Speaker "compression" is due to voice coil heating.
High average power heats the coil, raises impedance, the amplifier delivers less power.
The same effect will occur whether the crossover is active or passive, as long as the same average power is delivered.
Thats interesting. I have several similar results in CEpro2.0, but I have also many that show far better than this.
The track I mentioned in a previous post, has a range from 0dBfs to>-21dBfs.
I actually ripped the CD and went to the trouble of print screening it, only to find that my gallery still isnt working......
No!
They absolutely have to be present.
Otherwise we'd all be deaf as soon as a loud drummer comes along.
.. or we'd miss most of the music.
Without the peaks being incredibly loud (>120dB), the lower end of the music could well fall into the noise floor.
Someone singing softly could be as low as 60-70dB. A drum kit played hard can pass 120dB.
Turn the drum kit down to the ~80dB many of us listen at, and the singer that was singing softly is down at 20-30dB, being overpowered by distant traffic.
There is no option but to use some compression to bring the quietest and loudest bits closer together.
Chris
I've analyzed 1000s of tracks. -18dB RMS is typical of good material, some rock, jazz and especially classical CDs have more - 25 to 30dB average to peak inside one track. I have a hard time with recent recordings, they are just loud all the time. Even something as gentle as "The Civil Wars" has the life compressed out of it.
OK, sorry about the OT.
OK, sorry about the OT.
That's a form of compression, but I would like to know if there are other mechanisms at play, perhaps less well known. As has been stated, a driver's impedance changes dependent on the signal it's being driven with, or the recent history of the signal. Combine that with a series impedance (i.e. the crossover), and does that produce effects that you don't get with an active system, for example?
Is this a job for compression, or simply setting the sliders to the right levels for each performer? One messes up the dynamics, while the other allows the dynamics to come through unmolested, with only the relative level being artificial. If I were recording a live performance I would try to get away with as little dynamic compression as I could. Of course if most people are listening with headphones on the bus or walking down the street then they might complain if the dynamic range is too great. Optional artificial compression in the playback device would be best, I think.
Last edited:
Would it behave differently between active and passive? Would the mechanical limit begin to show up in the electrical characteristics of the driver which, combined with the passive crossover, would affect the sound differently?
if that really happens I guess you have exceeded driver limits and said goodbye to sound quality anyway
Active (voltage drive) should be slightly worse. Passive crossovers put impedance in series with the coil, reducing current variation with temperature a little.
Temperature effects will also change the reactive parts of the speaker impedance for various reasons and this will modify frequency and phase response with passive crossovers
limiters IMO are the last resort, a drum kit can be compressed just fine without brickwall limiters. Though i totally agree that 90db of dynamic range of a 120db drum hit is not necessarily desired in recorded material, let alone possible. Even the average spoken voice recorded at 55db @ 1m, reproduced at the same level, has 25db of dynamic range above the threshold of audibility (which i was taught as being 30db, may have since changed) thus the compression in recording helps reveal the low level detail which could be below the audibility threshold, at lower reproduction levels. I for one do not want to listen to a cd at domestic levels during a quiet intro, only to have my woofers explode when the track gets going. I think some classical fans mistake swells and shallows in movements as dynamics, i do not. I would regard the tight pizzicato of strings and their decay into reverberation as dynamics however.
I'm reasonably sure that it is possible to play a classical orchestral recording at home at the same level as you would hear it in the concert hall, and it sounds OK in the concert hall despite the lack of artificial compression and the audience shuffling and coughing. Would you absolutely need to compress it for other than aesthetic reasons?
And rock performances? Well, in a modest venue, the drums may not be amplified (i.e. you get the full dynamic range acoustically), while the vocals and guitars are set at the appropriate levels in the mix using the PA and the mixing desk (possibly limited or compressed, but maybe not). At the back of the room, the levels may not be much higher than you'd get in a domestic situation, but despite the chatter of the surrounding audience members you hear pretty much everything.
