DSP, Horns, HDMI and Home Theatre
There is plenty of information about active, multi-way systems and DSP.
It looks the way to do it, the time delay is particularly of interest for horns, to allow for the physical offset.
A typical DSP system like miniDSP, lets you avoid multiple D/A then A/D - just use the SPDIF inputs. Fine for CD players, if anyone remembers them
But it becomes complicated when one tries to expand to do quality Home Theatre.
Lossless audio is now available on Blu-Ray but the multi-channel output is ONLY available on HDMI.
To avoid a new receiver I can decode that on the player to put 8 channel PCM audio on the HDMI.
But MiniDSP don't have an HDMI to SDPIF converter.
Not impressive that their AVR products don't work with their core products.
Any one have any solutions?
David
There is plenty of information about active, multi-way systems and DSP.
It looks the way to do it, the time delay is particularly of interest for horns, to allow for the physical offset.
A typical DSP system like miniDSP, lets you avoid multiple D/A then A/D - just use the SPDIF inputs. Fine for CD players, if anyone remembers them
But it becomes complicated when one tries to expand to do quality Home Theatre.
Lossless audio is now available on Blu-Ray but the multi-channel output is ONLY available on HDMI.
To avoid a new receiver I can decode that on the player to put 8 channel PCM audio on the HDMI.
But MiniDSP don't have an HDMI to SDPIF converter.
Not impressive that their AVR products don't work with their core products.
Any one have any solutions?
David
Last edited:
I had the same concern - multiple a/d - d/a steps using miniDSP.
I think some AVRs at least will give you a digital PCM output of the main L and right C channels but not the surround.
But I asked myself is this a problem I need to worry about. I knew from my college days decades ago that there is a simple formula for SNR based on quantization noise, so I looked it up
In an ideal analog-to-digital converter, where the quantization error is uniformly distributed between −1/2 LSB and +1/2 LSB, and the signal has a uniform distribution covering all quantization levels, the Signal-to-quantization-noise ratio (SQNR) can be calculated from
SQNR = 6.02 Q in DB
Where Q is the number of quantization bits.
Plug in 24 bits you get 144 db. Each time you A/D D/A you degrade this by 6 db. There is plenty of headroom for doing this a 2nd time in the miniDSP. Its an unusual amplifier that has a noise floor even 110 db below the 1W output so the amplifier's noise will dominate and determine the overall SNR.
I think some AVRs at least will give you a digital PCM output of the main L and right C channels but not the surround.
But I asked myself is this a problem I need to worry about. I knew from my college days decades ago that there is a simple formula for SNR based on quantization noise, so I looked it up
In an ideal analog-to-digital converter, where the quantization error is uniformly distributed between −1/2 LSB and +1/2 LSB, and the signal has a uniform distribution covering all quantization levels, the Signal-to-quantization-noise ratio (SQNR) can be calculated from
SQNR = 6.02 Q in DB
Where Q is the number of quantization bits.
Plug in 24 bits you get 144 db. Each time you A/D D/A you degrade this by 6 db. There is plenty of headroom for doing this a 2nd time in the miniDSP. Its an unusual amplifier that has a noise floor even 110 db below the 1W output so the amplifier's noise will dominate and determine the overall SNR.
Unfortunately, what I failed to consider in my first post is that the SNR coming out of the AVR's preamp outputs is probably well below that theoretical 144 db and so re-sampling and converting might well be a significant degradation. This is where a pre-pro with differential outputs might be worthwhile - iff that degradation is common mode noise.
Exactly.
And with really efficient compression drivers (that aren't padded down with a passive crossover), it starts to bother me.
And it's just clumsy to convert back and forth.
I have had a bit of a search and it seems the problem has been discussed for years but I can't find one reasonable solution, unbelievable.
Best wishes
David
the problem may be licensing - its big bucks
I don't know if you can get anywhere at all with free/open software/protocols and HDMI - almost certainly not from a consumer universal player
antipiracy concerns usually prevent manufacturers form putting digital content out in the clear
I don't know if you can get anywhere at all with free/open software/protocols and HDMI - almost certainly not from a consumer universal player
antipiracy concerns usually prevent manufacturers form putting digital content out in the clear
Yes, but it's even worse - it seems the latest HDMI license explicitly restricts not only multi-way SPDIF but even any alternative multi-way output from disc players.
Not just antipiracy but anticompetitive.
I am just surprised there is no readily available hack or work-around.
Should be possible to access the I2S stream of the DACs in an HDMI product but presumably that is not so easy or it would be more common.
There is a temptation to do this, just on principle.
Anyone here know much about I2S?
