That's what I just said?
With just one compression driver, it's probably unlikely yes, as I already mentioned earlier.
With a big line array, I just know from experience that if you don't do things the right way, you WILL hear it.
There is no reasoning here, just some standard math?
btw, 24 bit will give 146dB max theoretical.
48 bits wil give 291dB, NOT 288dB
SNR = 6.02N + 1.76
Thank you for the detailed discussion, @JukkaM as well, thanks!
I think I understand the issue now.
This project will be using MiniDSP-HD and some inexpensive 2-channel Class D amps. So it will have none of the integrated self protection (idiot proofing) features that I am used to with Hypex Fusion. So thanks for making me aware of this. The amps do have a volume control knob, and I intend to take advantage of that knob
.But at one meter? We both know you were being silly.
If I built a line array or similar construction with many identical drivers for home use (1-3 meter listening distance), I'd wire most of them, if not all, in series, because that will retain voltage sensitivity of the system and not cause an overload into the electrics of my home. That way, all else being equal, the hiss would be equal to that of one driver.
In PA, I would wire more, but not necessarily all, in parallel, but also the listening distances are much greater. Even then I needed to attenuate tweeters relative to other drivers.
It depends if the DSP is fixed point or floating point throughout. If there is a fixed point component in the calculations 0dB cannot be exceeded without digital clipping. This cannot be undone by any gain change afterwards. It is very obvious if this is happening as digital clipping sounds truly awful. I had a Najda that did this and it was important to attenuate at the input to leave a safe margin, otherwise certain tracks would cause clipping.
If the DSP is floating point throughout then there is effectively unlimited headroom to boost as long as the signal is brought back under 0dB before output. In that case there is no difference between the filters if a corresponding change in gain is made before the output of the DSP.
I am not being silly?
I am talking from real world experience.
Besides, that is also not the point.
Some manufacturers like to give each (pair) of driver it's own amplifier to be able to control beaming (even more).
Last edited:
I did however made one mistake (yay for not being able to ever edit...)
For noise the drivers don't add up with a 20log but a 10log.
Which results in 14dB.
Or in the end 55dB of noise and not 68.6dB.
With this project, I will design the DSP filter so that 0 dB is the maximum. It seems like the safest way to move forward. The end user will simply need to supply a bit more gain in the preamp stage or at the amp (that very useful volume knob)...
Another things that have been decided - - the mid/tweeter CTC spacing is now fixed at 165 mm. I have reviewed all the data I collected and simulations I made, and this is the best geometry.
I am waiting for delivery of a MiniDSP 2x4 HD. I am also ready to begin constructing cabinets, which I will probably get started on in a few weeks.
j.
Another things that have been decided - - the mid/tweeter CTC spacing is now fixed at 165 mm. I have reviewed all the data I collected and simulations I made, and this is the best geometry.
I am waiting for delivery of a MiniDSP 2x4 HD. I am also ready to begin constructing cabinets, which I will probably get started on in a few weeks.
j.
I wonder if it is possible to know if MiniDSP2x4HD is fixed point of floating point... ?
The spec sheet says that it utilizes a 400 Mhz Sharc floating point processor... so perhaps ?
Last edited:
It's one of those, bit more searching and we wil know
https://www.analog.com/en/parametricsearch/11131#/
(I don't know all the variants on top of my head)
edit:
Oh it's this one;
https://www.analog.com/en/products/adsp-21489.html
Unfortunately a fixed point limitation can be introduced in software or other processes like sample rate converters, so to be sure it is better to test the device.
This is not any kind of barrier it just needs to be considered as part of the gain structure. Most DSP hardware has pre and post level adjustment, the input can be attenuated to avoid any clipping that the processing might cause.
In some cases with fixed point DSP the input always has to be attenuated a little as even basic mixer functions can cause an inter stage level clip.
In a similar way that inter sample overs can cause a problem when reconstructed back to analogue.
No need to be concerned just keep it in mind when setting up the system and run a few tests before turning it up to 11
That's actually a fair point.
Guess I am so used to think that noise isn't coherent.
But in the case, it's the same noise going into different drivers.
Which would imply it's coherent.
Construction of the cabinets started a few days ago. I am using a cabinet grade plywood from my local supplier. This is birch faced popular core 3/4 inch material. I am pretty happy with it so far. Of course it does not have the density or screw holding ability of Baltic birch plywood, but in the interest of being easily replicable, I want to use a commonly available material.
Final cabinet dimensions
The normalized internal dimension ratios are 1 : 1.58 : 3.69, with the depth being the smallest dimension.
Final cabinet dimensions
The normalized internal dimension ratios are 1 : 1.58 : 3.69, with the depth being the smallest dimension.
