In some ways its a step forward. The only need for analog inputs would be for something like phono. For digital audio, the data should only be converted back to analog one time, and that should be after all your DSP.
Don’t know if you will find this too expensive but I have been using Fouraudio DSP for a couple of years now and have found them to be really excellent with great support and a very good Win/Linux/Mac user interface.
Here’s an example of their standalone IIR/FIR
https://www.four-audio.com/en/products/hd2/
Here’s an example of their standalone IIR/FIR
https://www.four-audio.com/en/products/hd2/
here's the thing: I'm multi-amping my speakers and the DSP is the crossover, so all of my sources, preferably from my preamp, need to be routed through it.
CamillaDSP on a RaspberryPi 4 and a Motu Ultralite Mk5 USB Audio Interface. For around $US700 and an hour setting up the RPi you have a full featured web controlled DSP - FIR, PEQ, Gains, Mixers etc and the Motu UL5 will give you all in I/O you need. .
https://www.audiosciencereview.com/forum/index.php?threads/rpi4-camilladsp-tutorial.29656/
I can put a zipped SD card image up on Dropbox for the RPi that is working and would save you time.
I tri-amped my modified K-Horns and detailed the process here
https://www.diyaudio.com/community/...overs-room-correction-etc.349818/post-7213813
The CamillaDSP selects the source and is controlled by a remote
https://audiosciencereview.com/forum/index.php?threads/rpi4-camilladsp-tutorial.29656/post-1579025
Compared to miniDSP 2X4HD the CamillsDSP / Motu UL5 is silent and far more flexible.
Since these measurements were made I have experimented with "room curves" to give a bit more life to music.
this would be perfect if it had a GUI. I haven't done command line entry software since 1983
upon further reading, it does have a GUI, but it doesn't seem to be compiled? reading all of the command line entry required to get it going seems a bit onerous for the average user.
Very nice! And generous setup help offer with zip file
If i weren't into Q-sys, I think the RaspPi, with software of choice, is the way to go. Either that or a mini PC running Linux. (NO windows!)
I put JRiver on one a few years back, to do FIR convolution with it. Worked well. But nowhere near as easy to experiment with as Q-sys. And I experiment like crazy.
How do you do you EQ tuning work in Camilla? IIR, FIR, or combo?
Great results, however you're doing it
Last edited:
It has a GUI, it is pre-compiled, just need to download and unzip it. If you can copy / paste you can get this running. Once you have it up and running you can use the GUI to manage everything from any device with a web browser that is on the same network as the computer running CamillaDSP.
There are automated setup scripts -> https://github.com/HEnquist/camilladsp-setupscripts but depending on your specific hardware / OS they may not work correctly.
Yes. My main system consists of an Okto dac8 pro + RPi4 running CamillaDSP + Apple TV. I don't switch sources and only use the Apple TV as a source. Switching sources is relatively easy, typically done by changing CamillaDSP configurations which can be done via remote. Biggest issue with source switching is potential need to change the clock source / operating mode of the DAC.
The Apple TV remote is capable of turning the Okto (via IR), TV (via CEC) and Apple TV (via bluetooth) on/off and controls volume of the Okto. The Okto has a standby mode and will automatically turn on when it senses a signal, it also has a trigger out which turns on 4 amplifiers. After 10 minutes of inactivity, it goes back to standby, and the amplifiers shut off.
My office system uses a MOTU Ultralite Mk5 + RPi4 running CamillaDSP + Apple TV. Again, I don't switch sources here. I use a FLIRC IR USB device to control CamillaDSP volume and an OLED display for volume indication. This setup uses a Bobwire DAT1 and the optical output of the MOTU to provide trigger input to 3 amplifiers. Like the Okto setup it will turn on automatically if a signal is sensed and will turn off in about 1 minute of no signal.
Both remotes work great and will detect commands even if not directly aimed at the device via bouncing.
Michael
There are automated setup scripts -> https://github.com/HEnquist/camilladsp-setupscripts but depending on your specific hardware / OS they may not work correctly.
Yes. My main system consists of an Okto dac8 pro + RPi4 running CamillaDSP + Apple TV. I don't switch sources and only use the Apple TV as a source. Switching sources is relatively easy, typically done by changing CamillaDSP configurations which can be done via remote. Biggest issue with source switching is potential need to change the clock source / operating mode of the DAC.
The Apple TV remote is capable of turning the Okto (via IR), TV (via CEC) and Apple TV (via bluetooth) on/off and controls volume of the Okto. The Okto has a standby mode and will automatically turn on when it senses a signal, it also has a trigger out which turns on 4 amplifiers. After 10 minutes of inactivity, it goes back to standby, and the amplifiers shut off.
My office system uses a MOTU Ultralite Mk5 + RPi4 running CamillaDSP + Apple TV. Again, I don't switch sources here. I use a FLIRC IR USB device to control CamillaDSP volume and an OLED display for volume indication. This setup uses a Bobwire DAT1 and the optical output of the MOTU to provide trigger input to 3 amplifiers. Like the Okto setup it will turn on automatically if a signal is sensed and will turn off in about 1 minute of no signal.
