a 1 bit sigma-delta DAC can be used (and IS used often) to play 16-bit samples; by using oversampling, the 1-bit DAC at 256x oversampling performs as a 16-bit dac at 1x sampling rate.
I was considering a hypothetical track, where everything is put onto the CD with no sliders adjusted.
Track starts with someone singing softly, and then, further on, there's a drum solo, using a full-size stage kit, with a very heavy-handed drummer.
The drums might then quieten down to a more soft-handed rhythm again for the track to continue.
I'd estimate at least 50dB of dynamic range would be needed here, in order to reproduce the full dynamic swing in volume that was there originally.
There's no option but to make the quiet bits louder and the louder bits quieter. The system I played a drum kit recording through (mentioned a couple of pages back) would manage the full dynamics of that track, but using almost all of the 600w available into a 98dB@1w speaker isn't an option for domestic listening - its bonkers loud, and would likely deafen the listener in due course.
Chris
Thanks Flavio for the explanation, im really a novice in the field of ADC and DAC. Most of my experience is with PIC micros and coding to use the ADC functions. DACs are something that ive rudimentary theory. Ill have to read up to learn how they achieve 65536 divisions using 1 bit and x256 over sampling. I could probably use a 10bit ADC on a cheap PIC that im familiar with. Others have better ADCs, but again, id have to look in a book or two.
Last edited:
Oversampling is what you want... The main reason the CD player industry went to oversampling (first implemented as 4x the sampling rate on the first generation Philips CD players, which had a respectable 14-bit DAC chip) was that if the sampling frequency F s higher, then the images (artifacts produced by the DAC starting from frequency F/2 and higher) will be moved further away from the audible spectrum. This means that it will be easier to filter them out using analog filters.
When you go to the extreme (1-bit DAC and 256x oversampling), the problem is that any jitter on the sampling clock could potentially bring more trouble that the same jitter at 1x (no oversampling).
Or at least that's how i understood it.
Yes, but you have saved the driver to play again another day. Very important in the biz and why it's designed that way. A progressive suspension can you keep you out of trouble when the fellow behind the console has had a few too many. And that's pretty often.
Of course you can still burn up the voice coil.
I'll take your word on the figures, but I've still not grasped why the situation of listening at home is any different from listening to a live performance from a location that gives a similar SPL, e.g. in the audience in a classical concert. In the concert hall the listener is subjected to the full dynamic range yet can still hear the quiet bits and isn't deafened by the loud bits. There are differences of course, and maybe the listener in the concert hall has more information to work with which allows him to hear past his fellow coughing, shuffling audience members better, but at home he'll be listening in a quieter environment. His audio system can happily handle 90dB of dynamic range, say. Why is it imperative that the recording is compressed dynamically?
In the UK we have two large classical stations: Radio 3 and Classic FM. Classic FM is highly compressed and gives a superficially smooth presentation that, at first, sounds very sweet. But, after a little while it leaves you gasping for dynamics.
This compression or extreme limiting is what i was referring to earlier. In my experience radio is over compressed and DAB is no better, id guesstimate that its worse. Even with CD the compression is lesser, probably vinyl is also better (once you take out the dynamic range lost to rumble and crackle).
To the best of my knowledge 40db is a reasonable goal for vinyl or cd dynamic range, and a well executed 2:1 compression from 80db to 40db isnt half as objectionable.
Frankly i find the bandwidth of DAB through DTV box to be severely lacking on all but the flagship BBC stations, let alone the compression, and i liken it to 64kbit mp3 quality (i.e. Not listenable)
To the best of my knowledge 40db is a reasonable goal for vinyl or cd dynamic range, and a well executed 2:1 compression from 80db to 40db isnt half as objectionable.
Frankly i find the bandwidth of DAB through DTV box to be severely lacking on all but the flagship BBC stations, let alone the compression, and i liken it to 64kbit mp3 quality (i.e. Not listenable)
Last edited:
SY is correct, if splitting hairs is necessary. Obviously i was talking about log/log ratio and NOT absolute numerical ratio. But then people here are quick to point out others error (myself included) while considering themselves infallible (again myself included). I work with many with a SOH like this, so i got a hide like a crocodile.
for the easily audible bits.
If the noise floor whether in the room or in the recording is a further 20dB to 40dB below that results in a dynamic range of 60dB to 80dB.