The midrange sub-enclosure runs the full width of the cabinet. The interior dimensions are (inches) 10.25 W x 6 H x 3.75 D (260 x 152 x 95 mm). The normalized dimensional ratios are 1 : 1.60 : 2.73. The interior volume of this sub-enclosure is 3.8 liter, and it consumes 5.7 liter of total cabinet volume, including the thickness of its walls.
The interior of the sub-enclosure is lined with Noico 80 mil vibration damping material. This is a very soft butyl rubber with adhesive on one side, and aluminum foil on the other. I have used this on several previous projects.
Both the midrange and woofer will load the baffle with out-of-plane loads. Because the sub-enclosure spans the entire width of the cabinet, it serves as an effective brace. It will react the baffle load into the sidewalls as a shear load. Shear loading has a much lower potential for sonic impact than does out-of-plane loading. As I get further along in cabinet construction, I will explain the other bracing strategies.
I used construction adhesive to attach the sub-enclosure box to the inner baffle. I know from experience that the squeeze-out will obliterate any pencil lines on the inner baffle, making it difficult to accurately position the box. I used a pair of small nails as locator pins, and this allowed me to push the box up against the nails. I was able to get a consistent accurate positioning of the box even though I could not see my line. After the glue cured, I pulled the nails out.
j.
The interior of the sub-enclosure is lined with Noico 80 mil vibration damping material. This is a very soft butyl rubber with adhesive on one side, and aluminum foil on the other. I have used this on several previous projects.
Both the midrange and woofer will load the baffle with out-of-plane loads. Because the sub-enclosure spans the entire width of the cabinet, it serves as an effective brace. It will react the baffle load into the sidewalls as a shear load. Shear loading has a much lower potential for sonic impact than does out-of-plane loading. As I get further along in cabinet construction, I will explain the other bracing strategies.
I used construction adhesive to attach the sub-enclosure box to the inner baffle. I know from experience that the squeeze-out will obliterate any pencil lines on the inner baffle, making it difficult to accurately position the box. I used a pair of small nails as locator pins, and this allowed me to push the box up against the nails. I was able to get a consistent accurate positioning of the box even though I could not see my line. After the glue cured, I pulled the nails out.
j.
There‘s a simple carpenter’s trick for that. Drive some small nails almost fully into the edge you want to glue. Then cut off the nail heads so that a few millimeter of the remaining nail sticks out. When you position the edge on the panel, the nails will prevent drifting and lock the panel into place.
Not sure if anyone has recommended this yet, but I was contemplating a cheap 3 way monitor build with the Silver Flute W20RC38-08, Hivi DMB-A and SB 26STCN.
I recently tried the SF woofers as replacements for some older DM series B&W which were NLA and the overall size was close. I was shocked at how good they sounded being a $45 budget chinese cast frame driver. Build quality is excellent, even on par with most cheaper Scandanavian drivers.
Efficiency is almost 90dB for the 8 ohm version and 92dB for the 4 ohm. They can take a beating too without much power compression or TSP spec shift. The sound is smooth, but not veiled. Bass is excellent. Very controlled, accurate and clean, even close to xmax.
These haven't been that highly praised by Zaph according to his THD measurements. Being a very picky person myself, to my ears, this driver sounds way too good for the price. I'd put these up against any decent $100 8" driver crossed under 2k.
Drawbacks? Really only the mild peak at 4 to 5k, but with a proper LCR series notch, this wouldn't keep me from using them even in a decent 2 way. I turned my nose up at this driver before based on Zaphs review, but after using it myself, it's become my new favorite budget 8" along with the 6.5" version.
I recently tried the SF woofers as replacements for some older DM series B&W which were NLA and the overall size was close. I was shocked at how good they sounded being a $45 budget chinese cast frame driver. Build quality is excellent, even on par with most cheaper Scandanavian drivers.
Efficiency is almost 90dB for the 8 ohm version and 92dB for the 4 ohm. They can take a beating too without much power compression or TSP spec shift. The sound is smooth, but not veiled. Bass is excellent. Very controlled, accurate and clean, even close to xmax.
These haven't been that highly praised by Zaph according to his THD measurements. Being a very picky person myself, to my ears, this driver sounds way too good for the price. I'd put these up against any decent $100 8" driver crossed under 2k.
Drawbacks? Really only the mild peak at 4 to 5k, but with a proper LCR series notch, this wouldn't keep me from using them even in a decent 2 way. I turned my nose up at this driver before based on Zaphs review, but after using it myself, it's become my new favorite budget 8" along with the 6.5" version.
I tried that a few times, and I thought it was more effort than my way. It was very easy to tap a 3d finishing nail a half inch into the plywood, and then later pull it out with a pliers.
I don't own a nail gun and I don't have an air compressor in my shop.
When did Zaph measure a 8" Silver Flute? That model has been off my radar...
US$45 for a Chinese made driver is a premium product.... please share measurements!