If the noise floor is 90-odd dB below max level, 16 bits will not be the limitation- take a look at s/n of preamps, power amps, and mikes.
when looking on an instant signal the lowest and highest spl played in a fraction of time: do we have to take the S/N of the speakers with its 2nd, third, etc order harmonics (around -40 dB with a good driver !) Does this dynamic dB gap has to be taken in the equation ?
Well I have also a 85 dB efficienty speaker, so less sensible to the noise than a 100 dB horn system with the noise floor !
What is saying the datasheet of my pre : CD output S/N : 110 dB with 0.002 H distorsion (1.5V) between 20 Hz & 20K Hz
What is saying the datasheet of my Amp : better than 103 dB !
So a 103 dB S/N dac is good enough for the hifi as the amp is the limit ?? This is it ?
Thanks to help me to translate all of that in real life 🙂
(what is the limit of the ears to distinct two sounds in the shortest time width ? Millisecond ?)
Well I have also a 85 dB efficienty speaker, so less sensible to the noise than a 100 dB horn system with the noise floor !
What is saying the datasheet of my pre : CD output S/N : 110 dB with 0.002 H distorsion (1.5V) between 20 Hz & 20K Hz
What is saying the datasheet of my Amp : better than 103 dB !
So a 103 dB S/N dac is good enough for the hifi as the amp is the limit ?? This is it ?
Thanks to help me to translate all of that in real life 🙂
(what is the limit of the ears to distinct two sounds in the shortest time width ? Millisecond ?)
On the other hand so do bit-crushing guitar pedals.
As far as I know they introduce distortion by sampling using low bit rates but some also allow to vary the sample rate to reduce frequency response.
They may however sample at 24bit and then truncate without applying dither.
But then there was the Emulator II which was a lovely machine but the reason that whatever you recorded into it sounded not much like the real thing was not the noise level but the fact that it sampled at 8bit resolution.
If you drop down to 6bit PCM the sound quality of a Speak'n'Spell is about as good as it gets.
That said from everything I've read so far it seems that 16bit is a very adequate resolution for mass distribution.
As far as I can tell bit depth determines amplitude resolution AND noise level while the sample rate determines timing resolution AND frequency response.
Once you get above a certain point in either the resolution in amplitude or timing starts to exceed the abilities of human perception and one would need another good (or not so good) reason to push higher still.
Being generous I'd say that even the best trained and genetically gifted human ears are adequately covered resolution-wise by a 20bit 48kHz digital system and beyond 24/96 it becomes essentially a d**k-measuring contest.
@ Charles : Hey, good, I have two 20 bits AD1862 dac chips on the shelves 🙂 !
So I understand the more bits, it becomes not the worst as well to read you all, but not only needed like for instance 32 bits if you are not processing digital attenuation to swap a pre!
So well : 16 bits is more or less good enough, while 18, 20 or 24 bits chips will not hurt. And if I want to hear the 130 dB peak with 35 dB noise flor of the concert room, better for me to go Salle Pleyel or Opera than trying to get it at real volume (I speak of course of instant peaks, not trying to copy CapCanaveral launchings i.e 130 dB continuous)
So I understand the more bits, it becomes not the worst as well to read you all, but not only needed like for instance 32 bits if you are not processing digital attenuation to swap a pre!
So well : 16 bits is more or less good enough, while 18, 20 or 24 bits chips will not hurt. And if I want to hear the 130 dB peak with 35 dB noise flor of the concert room, better for me to go Salle Pleyel or Opera than trying to get it at real volume (I speak of course of instant peaks, not trying to copy CapCanaveral launchings i.e 130 dB continuous)
Last edited:
But given the number of non D-S DACs available now outside of instrumentation applications it's safe to say that low bit is proven to work! Seems industry has standardised on 6bits.
Don't forget the Fairlight CMI! But those were 8bit log DACs. Not sure who, but when the 16-bit series 3 came out someone complained it had lost the 'rock and roll' of the earlier models.
Noise, yes, amplitude resolution no unless the mastering used noiseless mikes and boards AND the engineer and mastering lab were both remarkably inept.
Yes, you'd be hard pressed to find one with better than 90dB S/N, and generally, it's a lot worse. To put in perspective, the best phono cartridges MIGHT have -80dB noise before being lowered into the record groove.
I live with 104dB efficiency horns, a Korn&Macway SP100 preamplifier (ultra rare) and power amp duties go to a heavily modded Trends TA10.1 (all caps, all resistors, and potentiometer upgraded). Max power is 2 x 8watts; I use a fraction of a watt average on most listening (80dB @ listening seat)
I would have a hard time living with 85dB/ w/m speakers 🙂
I think you will find that the antialiasing filter sets the timing resolution. The sample rate then has to be fast enough; making it faster than 'enough' adds no further timing resolution. Similarly for the frequency response.Charles Darwin said:
Totally off topic sorry :
If you were be Amped, your bass aera will be happier with at least 20W (on peaks demands) ; look at the blog section of Pano member : there is an excellent paper about that writed by la Revue du Son (assuming you speak french as well) : the 100 dB 360 Onken needs 50 W for the most demanding transcient peaks; so if you are 103 dB, you should need in the best of the world around 20 W !
I close the ( ) !😀
But I'm totally down : knowing our hifis are mainly rulled by the noise floor ! Me throwing all to the garbadge and go more often to live events !
