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 2

post #16 of 191
There is absolutely no evidence whatsoever that sound above 20kHz adds anything to music. If it can be perceived at all, it's just pressure, not sound. Frequency extension is a red herring. In fact, the frequencies below 10kHz are more important than those above by an order of magnitude. There is very little above 10kHz in music, some weak harmonics and cymbal crashes. In piano music, pretty close to zilch except at the extreme high end of te keyboard. And that is just low level harmonics.

The highest note on a keyboard is about 4kHz. 16kHz would be plenty to cover the audible harmonics for that.
Edited by bigshot - 12/20/12 at 3:16pm
post #17 of 191

One of my favorite Naxos recordings is this one...it actually sounds like a trumpeter standing in front of a piano...not sitting inside it!

 

"Gershwin for Trumpet (arr. J. Bartos and P. Breiner)" http://www.naxos.com/catalogue/item.asp?item_code=8.554302


Edited by sejarzo - 12/20/12 at 4:22pm
post #18 of 191
Quote:
Originally Posted by analogsurviver View Post

Yes, the dynamic range you cite is about correct. When listening/recording "normal" piano music. I was really taken by surprise just how loud or hard the piano can be played for real - about two years ago when recording a piece by a modertn compser that is extremely seldom performed due to the extreme demands on the part of the pianist.

Exceptions prove the rule.

 

Quote:
It is so demanding that the pianist refused to play it for me in order to set the recording levels right during the rehearsal. I did adjust for about - 4 dBFS ( please note when recording DSD, there is no brick wall at 0 dB as with PCM, there is yet 3 dB headroom over 0 dBFS available without objectionable distortion ) - in effect, giving me 7 dB headroom above anything played at the rehearsal.

Every recording engineer I know leaves at least 10 to 12 dB headroom from the highest peak to 0 dBFS. Granted, you couldn't adjust for the extremely loud part, but in that case you should have added even more headroom, maybe 20 dB.. instead you went for 7 dB???

 

Quote:
Come the concert ... all peak LEDs turned red ! When trying to ressurect that piece, a guesstimate of about + 8 dB was made, according to the general trend seen on the computer ...

Is there a proper recording of that piece? What's the name of it? It's kinda funny how it is performed "extremely seldom" and that there probably is no proper recording.. I'm not saying your story isn't true or is, but I cannot be sure. (See my signature for anecdotal evidence.)

 

Quote:
Regarding high frequency overtones - I might well have said upper midrange overtones. There are numerous mechanisms to distort the sound of the piano, from microphone through all amplification to the end transducer, be it headphone or speaker.

Ok, that's a different matter which I agree with. But some headphones can reproduce square waves (and some electronica tracks make use of them) very well. I'd argue that these tracks are a lot more demanding than the sound of a piano, having very low crest factors and lots and lots of bass.

 

Quote:
Please note that commercially available CDs which I assume you have used to measure their dynamic range, were subject to mastering, with further losses of fidelity along the way. The first to suffer is the dynamic range.

But I also checked the ones you suggested in #7 and no, the dynamic range is not greater. It's mostly in the ~40 to 50 dB range. I also checked some high-res "audiophile HD" files..

 

Quote:

Level of overtones ABOVE 20 kHz, which is accepted as high frequency limit of hearing of humans, is really minuscule - but it does make the difference. I guess it is on average below 1 % in amplitude vs "main sound" - but is the same as salt in soup - totally without it is soup not tasty for most people, too salted is undegestible.

Why is every analogy I encounter on this forum flawed? Salt is a very basic ingredient. Missing salt is like a roll-off at 10 kHz.

Let me try another analogy: the overtones above what is audible (an optimistic ~20 kHz) are like the decorative things at the edge of the plate that you do not eat.

 

 

Quote:
It is this extension above 20 kHz that allowed analog to survive and return in ever bigger numbers - CD just would not allow for the correct presentation of recorded acoustics, for which correct phase and therefore frequency response is mandatory - CD with the brick wall filtering above 20 kHz is inherently uncapable of doing it. 

No it is the FUD spread about digital audio by analog equipment manufacturers/advocates to keep it alive. Or all the distortions/noises analog equipment adds to the sound that some people like / are used to (they call it "warmth"). There are many more reasons, but audio quality or high-fidelity isn't one of them as much as vinyl lovers would like to tell us otherwise.

