sound different? That is a shocking discovery.

post #96 of 191
Quote:
Originally Posted by wuwhere View Post

So you like bass.

It's equalized so it can go from a whisper to a roar, depending on the music.
post #97 of 191
Quote:
Originally Posted by analogsurviver View Post

If you, or anyone else, is interested just how similar microphones with 25, 30, 40 and 50 kHz as specified high frequency response sound, contact Earthworks - couple of years back they used to have free CDs available with exactly the same music recorded under exactly the same conditions recorded with similar mics in exactly the same position with ever wider frequency response.

I'm sure they're very good demonstrations considering that CDs don't go above 21kHz.
post #98 of 191

@ xnor and bigshot:

 

I can understand your reaction - because you have never heard an audio system capable of reproducing such differences clearly. I spent most my listening with good earspeakers with response good to frequencies you both consider are irelevant to music to reproduction off vynil - and therefore have pretty good idea just how does sound a

 

1.) phono cartridge that is relatively (or next to totally, say less than + - 0,5 dB 20 Hz - 20 kHz) flat and rolls off rather sharply above say 22-23 kHz ( like a well designed MM will do ) 

 

2.) phono cartridge that is the same 20-20k but extends response to say 35 kHz with a more gentle rolloff ( like a medium good MC will do )

 

3.) phono cartridge that is even more flat 20-20k but extends its frequency response past 67 or so kHz ( test record with signals to 50 kHz recorded at 33 1/3 RPM but reproduced at 45 RPM ) ( like a really good MC or low impedance MM cartridge will do )

 

There was even a cartridge that is capable of 120 kHz response ( Technics EPC P100CMK4 ) - which I regrettably have not heard yet.

 

The sound differences in phono cartridges are directly analogous to sound differences in microphones - the better the extension in high frequency range, the better the reproduction of ambient or acoustic where music was recorded. The entire chain should ideally be very wide band, so that even that 120 kHz cartridge is the slowest component in the system. Technics could provide for such a system in early 80s, with its ribbon tweeters that went to 150 kHz - amplification was of course even faster. About five seconds of audio through such a system is all that it takes to convience people

with reasonably normal hearing for their age - no need to be able to hear pure sine tone at 20 kHz needed, not to mention outbating bats by about two times their hearing capabilities. As I mentioned, a vynil record mastered at half speed is capable of flat frequency response to 50 kHz - which does not mean it filters above that like brick wall filtering of CD, but rolls off at about 6 dB per octave, which means that on records can be recorded signals, not noise/distortion, in excess of 100 kHz - not linear, but around -6 to - 15 or so dB, depending on the HF behaviour of the cartridge.

 

Earthworks's CDs of course could not provide for the full benefit of response extended to various points above the 20k, all the way to 50k. But relative differences still hold true. Wish they could ( should ! for their own benefit ) make their recordings available also on SACDs

 

Since this is a thread about the piano - DPA, the famous Danish microphone manufacturer, issued a couple of years ago a disc featuring piano showcasing their microphones. http://www.dpamicrophones.com/en/Download/~/media/PDF/Download/grandpiano.pdf

http://www.dpamicrophones.com/en/Mic-University/Application-Guide/Grand-Piano.aspx

Although I do not agree with for my taste too close microphone positions, I certainly do agree with their choice of digital resolution of the disc - it is double layer SACD/CD playable on any CD player, but obviously playing back DSD recording as SACD on SACD players only. You can turn the 7 microphone test listened to either SACD or CD layers into 7 microphones on SACD vs CD layers/resolutions/FREQUENCY RESPONSES.

 

This also takes about 5 minutes to convience anybody with reasonably preserved hearing in the supremacy of SACD over CD - since this is head-fi,

I would recommend electrostatic headphones, say Stax Lambda series, driven preferably by amplifiers and not transformers that may limit the 41 kHz response of Lambda ( Pro, to be exact of the model, but others are close enough in the frequency response extension not to influence the outcome ).

But any reasonably good HPs should clearly differentiate between SACD and CD versions.

 

Needless to say, DPA strongly recommends listening to SACD layer, if possible; if it was you who put al the research and development in the mics and associated preamps, you would too try to make certain to present the sound quality of your product best you possibly can - and that is by no means CD with frequency response limited to just above 20k..

 

No idea if this DPA piano SACD/CD is still available; it may not hurt to ask.

 

Disclaimer: No affiliation with mic manufacturers - satisfied DPA customer (wish I could afford matched pair of Earthworks QTC50); would only like to present the effort that is being made in order to bring us closer to the real thing.