I saw measurements where the conductor of a Philharmonic when heeping the concert room quiet (spectators) have around 35 dB ! On some partition some peaks transcient can reach max 130 dB (it is not said if it is measured at the conductor place or middle of the room) !
130 db -35 db = 95 dB = close enough to the previous conversation ?
Do I putt the good math to compare with dB attenuation of a Hifi with a pot ? I just want to know if the difference of dynamic between the highest and lowest dynamic in short time (dynamic window) is good with a hifi like with an event (but I know it may varry because the F&M curve which change with spl pressure !)
Well if the concert middle level is for instance 85 dB ; the highest peaks at 125 dB most of the time with a nervous conductor : dynamic Windows is mainly 40 dB (I don't take here the spl below the midle level because they are rarely close on a partition) ! So if I listen at home at low level, for instance 60 dB, I know the dynamic gap will be at max 100 dB...
I have pain for the people of home theater whom want to reach THX !
Am I in the myth aera because of non understanding (which could be the fact!) ? Maybe I don't need a 100 dB spl speakers with a 50W Kaneda amp ? I don't understand finely when the noise floor is becomming a pain when talking about dynamic and DAC (lazy am I, I don't lissen too much my Rega Planar III but time to time !).
Well I believe many here answered already to that....
So, Thank you.
130 db -35 db = 95 dB = close enough to the previous conversation ?
Do I putt the good math to compare with dB attenuation of a Hifi with a pot ? I just want to know if the difference of dynamic between the highest and lowest dynamic in short time (dynamic window) is good with a hifi like with an event (but I know it may varry because the F&M curve which change with spl pressure !)
Well if the concert middle level is for instance 85 dB ; the highest peaks at 125 dB most of the time with a nervous conductor : dynamic Windows is mainly 40 dB (I don't take here the spl below the midle level because they are rarely close on a partition) ! So if I listen at home at low level, for instance 60 dB, I know the dynamic gap will be at max 100 dB...
I have pain for the people of home theater whom want to reach THX !
Am I in the myth aera because of non understanding (which could be the fact!) ? Maybe I don't need a 100 dB spl speakers with a 50W Kaneda amp ? I don't understand finely when the noise floor is becomming a pain when talking about dynamic and DAC (lazy am I, I don't lissen too much my Rega Planar III but time to time !).
Well I believe many here answered already to that....
So, Thank you.
Last edited:
Grab the best classical recordings you can find and measure the dynamic range (using software like foobar dynamic range plug in or dynamicsanalysis). That 95dB stars to look pretty safe. 😀 And, of course, vinyl or tape aren't even close. Ditto unweighted noise from mikes.
Noise shaping reduces the audible floor even more. TBH, I have never heard anyone complain that CDs are too noisy.
Noise shaping reduces the audible floor even more. TBH, I have never heard anyone complain that CDs are too noisy.
Same thing happened with delays.
When there were only tape delays everybody moaned about the diminishing sound quality.
They developed BBDs but it still wasn't good enough so along came digital delays which did not degenerate the sound with each repetition and suddenly they are 'too clinical' and 'anodyne'.
Thing is when it comes to (playing not replaying) music we need a minimum level of dirt, inaccuracies and other surprises to keep us interested.
The only thing as boring as perfection is pure chaos, the interesting stuff happens in the region of transition.
PS: Aren't the functioning 1bit systems based on PWM rather than PCM?
Yes clearly most of the audiophiles don't know how the noise translates here : if this is for instance distorsion with worst mid-trebles (like a feeling of no wanted odd harmonics)... it's not a buzzzzzzzzzzzzz indeed !
May be you'd like to correct this wiki page then:
https://en.wikipedia.org/wiki/Digital-to-analog_converter#DAC_types
"DAC performance[edit]
DACs are very important to system performance. The most important characteristics of these devices are:
Resolution
The number of possible output levels the DAC is designed to reproduce. This is usually stated as the number of bits it uses, which is the base two logarithm of the number of levels. For instance a 1 bit DAC is designed to reproduce 2 (21) levels while an 8 bit DAC is designed for 256 (28) levels. Resolution is related to the effective number of bits which is a measurement of the actual resolution attained by the DAC. Resolution determines color depth in video applications and audio bit depth in audio applications.
Dynamic range
A measurement of the difference between the largest and smallest signals the DAC can reproduce expressed in decibels. This is usually related to resolution and noise floor."
This one is a good and interesting reading.
Dick Burwen music and studio room,where he tried to make the most realistic sound,with recording in the same room and get the same playback after it. It was build in the early 80's so that should take in to account. I'm pretty happy if I like the sound of my own system, never really bothered to make it 100% perfect.
Dick Burwen's Sound System
by the way I do listen a bit of moded Revox B225 14 bit (tda1540) and DacEnd2 AD1865.
Dick Burwen music and studio room,where he tried to make the most realistic sound,with recording in the same room and get the same playback after it. It was build in the early 80's so that should take in to account. I'm pretty happy if I like the sound of my own system, never really bothered to make it 100% perfect.
Dick Burwen's Sound System
by the way I do listen a bit of moded Revox B225 14 bit (tda1540) and DacEnd2 AD1865.
Last edited:
- Status
- Not open for further replies.