Both remotes work great and will detect commands even if not directly aimed at the device via bouncing.
Michael
Sonically I don't think there is much difference.
Ultralite Mk5 has higher residual noise than the Okto, about 7 uV vs 3.5 uV which may be more noticeable with higher sensitivity drivers and/or higher gain amplifiers.
Based on testing with a Cosmos ADC which has low input impedance that varies by input sensitivity setting (lower impedance at lower voltage input settings), the Ultralite Mk5 does not like driving low impedances and the result is higher distortion. The Okto doesn't care, and the distortion results are low and do not change regardless of what input voltage setting is used on the Cosmos.
Okto is more consumer friendly as it has built-in IR volume control, a nice large screen with volume indication and a trigger output. Ultralite Mk5 has a small knob which can control volume of the analog output channels but no IR volume control, my work around for this is to use CamillaDSP volume control as mentioned previously.
Ultralite Mk5 has tons of I/O (8 channel analog input, ADAT/TOSLINK input, SPDIF input / 10 channel analog output, ADAT/TOSLINK output, SPDIF output, headphone output) and I use it for all sorts of testing purposes.
Both are designed to take an input, route it through a computer for processing and output the processed signal, which is an awesome capability.
In comparison to something like a Topping DM7 which is only an 8 channel USB DAC with no other I/O functionality, I think both have a significant advantage as they can be clock sync'd to a digital input. This is nice as you can use one device as both the capture and playback device in CamillaDSP without resampling or rate adjust which generally allows for lower chunk sizes resulting in lower latency. They also both have a digital output which allows you to clock other devices which can be useful when using a separate audio interface for acoustic measurements.
Both are great options. I think the Okto is a better device for most people unless you need the additional I/O of the MOTU.
Michael
Ultralite Mk5 has higher residual noise than the Okto, about 7 uV vs 3.5 uV which may be more noticeable with higher sensitivity drivers and/or higher gain amplifiers.
Based on testing with a Cosmos ADC which has low input impedance that varies by input sensitivity setting (lower impedance at lower voltage input settings), the Ultralite Mk5 does not like driving low impedances and the result is higher distortion. The Okto doesn't care, and the distortion results are low and do not change regardless of what input voltage setting is used on the Cosmos.
Okto is more consumer friendly as it has built-in IR volume control, a nice large screen with volume indication and a trigger output. Ultralite Mk5 has a small knob which can control volume of the analog output channels but no IR volume control, my work around for this is to use CamillaDSP volume control as mentioned previously.
Ultralite Mk5 has tons of I/O (8 channel analog input, ADAT/TOSLINK input, SPDIF input / 10 channel analog output, ADAT/TOSLINK output, SPDIF output, headphone output) and I use it for all sorts of testing purposes.
Both are designed to take an input, route it through a computer for processing and output the processed signal, which is an awesome capability.
In comparison to something like a Topping DM7 which is only an 8 channel USB DAC with no other I/O functionality, I think both have a significant advantage as they can be clock sync'd to a digital input. This is nice as you can use one device as both the capture and playback device in CamillaDSP without resampling or rate adjust which generally allows for lower chunk sizes resulting in lower latency. They also both have a digital output which allows you to clock other devices which can be useful when using a separate audio interface for acoustic measurements.
Both are great options. I think the Okto is a better device for most people unless you need the additional I/O of the MOTU.
Michael
Thanks for your extensive reply. I heard OktoDAC before and I liked it. DM7 less so. Still looking for best 8 channel solution, will probably go for an ES9039 PRO 8 channel AckoDAC once available. Would actually like to have 8 channel Tambaqui or MU2, but it does not exist and even if it was it would be very expensive. Also interested in these DSC2.6 DSD DACs, but this does not scale up yet to 8 channels.
So far I run on a pair of Hypex FA253, which already sounds very great as long as you setup the gain structure properly and ensure a good digital source.
More on topic, an option could also be the new MiniDSP 8 channel HT with PC processing (or internal DSP if required)...
So far I run on a pair of Hypex FA253, which already sounds very great as long as you setup the gain structure properly and ensure a good digital source.
More on topic, an option could also be the new MiniDSP 8 channel HT with PC processing (or internal DSP if required)...
If you're going to cast a shadow on some piece of gear (in this case, the Topping DM7 8-channel DAC), then I think it's not at all fair unless you're much more verbose on the equipment used, the conditions of the test (including room acoustics and loudspeakers used), and exactly subjectively what you heard that displeased you.
Also, I have to say that it doesn't inspire confidence when you say something like "...Still looking for best 8 channel solution...". That doesn't appear to something that, say, an engineer would say.