 

It's funny that you mention phase and frequency response. The lowpass filter in DACs has linear phase. The frequency response is absolutely flat from at least 20 Hz to 20 kHz, unlike vinyl where "Frequency deviations of 5-10 dB or greater are not uncommon in the 20 kHz range for many records." and "playback of ultrasound frequencies is still not guaranteed" for a plethora of reasons. And btw, the filter has to attenuate stuff at and above 22.05 kHz, not at 20 kHz.

 

Quote:
If you do not believe me - try it by yourself. If you want to improve the quality of your bass, the right thing to do, provided your woofers are at least respectable in the first place, is to add - supertweeter. Of course, you have to feed it with a source that extends above 20 kHz - ruling out the CD.

With analog and emerging high resolution digital supporting frequency response above 20 kHz, it does make sense - just check the catalog of Tannoy, for example; they would not be offering supertweeters if there were no fire behind this smoke, they have far too good reputation to jeopardize it by someone calling them just greedy for offering supertweeters for nothing but profit to themselves and no gain to the buyers.

This is a pretty weak argument. Denon sells an "audiophile ethernet cable" for $500 that is a measly 1.5m long. I guess I do not need to mention that most of these ridiculous products are targetet at "engrained" audiophiles, not pro-audio/studios ... They are not jeopardizing their reputation by serving a small but highly profitable segment. There actually doesn't have to be an audible gain to the buyers, bias and placebo effects are strong enough to justify such purchases. It even works for companies that sell cheap cables at crazy prices.

Do you have any independently confirmed ABX tests to link to that show the effect of supertweeters? I read somewhere that the high-frequency noise can cause distortions in the audible frequency range, but this sounded more like a hypothesis than anything else.

Also, by adding such a supertweeter you also have to add another crossover, causing possible audible phase and frequency response distortion. Ouch!

 

Quote:
It is funny how people think that an audio device will sound sharper/harsher if it has more extended frequency response. Exactly opposite is true - the more extended the rsponse, the more naturally rounded "nothing to comment upon" sound emerges. Want proof ? Go to the 

 

http://www.2l.no/hires/index.html

 

You can download the same music in multitude of formats/resolutions. The best you can do at home is to listen to DSD download played on DSD recorder or DSD DAC - no DXD devices commercially available that are known to me. And you can proceed from there to PCM - from 192/24 to lower resolutions.- at least they had it available not so long ago.

 

Or go to Linn Records http://www.linnrecords.com/linn-downloads.aspx and download various PCMs of the same piece of music - from 192/24 down to the MP3 - you can play these with foobar2000 - and hear the difference for yourself. The lower the resolution, the sharper/harsher the sound.

I've looked at the 2l test files years ago, and did again today. DXD or DSD, I've tried both. If you subtract a resampled to 44.1 kHz file from the high sample-rate file and look at the remaining signal, there are a couple of things to note:

- the rms amplitude of the remaining signal is extremely low, some files showed about -80 dB average to below -100 dB minimal

- even if these remaining signal components were in the range of audible frequencies, they would be masked big time

- except for a few blips of weird noise, it is just white noise ... *pfffffffffffffffffffffffft* (made audible by shifting the frequencies)

 

I don't trust Linn anymore, after they've published differently mastered files of the supposedly same track (44.1/16 and 24/96 as far as I can remember) for people to hear the difference between the formats. Jeez, even the waveforms showed clearly visible differences.


Edited by xnor - 12/20/12 at 3:53pm
post #19 of 191
Quote:
Originally Posted by bigshot View Post

There is absolutely no evidence whatsoever that sound above 20kHz adds anything to music. If it can be perceived at all, it's just pressure, not sound. Frequency extension is a red herring. In fact, the frequencies below 10kHz are more important than those above by an order of magnitude. There is very little above 10kHz in music, some weak harmonics and cymbal crashes. In piano music, pretty close to zilch except at the extreme high end of te keyboard. And that is just low level harmonics.
The highest note on a keyboard is about 4kHz. 16kHz would be plenty to cover the audible harmonics for that.

The most important band of frequencies to get right is from 100 Hz to 10 kHz - and I always prefer this range done right than more extendeed range done poorly.