 

@ bigshot: the speaker system described listing only the multiple drivers covering various portions of the audible spectrum without specifiying properly designed passive, but preferably active crossovers, is not likely to offer coherent reproduction. It is certainly a far cry from probably the best conceived, designed and executed audio system in the world today:

 

http://www.royaldevice.com/custom31.htm

post #99 of 191
I grow very weary. It just isn't worth commenting on reams of text. I have no idea where to start. All I can say is that you're in the wrong forum for this sort of stuff. It would go over much better with those who don't have any idea how sound reproduction works.
post #100 of 191
Quote:
Originally Posted by analogsurviver View Post

 

Yes, transducers DO have 'a sound-signature' ..


Edited by AKG240mkII - 12/30/12 at 5:11am
post #101 of 191
Quote:
Originally Posted by AKG240mkII View Post

Yes, transducers DO have 'a sound-signature' ..

Point you seem not to understand or ignore on purpose, in example of Earthworks mics at least, is the fact that mics featured on those CDs are designed by the same man, are essentially the same deesign, the main difference being the high frequency extension. There are other differences, such as noise figures, which normally tend to go better with increase of cost, but are not nearly as directly audible as differences in HF extension.

It is a test that leaves all other variables save frequency response at an absolute lowest level achievable with real world designs.

 

Any real transducer to date DOES have "a sound signature" - so does electronics, despite being in the worst case scenario at least an order of magnitude better than the best transducer available. Yet - if people compare amps and such by listening through said at least one order of magnitude worse transducers than the amps driving them and can still differentiate among amps and describe their sonic character - the same can be concluded for mics with response above 20 kHz,, even if played on CDs. Still, a marked difference can be heard. You really should try it for yourself - if you then claim you can not hear any difference, we might have still topic for debate. A priori dismissing the possibility of audibility of high frequency extension differences in microphones with various responses past 20 kHz without ever giving it a listening try no longer leaves this topic open. 

 

Faster transducer , if not seriously flawed in any way, will always sound more natural  than the slower one with comparable characteristics in the "midband". In real life, you get to hear it all an instrument or voice has - no part is filtered out by any other means but your ears.  Real listening is not limited by technological means - I see it hard why is so simple fact so hard to understand for some people Any real mic/preamp will tend to limit the amount of information reaching your ears - the fastest one will the best approach the sound heard live.

post #102 of 191
Quote:
Originally Posted by analogsurviver View Post

Point you seem not to understand or ignore on purpose, in example of Earthworks mics at least, is the fact that mics featured on those CDs are designed by the same man, are essentially the same deesign, the main difference being the high frequency extension. There are other differences, such as noise figures, which normally tend to go better with increase of cost, but are not nearly as directly audible as differences in HF extension.

It is a test that leaves all other variables save frequency response at an absolute lowest level achievable with real world designs.

You do not seem to understand the differences between polar patterns and that there is a clear difference in frequency response within the audible range.

post #103 of 191
Quote:
Originally Posted by analogsurviver View Post

Point you seem not to understand or ignore on purpose, in example of Earthworks mics at least, is the fact that mics featured on those CDs are designed by the same man, are essentially the same deesign, the main difference being the high frequency extension. There are other differences, such as noise figures, which normally tend to go better with increase of cost, but are not nearly as directly audible as differences in HF extension.

It is a test that leaves all other variables save frequency response at an absolute lowest level achievable with real world designs.

 

Any real transducer to date DOES have "a sound signature" - so does electronics, despite being in the worst case scenario at least an order of magnitude better than the best transducer available. Yet - if people compare amps and such by listening through said at least one order of magnitude worse transducers than the amps driving them and can still differentiate among amps and describe their sonic character - the same can be concluded for mics with response above 20 kHz,, even if played on CDs. Still, a marked difference can be heard. You really should try it for yourself - if you then claim you can not hear any difference, we might have still topic for debate. A priori dismissing the possibility of audibility of high frequency extension differences in microphones with various responses past 20 kHz without ever giving it a listening try no longer leaves this topic open. 

 

Faster transducer , if not seriously flawed in any way, will always sound more natural  than the slower one with comparable characteristics in the "midband". In real life, you get to hear it all an instrument or voice has - no part is filtered out by any other means but your ears.  Real listening is not limited by technological means - I see it hard why is so simple fact so hard to understand for some people Any real mic/preamp will tend to limit the amount of information reaching your ears - the fastest one will the best approach the sound heard live.

The argument here might be valid if the listening test was conducted in a way that tested for the effect of HF response only, and eliminated all other influences, i.e. the tester's knowledge  that the mic under test is supposed to have extended high frequency response.  You also need two test mics with identical in-band response, but one with extended HF response so that you are only testing for that.  Otherwise what you have is a test of the total microphone experience, which includes the total frequency response - in band and above, the use, handling, appearance, knowledge of specs and reputation, etc.  That's not a scientific test, and would prove nothing about the effect only of extended HF response alone.  What you have is more like a focus group evaluating the total mic experience.