If the recording goes silent I expect it to "sound" silent. Then I can hear when it all starts up again. That starting up could be well below your arbitrarily set value of 40dB for dynamic range or it could be within your 40dB.
Rooms can easily achieve 70dB of dynamic range and some may exceed a range 90db, (max signal to ambient noise).
Recordings can also match that achievable range.
Why should we settle for less?
Last edited:
I wouldn't call 34dB splitting hairs, but OK.
Andrew, we settle for less because we have to. Recordings are not made for us exclusively. They have to sell people like my mother who, although she attends concerts regularly, still listens to jazz and classical on a small radio/CD player - and likes it. I see people ride the volume control of the radio and the TV all the time. For them, there isn't enough compression.
I'm with you, I'd like wider dynamic range. But apart from a few "Audiophile" type masters, we won't get it.
I have heard a few recordings of concerts I've attended with little or no compression applied. I liked them.
I ride the TV volume all the time because there is too much compression!
Watching movies is fine but the ads in the breaks are very heavily compressed so that the rms level is much higher than the movie soundtrack so I have to turn the volume down or off during ad breaks.
I had similar issues with the contents of my itunes folder (I have it on shuffle nearly all of the time) but some of the cds I ripped are heavily compressed and I had to turn the volume down for those.
Now I just have the offending tunes (Green Day, The Stooges 'Raw Power' etc) permanently muted.
Last edited:
CDs claim 96dB of signal to noise, that is at the top end of what we need to hear silence in the silent bits and achieve very loud, when the music requires it.
It also happens to fit the model I stated
We don't need a "noise war". The recording/playback mediums we have are easily capable of giving very good dynamic range. Then the user/operator can choose how loud with the vol knob.
It also happens to fit the model I stated
We don't need a "noise war". The recording/playback mediums we have are easily capable of giving very good dynamic range. Then the user/operator can choose how loud with the vol knob.
Last edited:
Or now, in the day of ever cheaper DSPs, the CD itself could be uncompressed and it's up to the user how much compression they want to use!
We don't need the loudness war, but we got it anyway.
The mute button is the first thing to wear out on the TV remote at our house.
that is the most annoying thing about TV through loudspeakers. Youd think it wasnt difficult to get the levels matched
AndrewT,
Regarding your comments of dynamic range possible in room, i largely agree.
However, unless ive forgotten a pivotal fact, the threshold of hearing also comes into play, and is subtractive on the total dynamic range.
I cant recall if the threshold is 20db or 30db, but with a reproduction level of 90db youre left with 60db to 70db of AUDIBLE range at the bottom of that scale. I.e ive never cranked an amp with an SNR of 100db to a level where i could hear the noisefloor of the amp alone. But then id have to be at 120db.
The same applies here, if totally uncompressed music with a range of 96db is played at equivalent SPL, 20db of that is lost below audibility. Play quieter and more low level signal is inaudible, play louder and more is revealed but now realistic amplitude is lost (as also for reduced levels)
Assuming 20db is correct for that threshold, im running to look it up as i speak, save the accuracy police the job of pointing out my hand waving minds eye approximation in speech.(or text)
AndrewT,
Regarding your comments of dynamic range possible in room, i largely agree.
However, unless ive forgotten a pivotal fact, the threshold of hearing also comes into play, and is subtractive on the total dynamic range.
I cant recall if the threshold is 20db or 30db, but with a reproduction level of 90db youre left with 60db to 70db of AUDIBLE range at the bottom of that scale. I.e ive never cranked an amp with an SNR of 100db to a level where i could hear the noisefloor of the amp alone. But then id have to be at 120db.
The same applies here, if totally uncompressed music with a range of 96db is played at equivalent SPL, 20db of that is lost below audibility. Play quieter and more low level signal is inaudible, play louder and more is revealed but now realistic amplitude is lost (as also for reduced levels)
Assuming 20db is correct for that threshold, im running to look it up as i speak, save the accuracy police the job of pointing out my hand waving minds eye approximation in speech.(or text)
Last edited:
Nice discussion, a bit sidetracked but interesting.
Regarding dynamics contents, for music and relative to compression and 'loudness war' it seems that most mastering engineer, at this time,
try to go back a (big) bit about rms levels.