 

But it is not only cymbal crashes - weak for sure, but important harmonics of say violin that help us differentiate between say an Amati or Stradivari or any violin for that matter, are almost all above 10 kHz.

 

It is low, really low levels of harmonics in piano going way higher than normally accepted - giving the illusion there is real instrument playing instead of reproduced one. When passing by a music school, you can always tell there is a real instrument playing - no CD with its limited bandwidth to 20 kHz can give you this kind of realism.

post #20 of 191
Quote:
Originally Posted by analogsurviver View Post

The most important band of frequencies to get right is from 100 Hz to 10 kHz - and I always prefer this range done right than more extendeed range done poorly.

 

But it is not only cymbal crashes - weak for sure, but important harmonics of say violin that help us differentiate between say an Amati or Stradivari or any violin for that matter, are almost all above 10 kHz.

 

It is low, really low levels of harmonics in piano going way higher than normally accepted - giving the illusion there is real instrument playing instead of reproduced one. When passing by a music school, you can always tell there is a real instrument playing - no CD with its limited bandwidth to 20 kHz can give you this kind of realism.

 

Maybe it is not the limited bandwidth or resolution of the CD, but how the CD was mastered...

post #21 of 191
Take an equalizer and do a highpass filter and find out what that stuff you seem to think is so vitally important actually sounds like. It ain't the difference between violins, I can guarantee you that. Above 20kHz it sounds like...












nothing!
Edited by bigshot - 12/20/12 at 5:39pm
post #22 of 191
Quote:
Originally Posted by Strangelove424 View Post

I agree that it's the recordings that suck not the transducer. I was appalled when I first got good equipment and heard what sounded like severe clipping on some Nora Jones songs (mainly from her early Come Away With Me album). The piano note attacks sound harsh and really badly distorted. But it does that at any volume, so it's not my equipment clipping. I'm no expert, but it sounds to me like someone wasn't watching the levels on their mic (and they were probably using just one).

 

Is that with headphones? I only got that in my car, and when I got a decent receiver with individual time alignment settings for each tweeter and midwoofer, it's almost completely gone; nothing like it on my headphones and home set-up (unless the noise floor is just much lower so I'm not cranking it up enough, especially with drivers right outside my ear canals).

post #23 of 191
Quote:
Originally Posted by ProtegeManiac View Post

 

Is that with headphones? I only got that in my car, and when I got a decent receiver with individual time alignment settings for each tweeter and midwoofer, it's almost completely gone; nothing like it on my headphones and home set-up (unless the noise floor is just much lower so I'm not cranking it up enough, especially with drivers right outside my ear canals).


I think Strangelove424 is referring to a recording problem dude. It would sound suboptimal through any decent rig...

post #24 of 191
Quote:
Originally Posted by ProtegeManiac View Post

 

Is that with headphones? I only got that in my car, and when I got a decent receiver with individual time alignment settings for each tweeter and midwoofer, it's almost completely gone; nothing like it on my headphones and home set-up (unless the noise floor is just much lower so I'm not cranking it up enough, especially with drivers right outside my ear canals).

 

Yep, headphones, DT880s. My speakers kind of gloss it over, but it’s still there somewhat. I hear it on Don’t Know Why on the first piano note, and it pops up a few more times throughout the album. It’s at dynamically challenging points, times she strikes down hard on the keys. Makes it sound like she’s on one of those toy pianos that has steel plates instead of strings - that is truly the best way I can describe it. In general though, that album doesn't sound well mastered or recorded to me anymore, not just the piano instrumentals. Her voice also sounds compressed and distorted on occasion. This was one of those albums that I wanted to dive into detail-wise with my headphones and when I finally got a chance to do that I was a bit disappointed.       

post #25 of 191

What makes piano sound so hard to reproduce?

The answer is harmonics.  If you double a frequency you get an octave.  A piano splits that octave into 12 steps.  If we were thinking like engineers we would split the frequency into 12 even steps or steps based on math.  Thats not what happened, when they invented the piano they realized some of the 12 steps sounded really bad when made even in terms of frequency. (HZ). They tried to play music like this for a while, its called even temperment and sounds horrible to our ears. 

 

PULL UP, reweind, for example if I call A 440HZ the octave up is 880Hz and all the notes in between fall in that range, you would think at a consistent rate.