 

The same holds for the supposed "signature" of electronics, only that one's been at least partially been done and published - power amps, preamps, etc. And pretty much we all know the results. 

post #104 of 191
Attributing a sound signature to inaudible frequencies is patently absurd.
post #105 of 191
Quote:
Originally Posted by analogsurviver View Post

@ xnor and bigshot:

 

I can understand your reaction - because you have never heard an audio system capable of reproducing such differences clearly. I spent most my listening with good earspeakers with response good to frequencies you both consider are irelevant to music to reproduction off vynil - and therefore have pretty good idea just how does sound a

 

1.) phono cartridge that is relatively (or next to totally, say less than + - 0,5 dB 20 Hz - 20 kHz) flat and rolls off rather sharply above say 22-23 kHz ( like a well designed MM will do ) 

 

2.) phono cartridge that is the same 20-20k but extends response to say 35 kHz with a more gentle rolloff ( like a medium good MC will do )

 

3.) phono cartridge that is even more flat 20-20k but extends its frequency response past 67 or so kHz ( test record with signals to 50 kHz recorded at 33 1/3 RPM but reproduced at 45 RPM ) ( like a really good MC or low impedance MM cartridge will do )

 

There was even a cartridge that is capable of 120 kHz response ( Technics EPC P100CMK4 ) - which I regrettably have not heard yet.

 

The sound differences in phono cartridges are directly analogous to sound differences in microphones - the better the extension in high frequency range, the better the reproduction of ambient or acoustic where music was recorded. The entire chain should ideally be very wide band, so that even that 120 kHz cartridge is the slowest component in the system. Technics could provide for such a system in early 80s, with its ribbon tweeters that went to 150 kHz - amplification was of course even faster. About five seconds of audio through such a system is all that it takes to convience people

with reasonably normal hearing for their age - no need to be able to hear pure sine tone at 20 kHz needed, not to mention outbating bats by about two times their hearing capabilities. As I mentioned, a vynil record mastered at half speed is capable of flat frequency response to 50 kHz - which does not mean it filters above that like brick wall filtering of CD, but rolls off at about 6 dB per octave, which means that on records can be recorded signals, not noise/distortion, in excess of 100 kHz - not linear, but around -6 to - 15 or so dB, depending on the HF behaviour of the cartridge.

 

Earthworks's CDs of course could not provide for the full benefit of response extended to various points above the 20k, all the way to 50k. But relative differences still hold true. Wish they could ( should ! for their own benefit ) make their recordings available also on SACDs

 

Since this is a thread about the piano - DPA, the famous Danish microphone manufacturer, issued a couple of years ago a disc featuring piano showcasing their microphones. http://www.dpamicrophones.com/en/Download/~/media/PDF/Download/grandpiano.pdf

http://www.dpamicrophones.com/en/Mic-University/Application-Guide/Grand-Piano.aspx

Although I do not agree with for my taste too close microphone positions, I certainly do agree with their choice of digital resolution of the disc - it is double layer SACD/CD playable on any CD player, but obviously playing back DSD recording as SACD on SACD players only. You can turn the 7 microphone test listened to either SACD or CD layers into 7 microphones on SACD vs CD layers/resolutions/FREQUENCY RESPONSES.

 

This also takes about 5 minutes to convience anybody with reasonably preserved hearing in the supremacy of SACD over CD - since this is head-fi,

I would recommend electrostatic headphones, say Stax Lambda series, driven preferably by amplifiers and not transformers that may limit the 41 kHz response of Lambda ( Pro, to be exact of the model, but others are close enough in the frequency response extension not to influence the outcome ).

But any reasonably good HPs should clearly differentiate between SACD and CD versions.

 

Needless to say, DPA strongly recommends listening to SACD layer, if possible; if it was you who put al the research and development in the mics and associated preamps, you would too try to make certain to present the sound quality of your product best you possibly can - and that is by no means CD with frequency response limited to just above 20k..

 

No idea if this DPA piano SACD/CD is still available; it may not hurt to ask.

 

Disclaimer: No affiliation with mic manufacturers - satisfied DPA customer (wish I could afford matched pair of Earthworks QTC50); would only like to present the effort that is being made in order to bring us closer to the real thing.

 

@ bigshot: the speaker system described listing only the multiple drivers covering various portions of the audible spectrum without specifiying properly designed passive, but preferably active crossovers, is not likely to offer coherent reproduction. It is certainly a far cry from probably the best conceived, designed and executed audio system in the world today:

 

http://www.royaldevice.com/custom31.htm

 

Just gotta say, a very thought provoking post!

thx

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › What makes piano sound so hard to reproduce?