FIR HFD5.0 release notes: FIR implemented (4500 taps for one FIR filter at the input or 1500 taps per channel on the output)
FIR HDF 5.0 Help PDF: To create FIR filters, please use 93750Hz as the samplerate, as this is the rate at which the DSP processes the audio
Is this even a useful number of taps?
I have two FA503 Week 2 2023 with Firmware 5.x and ADAU1452 on 3 way Vox 253MTI floor standers. So I might try this on a preset for music. Leaving my existing IIR preset for TV.
If I remember correctly the best approach is to use FIR to linearise the phase and response of the drivers then combine this with IIR filters for crossovers and time align with delay.
Edit to add sampleratefor FIR usefulness calculation
FIR HDF 5.0 Help PDF: To create FIR filters, please use 93750Hz as the samplerate, as this is the rate at which the DSP processes the audio
Is this even a useful number of taps?
I have two FA503 Week 2 2023 with Firmware 5.x and ADAU1452 on 3 way Vox 253MTI floor standers. So I might try this on a preset for music. Leaving my existing IIR preset for TV.
If I remember correctly the best approach is to use FIR to linearise the phase and response of the drivers then combine this with IIR filters for crossovers and time align with delay.
Edit to add sampleratefor FIR usefulness calculation
Last edited:
I tried it this weekend. Works like it should! I created the FIR files in rePhase. I set up a 4-way system using FA253 for the mains and FA252 for each woofer system. XO at 250 / 48 dB/oct L-R, and 2500 and 4500 brick wall. All linearPhase.
There is the same number of taps on each output if you move FIR filters to output, and that number of taps is required. You can’t move the taps around like on MinDSP. The advantage of this is that all channels have the same delay, but of course 1500 taps at 93.75 kHz samplerate won’t get you high Q filters or brick wall filters at low frequencies. Other downsides are that the FIR files DON’T follow the config XML files, meaning uploading the same config to an other module requires you to manually add and upload all the FIR files as well. It also takes a very long time to upload the config when FIR is used. When it comes to the performance of the DSP and amps ; simply superb.
What do you think about connecting a transformer coupled ribbon tweeter directly to the NC100HF of the Fusionamp?
Some say don't do this because such a tweeter have almost dead short DC resistance (mine have 0,06 Ohm DCR).
I asked Hypex and they said this:
"I don’t expect any issues with this configuration but the amp is DC coupled so as a precaution you could wire a capacitor in series."
But as far as I know, Ncore amps have DC output protection, so I don't know what to think.
I want to avoid the series capacitor because in my opinion every cap has a sound signature independently on the transfer function.
Some say don't do this because such a tweeter have almost dead short DC resistance (mine have 0,06 Ohm DCR).
I asked Hypex and they said this:
"I don’t expect any issues with this configuration but the amp is DC coupled so as a precaution you could wire a capacitor in series."
But as far as I know, Ncore amps have DC output protection, so I don't know what to think.
I want to avoid the series capacitor because in my opinion every cap has a sound signature independently on the transfer function.
I don't believe in capacitor sound, I just hear their sound character. I would be happy if there was no audible difference between them, but unfortunately there is.
@sheeple
Yes, this is what I would like to know. I used this tweeter for a while without cap and no problem occured but I don't know is this safe or not in the long term.
Last edited:
I think the Hypex will do just fine without any cap, if you make sure there is absolutely no signal in the frequency range where the impedance is very low. Using a cap will protect the tweeter of course. I still use fuses on my Magnepan ribbon tweeters - it’s saved me from burning them a couple of times …..
If something happens with the amp, or if you accidentally send full range signal to the tweeter, a cap will protect it.
If something happens with the amp, or if you accidentally send full range signal to the tweeter, a cap will protect it.
I am sure there is a difference in sighted listening, but how about a proper double blind (I know, always tricky to do with speakers)?
Let's say an attentive listener compared two identical active speaker systems. Both systems use DSP filtering and crossovers for the woofer and tweeter. The only difference is that one system has the tweeter directly connected to its amp, and the other system has a capacitor between the tweeter and its amp. In my opinion, it is very plausible that the listener might perceive a difference, even if EQ was used to make the the frequency response as similar as possible between them.
The impedance load presented to the amplifier is different, and this might change the character somewhat. Even with EQ, the presence of the capacitor will alter the acoustical phase of the tweeter. So I am not surprised that YSDR can hear a difference.
The impedance load presented to the amplifier is different, and this might change the character somewhat. Even with EQ, the presence of the capacitor will alter the acoustical phase of the tweeter. So I am not surprised that YSDR can hear a difference.
Hey Julf!
Maybe do you remember when I did a double blind test comparing the analog and the digital input of the Fusionamp. In that blind test I failed to identify which input I heard, but in normal usage that test didn't changed anything because I still likes the sound of the digital input better.
Same with the capacitors. Although I didn't blind test that and I see no point doing so because blind test results doesn't have an effect on me if I normally listening to music etc. Even though I believe in the results of blind tests, although recently less and less.
Maybe do you remember when I did a double blind test comparing the analog and the digital input of the Fusionamp. In that blind test I failed to identify which input I heard, but in normal usage that test didn't changed anything because I still likes the sound of the digital input better.
Same with the capacitors. Although I didn't blind test that and I see no point doing so because blind test results doesn't have an effect on me if I normally listening to music etc. Even though I believe in the results of blind tests, although recently less and less.