 

So in the baroque period or about then musicians invented what we call the "well tempered" scale.  This means that the difference in HZ between each of these 12 steps is not equal, not in proportion or linear ratio.   Now think about that and think about how much more complex this makes the entire harmonic spectrum compared to an instrument that gets to play one note at a time at one frequency.  Pianos have to have all the notes sounding correct through all the octaves yet the octaves are split unevenly, i.e. there is more difference in HZ between some half steps than others. (a half step is the distance to the next closest key)

 

So an interval on a piano is not exactly correct in terms of HZ.  Its tuned by ear.  Some of the 12 even steps are Larger than others.  Just imagine the harmonic complexity, now add chords!

 

The piano is a beast and a great instrument to study sound and music theory and where they clash.  If you are someone who thinks in HZ you should really look into this.

 

The "well tempered" piano is the reason we have the entire harmonic system used in music today.  If the early piano makers evenly divided the octave among HZ and made even tempered pianos music would sound way way different today.  This is why strings tune to the piano...not only can't the piano retune but the intervals (distance between notes) are not consistent so string players (in chamber music) will slide their finger so its in tune.

 

This makes the overtones and the harmonics of a piano incredibly complex.  This is why piano tuning is a profession. 

 

Someone may have a better answer but this is my opinion based on my studies.

 

Quote:
 

Edited by pabloaugustus - 12/20/12 at 11:08pm
post #26 of 191
Quote:
Originally Posted by pabloaugustus View Post

What makes piano sound so hard to reproduce?

The answer is harmonics.  If you double a frequency you get an octave.  A piano splits that octave into 12 steps.  If we were thinking like engineers we would split the frequency into 12 even steps or steps based on math.  Thats not what happened, when they invented the piano they realized some of the 12 steps sounded really bad when made even in terms of frequency. (HZ). They tried to play music like this for a while, its called even temperment and sounds horrible to our ears. 

 

PULL UP, reweind, for example if I call A 440HZ the octave up is 880Hz and all the notes in between fall in that range, you would think at a consistent rate.

 

So in the baroque period or about then musicians invented what we call the "well tempered" scale.  This means that the difference in HZ between each of these 12 steps is not equal, not in proportion or linear ratio.   Now think about that and think about how much more complex this makes the entire harmonic spectrum compared to an instrument that gets to play one note at a time at one frequency.  Pianos have to have all the notes sounding correct through all the octaves yet the octaves are split unevenly, i.e. there is more difference in HZ between some half steps than others. (a half step is the distance to the next closest key)

 

So an interval on a piano is not exactly correct in terms of HZ.  Its tuned by ear.  Some of the 12 even steps are Larger than others.  Just imagine the harmonic complexity, now add chords!

 

The piano is a beast and a great instrument to study sound and music theory and where they clash.  If you are someone who thinks in HZ you should really look into this.

 

The "well tempered" piano is the reason we have the entire harmonic system used in music today.  If the early piano makers evenly divided the octave among HZ and made even tempered pianos music would sound way way different today.  This is why strings tune to the piano...not only can't the piano retune but the intervals (distance between notes) are not consistent so string players (in chamber music) will slide their finger so its in tune.

 

This makes the overtones and the harmonics of a piano incredibly complex.  This is why piano tuning is a profession. 

 

Someone may have a better answer but this is my opinion based on my studies.

 

A great post - thank you for this explanation.

 

My experience with piano tuners is the following : the minute after the rehearsal for the evening concert is finished, when all I want is peace and quiet for about a couple of minutes, to find if there are any vestiges of hum etc in the recording setup, a piano tuner appears and starts "banging" on that piano like there is no tomorrow. It is a hard profession - more often than not, piano tuners are called in only after pianist(s) declared the piano unfit for the concert - piano tuners cost money and organizers of concerts tend to cut expenses wherever possible. That they generally do menage to bring the piano up to desired tuning under such conditions and under so short notice is quite extraordinary.

 

Any audio device will alter that complex harmonic structure to a point - some more, some less -  the better the master, less possibility for something to go wrong later in the (post)production. So, I record with the greatest bandwidth I can possibly lay my arms on. Better to correctly record "nothing" and later throw that away than completely fail to record that last n-th harmonic in the first place - in that case, it is lost for good, no after the fact can bring it back.