Why do I make comments here? My experience with the lower power version of the Hypex FusionAmp, FA122 (i.e., the better choice for fully horn-loaded loudspeakers with extremely high sensitivity), was the worst sounding solid state amplifier that I've heard to date on my K402/TAD TD-4002 HF drivers and KPT-KHJ-LF bass bins (i.e., fully horn loaded with nominal 105 dB/1m sensitivity), which is the subject of this thread: fully horn-loaded loudspeakers using DSP crossovers. It seems that the people using these FusionAmps are not using them on fully horn-loaded loudspeakers, where the first few milliwatts of output are what you're actually listening to 99% of the time.
When you say that you use Hypex FusionAmps with rated 250 w/channel output capability (FA253) without also discussing the other issues you accepted for using such high power amplifiers on fully horn-loaded loudspeakers, then I must suspect all the advice given.
JMTC.
Chris
Save the preamp for phono and or tape. Stay digital for a long as possible. Digital volume controls may or may not be ideal in how they implemented however. Probably better to have some switchable and or relay controlled analog attenuators after you dacs. Of course there will aways be some capacitance to drive after the attenuators, so good analog buffers may be best before going into power amps.
The point here is that going to multi-amped DSP systems looks very appealing in some ways, but its not so cheap or easy to do well. Multiple A/D/A conversions kill your audio quality whether you have noticed it yet or not. Its serious degradation I wouldn't want to accept for my own system. I know because I have done the A/B tests with very good data converters. And they are far, far better than any put in low-cost DSP boxes.
Last edited:
Not sure where the slight hostility comes from, but I will give it a 2nd round...
I elaborated more on Topping DM7 sound in a different thread. We listened to this DAC with various people, sound systems and rooms and concluded the sound is somewhat sterile and unnatural. I wanted to keep this part short as it appears there is triggered sometimes some unbelief & discussion regarding this point as the DAC is considered to have very good objective measurements. Fine if somebody can enjoy it, but it is just not for me.
Perhaps I should have said best fitting solution for me. I know "best" is not very descriptive & SMART. My overall targets would be both good objective and subjective qualities, suitability for digital volume control (ideally with some relais on output for range switching), natural & open sound, neutral with perhaps a touch of extra warmth, good stereo placement etc.
Well I did not refer to suitability for very high sensitivity speakers. I could indeed imagine there are better solutions for that kind of systems. Still, I think it would be possible to make it work decently. I said with the proper gain structure. Did you try to mod the FA122 to low its gain? I lowered the gain of my FA253 with about 4-6x (do not remember the exact number now). So it is also not 250W capable anymore. On top of this, I added a series resistor on mid/treble drivers to get better system SNR. I use an ESS AMT1 of ~98 dB/W, and a mid driver of ~93 dB/W which work ok this way.
Fedde
That's the key that unlocks the puzzle, I think. If you didn't take the time to read the OP's problem statement, somewhat carefully, I think you will wind up mixing non-horn-loaded DSP crossover (including multichannel DAC) solutions with fully horn-loaded solutions. I've found that they generally don't mix well.
Which thread? (Link please.) That was the point of the comment.
Not hostility at all but rather attempting to wake up some from sleep walking into the kind of soft thinking that leads some into issues and confusion--and eventual disappointment and disillusionment, since part of your proposed solution(s) I have good reason to believe really do not apply to the actual domain of application: horn-loaded efficiency (in case you didn't pick that up on that point already).
This is probably yet another example why some here have called for a separate "horn loaded" area (noting again the extreme sensitivity of one or more mods to this issue), to avoid such sleep-walking episodes.
YMMV.
Chris
The thread is called next level DSP crossover, not specifically mentioning high efficiency. I read the OP and already mentioned that my comments were somewhat off topic though.
Still, I think there are options to scale up solutions like Fusion amps towards high efficiency. The interesting question to me would whether the benefits of high efficiency speakers would remain with a higher power amp solution with gain reductions and post attenuation (series resistor or divider network)...
Anyhow, with modern DAC chips there is quite some more SNR headroom than previous generation DACs like the one used in Fusion.
Still, I think there are options to scale up solutions like Fusion amps towards high efficiency. The interesting question to me would whether the benefits of high efficiency speakers would remain with a higher power amp solution with gain reductions and post attenuation (series resistor or divider network)...
Anyhow, with modern DAC chips there is quite some more SNR headroom than previous generation DACs like the one used in Fusion.
Based on what I heard, I wouldn't touch FusionAmps with a 10-ft pole. You can ask me why I take such a stance if you don't understand why I would say this. I don't actually bite.
This is yet another reason why some here have called for a separate horn-loaded area, and it is probably the root reason why I spoke up in the way that I did. If you haven't figured out why horn loading cannot be replaced by direct radiators and more power in terms of hi-fi sound quality--you probably never will.
(I also have to say that I now doubt your assessments of sound quality based on the above quote. Come back for your free "subjective listening redemption coupon" when you figure it out.)
Chris