I'm making a distinction between compression (where the gain is constantly adjusted to compress dynamic range), and multi-track mixing where the vocals may be set at an artificially high level compared to the drums (just like singing through an amp against unamplified drums), but it's set at a more-or-less constant gain.
The use of some compression during the recording/mixing process is almost inevitable.
Especially vocals and electric bass are prone to substantial and unwanted changes in level, either by moving towards or away from the mic or plucking a string slightly harder than intended.
So some compression is needed but it does become a nuisance if to much of it is applied to make one cd sound louder than another.
In the early days of cd as a medium an rms value of -18dBFS was recommended and it sounded good. This was then revised down to -12dBFS rms and it was still ok but since the Loudness Wars kicked in this has gone down to -6dB or less and that is a horrible state of affairs. Almost everything mixed at those levels is only good for background muzak and even then it sounds bad.
Especially vocals and electric bass are prone to substantial and unwanted changes in level, either by moving towards or away from the mic or plucking a string slightly harder than intended.
So some compression is needed but it does become a nuisance if to much of it is applied to make one cd sound louder than another.
In the early days of cd as a medium an rms value of -18dBFS was recommended and it sounded good. This was then revised down to -12dBFS rms and it was still ok but since the Loudness Wars kicked in this has gone down to -6dB or less and that is a horrible state of affairs. Almost everything mixed at those levels is only good for background muzak and even then it sounds bad.
im sure an orchestral piece COULD be reproduced at real levels without compression, like for e.g. A monitor. Its just that 110db dynamic range is less RECORDABLE, so recording compression is required. I highly doubt you have any recordings with a 100db dynamic range. Even 80db is too high, taking threshold of hearing into account, compressing 2 to 1 makes it feasible, just.
Charles hit on an important point. Modern recordings with next to no dynamic range are designed to be muzak. Music which sounds the same no matter what you play it through. To most folks 'high fidelity' is just not important anymore.
Why not build a HQ expander? That would certainly increase dynamics, but likely decrease the lifespan of the rest of the chain...
Charles hit on an important point. Modern recordings with next to no dynamic range are designed to be muzak. Music which sounds the same no matter what you play it through. To most folks 'high fidelity' is just not important anymore.
Why not build a HQ expander? That would certainly increase dynamics, but likely decrease the lifespan of the rest of the chain...
Last edited:
I don't think there are any recordings with 100dB dynamic range in existence.
The original -18dBFS recommendation allows for about 36dB of dynamic range in a song not counting fade-outs.
Also 16bit do not allow enough resolution, at (very) low levels the response becomes audibly stepped. This can be heard in long reverb tails for example. 24 bit is much better in this respect.
The original -18dBFS recommendation allows for about 36dB of dynamic range in a song not counting fade-outs.
Also 16bit do not allow enough resolution, at (very) low levels the response becomes audibly stepped. This can be heard in long reverb tails for example. 24 bit is much better in this respect.
I think it is a bit more complex than that.
This might be a bit closer to what I mean:
Roger Nichols: Across The Board
Btw 8bit recordings sound really bad although it is used as an effect sometimes.
There are bit crunching plug ins available for this purpose.
This might be a bit closer to what I mean:
Roger Nichols: Across The Board
Btw 8bit recordings sound really bad although it is used as an effect sometimes.
There are bit crunching plug ins available for this purpose.
yes of course, i forget that when they say 16 bit, what they mean is 1 bit sigma delta DAC or something of the like.
Not true 16 bit, where there are 2^16 divisions.
As I said before, I have never heard that effect on anything but poor quality DTV, where piano (and everything else for that matter) seems stepped around a couple of Hz. Never heard the effect in 16bit PCM, regardless what bandwidth I used.
24bit DOES sound cleaner, but ive never heard anything I could precisely describe and attribute anything in particular.
Not true 16 bit, where there are 2^16 divisions.
As I said before, I have never heard that effect on anything but poor quality DTV, where piano (and everything else for that matter) seems stepped around a couple of Hz. Never heard the effect in 16bit PCM, regardless what bandwidth I used.
24bit DOES sound cleaner, but ive never heard anything I could precisely describe and attribute anything in particular.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.