 

And to wet anyone's appetite for high frequencies above 20 kHz - when recording harpsichord about a year ago, a fellow recording engineer put that recording on a very good spectrum analyzer that goes far higher than 20 kHz. Harpsichord clearly had harmonics up to 30 - 40 kHz or there about, it was not constant HF ( above audio ) noise that is the unfavourable characteristic of DSD, it was in tune with the notes being played. The difference in sound from the DSD master and issued CD is quite marked - provided that amplification and speakers or headphones used support frequency response extended way above 20 kHz.

post #27 of 191
Quote:
Originally Posted by ultrabike View Post

 

Maybe it is not the limited bandwidth or resolution of the CD, but how the CD was mastered...

Up to a point. CD resolution is "enough" - but only if recording "direct to CD-R", no further mastering of any kind ( except for editing the recording to movements, songs etc ). And provided music does not exceed the dynamic range of 96 dB - no go for Mahler's 2nd symphony, where there are jumps in dynamic range from the noise floor of the recording venue ( musicians/singers are dead silent at few spots within the composition ) to practically pain threshold ( in the vicinity of 120 dB, depending on orchestra and choir, hall etc, but all the stops are usually pulled out in the finale ). 

 

Very good recordings are possible using this "direct to CD-R" technique - but they still pale in comparison with DSD recording made paralel to it, both recorders fed from the same mike/preamp setup. DSD has 120 dB ( or there about, an odd dB up or down ) dynamic range and extends fairly flat to at least 50 kHz, with GENTLY rolled off response continuing past the 100 kHz . In case od DSD, the limits of microphone noise and frequency response are likely to set in before DSD recorder runs out of steam.


Edited by analogsurviver - 12/21/12 at 12:42am
post #28 of 191
Quote:
Originally Posted by pabloaugustus View Post

What makes piano sound so hard to reproduce?

The answer is harmonics.  If you double a frequency you get an octave.  A piano splits that octave into 12 steps.  If we were thinking like engineers we would split the frequency into 12 even steps or steps based on math.  Thats not what happened, when they invented the piano they realized some of the 12 steps sounded really bad when made even in terms of frequency. (HZ). They tried to play music like this for a while, its called even temperment and sounds horrible to our ears. 

 

PULL UP, reweind, for example if I call A 440HZ the octave up is 880Hz and all the notes in between fall in that range, you would think at a consistent rate.

 

So in the baroque period or about then musicians invented what we call the "well tempered" scale.  This means that the difference in HZ between each of these 12 steps is not equal, not in proportion or linear ratio.   Now think about that and think about how much more complex this makes the entire harmonic spectrum compared to an instrument that gets to play one note at a time at one frequency.  Pianos have to have all the notes sounding correct through all the octaves yet the octaves are split unevenly, i.e. there is more difference in HZ between some half steps than others. (a half step is the distance to the next closest key)

 

So an interval on a piano is not exactly correct in terms of HZ.  Its tuned by ear.  Some of the 12 even steps are Larger than others.  Just imagine the harmonic complexity, now add chords!

 

The piano is a beast and a great instrument to study sound and music theory and where they clash.  If you are someone who thinks in HZ you should really look into this.

 

The "well tempered" piano is the reason we have the entire harmonic system used in music today.  If the early piano makers evenly divided the octave among HZ and made even tempered pianos music would sound way way different today.  This is why strings tune to the piano...not only can't the piano retune but the intervals (distance between notes) are not consistent so string players (in chamber music) will slide their finger so its in tune.

 

This makes the overtones and the harmonics of a piano incredibly complex.  This is why piano tuning is a profession. 

 

Someone may have a better answer but this is my opinion based on my studies.

 

It's not that "some of the 12 steps sounded bad". It has to do with the combination of these steps, which sometimes resulted in impure harmony. An equal temperament will yield a perfect fifth (C to G) but has an awful M3 (C to E). What is great about equal temperament is that it allows one to play or modulate to many keys, and all these keys will sound "tolerable" to the ears. Of course equal temperament cannot compete with unequal temperaments when playing pieces that have G major or C major keys (just to name a few). This is the reason why so many Baroque pieces have the typical major/ minor keys, rather than something like C sharp major or G sharp minor. In short, the equal temperament is the jack of all trades. Unequal temperaments are great with a few keys, awful with others. 