Bob Katz (Digital Domain mastering studio owner/engineer and a great professor with everything mastering related) gave a proposition to limit
over compression in mastering stage: K-system. Given the music style and destination (CD, Broadcast) he give three reference meter calibration K-20, K-14
and K-12 (repsectively for acoustic music (classical), pop music and broadcast materials).
Basically it's a calibration procedure of monitor system, -20dfs pink noise generated to give 83db spl for each speaker at sweet spot (listening position),
giving a way to verify the loudness of material and judge it's subjective level. Very effective way of work i find.
Only drawback of principle (in case of K-20 e.g.) a symphonic orchestra forte could sound at same 'level' than an flute solo if two tracks are put
side by side... But it has numerous advantage: you know which dynamic margin you use for your playback system (max peak are 20db over the
reference level thus 106db for K-20 at listening position ) direct comparison of different styles in using the volume control giving info on macro/micro
dynamics in a recording, repeatability,etc,etc...
Well 20db margin can seems low dynamics for loud passages for classical but in practice and as Pano noted for 'johnny six-pack' it's enough...
IIf you want zero compression, high quality recordings, go Chesky records... But you need an 'hell good' playback system for those recording to
sound well (and like the style, mellow Jazz).
For radio sound compression (or broadcast in general)situation is a two way path: first technical: in analogue fm radio you NEED compression or you'll cover stations near your
numbers (107.5 and 106.8 e.g.).
Second path psychoacoustic to help business: the louder the more 'hooked' are listeners to radio station. The more captive the listeners, the more
money you can sell commercials... Same in TV.
By now that extremes had been reached for many years (<6db rms nearly white noise) and nuisance are greater than gains (and with digital broadcast media)
new norms appears like ITU-R BS.1770 / EBU R128 to limit this (especially for TV). But it's long way to go back...
Well sorry for the OT...
Let's go back to original subject: does anyone use FIR in active config? What are advantages of Brickwall filters compared to classic configuration?
Are they really worth the money and which method to parameter filters like this?
Regarding dynamics contents, for music and relative to compression and 'loudness war' it seems that most mastering engineer, at this time,
try to go back a (big) bit about rms levels.
Bob Katz (Digital Domain mastering studio owner/engineer and a great professor with everything mastering related) gave a proposition to limit
over compression in mastering stage: K-system. Given the music style and destination (CD, Broadcast) he give three reference meter calibration K-20, K-14
and K-12 (repsectively for acoustic music (classical), pop music and broadcast materials).
Basically it's a calibration procedure of monitor system, -20dfs pink noise generated to give 83db spl for each speaker at sweet spot (listening position),
giving a way to verify the loudness of material and judge it's subjective level. Very effective way of work i find.
Only drawback of principle (in case of K-20 e.g.) a symphonic orchestra forte could sound at same 'level' than an flute solo if two tracks are put
side by side... But it has numerous advantage: you know which dynamic margin you use for your playback system (max peak are 20db over the
reference level thus 106db for K-20 at listening position ) direct comparison of different styles in using the volume control giving info on macro/micro
dynamics in a recording, repeatability,etc,etc...
Well 20db margin can seems low dynamics for loud passages for classical but in practice and as Pano noted for 'johnny six-pack' it's enough...
IIf you want zero compression, high quality recordings, go Chesky records... But you need an 'hell good' playback system for those recording to
sound well (and like the style, mellow Jazz).
For radio sound compression (or broadcast in general)situation is a two way path: first technical: in analogue fm radio you NEED compression or you'll cover stations near your
numbers (107.5 and 106.8 e.g.).
Second path psychoacoustic to help business: the louder the more 'hooked' are listeners to radio station. The more captive the listeners, the more
money you can sell commercials... Same in TV.
By now that extremes had been reached for many years (<6db rms nearly white noise) and nuisance are greater than gains (and with digital broadcast media)
new norms appears like ITU-R BS.1770 / EBU R128 to limit this (especially for TV). But it's long way to go back...
Well sorry for the OT...
Let's go back to original subject: does anyone use FIR in active config? What are advantages of Brickwall filters compared to classic configuration?
Are they really worth the money and which method to parameter filters like this?
Not the CD but digital format allow for the use of metadata including infos relative to rms/peak level. But and it's a big but, no distributors or hardware corporation wants to find an agreement to use them...
Personnal mastering is not really an answers: skills needed to do something correct (listenable) are great and difficult to gain...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.