 

There are unequal temperaments that somewhat rival equal temperaments, and to list one: the well-temperament. The well-temperament is a temperament designed by J.S. Bach so that all keys will sound tolerable to the ears (like equal temp.) but will still maintain their own unique color (unlike equal temp., which divides every semitone / half-step equally, making every key sound the same). 

 

The topic of the overtones is even more complex than the topic of chords/ temperament. I do not know enough about the harmonic series to discuss it in detail.

 

I believe headphones are good at reproducing intonation, meaning how harmonically complex a piano is due to temperament isn't the main issue.  I think the real issue here is timbre, which is clearly much harder to reproduce. 

post #29 of 191

Quote:
Originally Posted by pabloaugustus View Post

What makes piano sound so hard to reproduce?

The answer is harmonics.  If you double a frequency you get an octave.  A piano splits that octave into 12 steps.  If we were thinking like engineers we would split the frequency into 12 even steps or steps based on math.  Thats not what happened, when they invented the piano they realized some of the 12 steps sounded really bad when made even in terms of frequency. (HZ). They tried to play music like this for a while, its called even temperment and sounds horrible to our ears. 

 

PULL UP, reweind, for example if I call A 440HZ the octave up is 880Hz and all the notes in between fall in that range, you would think at a consistent rate.

 

So in the baroque period or about then musicians invented what we call the "well tempered" scale.  This means that the difference in HZ between each of these 12 steps is not equal, not in proportion or linear ratio.   Now think about that and think about how much more complex this makes the entire harmonic spectrum compared to an instrument that gets to play one note at a time at one frequency.  Pianos have to have all the notes sounding correct through all the octaves yet the octaves are split unevenly, i.e. there is more difference in HZ between some half steps than others. (a half step is the distance to the next closest key)

 

So an interval on a piano is not exactly correct in terms of HZ.  Its tuned by ear.  Some of the 12 even steps are Larger than others.  Just imagine the harmonic complexity, now add chords!

 

The piano is a beast and a great instrument to study sound and music theory and where they clash.  If you are someone who thinks in HZ you should really look into this.

 

The "well tempered" piano is the reason we have the entire harmonic system used in music today.  If the early piano makers evenly divided the octave among HZ and made even tempered pianos music would sound way way different today.  This is why strings tune to the piano...not only can't the piano retune but the intervals (distance between notes) are not consistent so string players (in chamber music) will slide their finger so its in tune.

 

This makes the overtones and the harmonics of a piano incredibly complex.  This is why piano tuning is a profession. 

 

Someone may have a better answer but this is my opinion based on my studies.

 

 

Thanks! I honestly didn't (and still don't) know much about the complexity that goes into the sound of a piano note. I'm not a piano expert, and I have a lot of respect for the artists and craftsmen that produce and play this and other instruments. However, note that a decent headphone should be able to produce all frequencies from maybe 80Hz to 20kHz at the same time with little distortion. TOTL headphones might do so all from 20Hz to maybe 30kHz at the same time, with distortion levels bellow 1%.

 

 

Quote:
Originally Posted by analogsurviver View Post

Up to a point. CD resolution is "enough" - but only if recording "direct to CD-R", no further mastering of any kind ( except for editing the recording to movements, songs etc ). And provided music does not exceed the dynamic range of 96 dB - no go for Mahler's 2nd symphony, where there are jumps in dynamic range from the noise floor of the recording venue ( musicians/singers are dead silent at few spots within the composition ) to practically pain threshold ( in the vicinity of 120 dB, depending on orchestra and choir, hall etc, but all the stops are usually pulled out in the finale ). 

 

Very good recordings are possible using this "direct to CD-R" technique - but they still pale in comparison with DSD recording made paralel to it, both recorders fed from the same mike/preamp setup. DSD has 120 dB ( or there about, an odd dB up or down ) dynamic range and extends fairly flat to at least 50 kHz, with GENTLY rolled off response continuing past the 100 kHz . In case od DSD, the limits of microphone noise and frequency response are likely to set in before DSD recorder runs out of steam.

 

 

DSD suffers from the limitations of requiring a single-bit delta sigma. It's not a bad technology, but it does have it's limitations. Also, I'm not a remastering engineer, but doesn't the remastering process extend the precision to more than 16 bits during editing and later round back down to 16 bits if necessary? If the song has a dynamic range of 120 dB, can you hear the passages that fall bellow 96 dB from the peak?


Edited by ultrabike - 12/21/12 at 1:38am
post #30 of 191
Quote:
Originally Posted by analogsurviver View Post
Any audio device will alter that complex harmonic structure to a point - some more, some less -  the better the master, less possibility for something to go wrong later in the (post)production. So, I record with the greatest bandwidth I can possibly lay my arms on. Better to correctly record "nothing" and later throw that away than completely fail to record that last n-th harmonic in the first place - in that case, it is lost for good, no after the fact can bring it back.

Scnr, but instead of going for greatest bandwidth you should concentrate on recording stuff so that it doesn't clip.

 

There's nothing wrong with recording at higher sample rates and it is de facto standard to use higher bit depths, but such formats are not needed for playback. As I've said before, the stuff up there is extremely low in level, easily masked, mostly noise and most importantly: inaudible.

 

Quote:

And to wet anyone's appetite for high frequencies above 20 kHz - when recording harpsichord about a year ago, a fellow recording engineer put that recording on a very good spectrum analyzer that goes far higher than 20 kHz. Harpsichord clearly had harmonics up to 30 - 40 kHz or there about, it was not constant HF ( above audio ) noise that is the unfavourable characteristic of DSD, it was in tune with the notes being played. The difference in sound from the DSD master and issued CD is quite marked - provided that amplification and speakers or headphones used support frequency response extended way above 20 kHz.

How did you determine that these harmonics were not bursts of noise and there was no constant HF noise? Because I just checked a 24/96 recording of a harpsichord and the HF rms amplitude is -80 dB on average to well below -120 dB minimum. It IS constant HF noise plus bursts of noise.

 

You can repeat your claims of the differences you hear between DSD and CD all you want, it's not going to change anything. Again, do you have anything to back up your claims other than anecdotes?

 

Quote:
Originally Posted by analogsurviver View Post

Up to a point. CD resolution is "enough" - but only if recording "direct to CD-R", no further mastering of any kind ( except for editing the recording to movements, songs etc ). And provided music does not exceed the dynamic range of 96 dB - no go for Mahler's 2nd symphony, where there are jumps in dynamic range from the noise floor of the recording venue ( musicians/singers are dead silent at few spots within the composition ) to practically pain threshold ( in the vicinity of 120 dB, depending on orchestra and choir, hall etc, but all the stops are usually pulled out in the finale ).

Seriously, where are you pulling those claims out of? The dynamic range of concert halls doesn't exceed 80 dB (Eargle, John; 2005; Handbook of Recording Engineering) and all the recordings I checked of Mahler's 2nd do rarely exceed 70 dB.

 

16 bit has clearly enough dynamic range. Again a comparison to vinyl: "Under no legitimate circumstances will the dynamic range of vinyl ever exceed the dynamic range of CD, under any frequency, given the wide performance gap and the physical limitations of vinyl playback."

 

Quote:
Very good recordings are possible using this "direct to CD-R" technique - but they still pale in comparison with DSD recording made paralel to it, both recorders fed from the same mike/preamp setup. DSD has 120 dB ( or there about, an odd dB up or down ) dynamic range and extends fairly flat to at least 50 kHz, with GENTLY rolled off response continuing past the 100 kHz . In case od DSD, the limits of microphone noise and frequency response are likely to set in before DSD recorder runs out of steam.

PCM recording is usually done with 24 bits and possibly higher sample rates (88.2, 96 or even 192 kHz). You do not have to record directly to CD-R. All you need is a (quality) resampler and dither for the conversion to the CD format.

 

 

Quote:
Originally Posted by ultrabike View Post
DSD suffers from the limitations of requiring a single-bit delta sigma. It's not a bad technology, but it does have it's limitations. Also, I'm not a remastering engineer, but doesn't the remastering process extend the precision to more than 16 bits during editing and later round back down to 16 bits if necessary? If the song has a dynamic range of 120 dB, can you hear the passages that fall bellow 96 dB from the peak?

Yes, editing and even recording is done at higher bit depths than 16 bit (DSP uses as high as 80 bit floating point). Yes, you can hear stuff below the dither noise floor, you can for example hear a 1 kHz sine wave at -100 dBFS quite clearly on a properly dithered 16-bit file.


Edited by xnor - 12/21/12 at 9:08am
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?