Best wishes
David
Have you seen this?
nanoAVR HDA | MiniDSP
At the moment I use PS3 and HTPC with LPCM out into the nanoAVR HD (the first one, with no analogue outs) into a Denon AVR (LPCM out into the Denon) and drive my B&C DE250 from the amp in my AVR. No worries. For vanity's sake I'm going to upgrade the power amp. Because I'm running 3 way active and want surround sound, I'm going to do pre-out into ADC (PCM9211) and miniSHARC + miniDAC8. It may be possible for me to use my HTPC with both USB out and HDMI out. In that config I could feed a USB to i2s board into the PCM9211 i2s port. I'll leave that option open but not experiment with it until a bit later.
With that I can run 5ch for centre/surrounds and .1 for subwoofer off the nanoAVR HD + Denon AVR3808 and up to 4 way active front channels off the miniSHARC+miniDAC8.
A blog diagram of what I'm building is in my blog post here - Active surround sound HiFi
Starting from scratch it would be simpler to have a nanoAVRHD product that can do more channels but that would require HDMI 2.0 which has very few chipsets available on market at the moment from my understanding.
Have you worked out how many channels you need for active surround sound hifi ... 8ch doesn't cut it. I will have 14 but I'm worried that's not enough, for later on. My wife says our living room is getting crowded by my sh1t.
Chris
nanoAVR HDA | MiniDSP
At the moment I use PS3 and HTPC with LPCM out into the nanoAVR HD (the first one, with no analogue outs) into a Denon AVR (LPCM out into the Denon) and drive my B&C DE250 from the amp in my AVR. No worries. For vanity's sake I'm going to upgrade the power amp. Because I'm running 3 way active and want surround sound, I'm going to do pre-out into ADC (PCM9211) and miniSHARC + miniDAC8. It may be possible for me to use my HTPC with both USB out and HDMI out. In that config I could feed a USB to i2s board into the PCM9211 i2s port. I'll leave that option open but not experiment with it until a bit later.
With that I can run 5ch for centre/surrounds and .1 for subwoofer off the nanoAVR HD + Denon AVR3808 and up to 4 way active front channels off the miniSHARC+miniDAC8.
A blog diagram of what I'm building is in my blog post here - Active surround sound HiFi
Starting from scratch it would be simpler to have a nanoAVRHD product that can do more channels but that would require HDMI 2.0 which has very few chipsets available on market at the moment from my understanding.
Have you worked out how many channels you need for active surround sound hifi ... 8ch doesn't cut it. I will have 14 but I'm worried that's not enough, for later on. My wife says our living room is getting crowded by my sh1t.
Chris
Last edited:
Yes, I have looked at this but don't quite see how to fit it into my plans.
I plan a modular system where each speaker has a dedicated power-pack with amps matched to drivers and would prefer each to have its own crossover.
Possibly passive crossovers in the surrounds/rears for simplicity.
That is affordable with a miniDSP 4 way in the R, L and C speakers but then I can't connect the SPDIF to HDMI so I end up with D to A then A back to D to A.
Clumsy.
However I will look more carefully at your system and maybe have a rethink.
Fully HDMI is one possibility I suppose.
Does the PS3 handle the lossless BluRay audio formats?
Thanks for the ideas.
Best wishes
David
Had a look at your pics.
It looks like it converts HDMI to A in the receiver then A to D for the miniSharc then D to A at the miniDAC?
Last edited:
That's a lot of boxes. I'd try to look for a USB or thunder bolt HDMI input/output card and then run the 8 LPCM channel into OSX and, I think, you may have up to 16 output channels available. Then I'd run two AVR receivers on two HDMI cables for 16 channels of digitally derived info.
Three HDMI cables, one mac and two AVRs for 16 amplified speaker ports. Requires some research tho!
My impression is that 'sending' HDMI requires no licence but all 'receiving' ICs need licensing.
Three HDMI cables, one mac and two AVRs for 16 amplified speaker ports. Requires some research tho!
My impression is that 'sending' HDMI requires no licence but all 'receiving' ICs need licensing.
Last edited:
Part of the idea for dedicated power packs for each speaker is to minimize this problem.
Power packs built in to each speaker result in a tidy installation and each amp can be optimized for its own driver.
This is pretty standard for professional audio.
It's the one part of the system that I am committed to.
Possibly I may drive the surrounds/rears from my receiver.
Best wishes
David
Yes, because I want active speakers + surround sound.
HDMI only supports 8ch which will never be sufficient. And a single DSP module doing more than 8ch output will reduce the power of the DSP available for each channel.
I have the option to (after building the first prototype in the next month or so) do dual audio output from my HTPC with 8ch over HDMI and 2ch over USB to feed the miniSHARC + miniDAC8. This would give me a single DtoA conversion for each signal, but I am unsure of the effect of clock drift between the USB data and the HDMI data. So I'm not committing myself to that solution for dual outputs from the PC.
I don't think I can get dual digital outputs from PS3 or other similar consoles though HDMI audio + optical audio would be nice I'm not sure the config allows that (plus 192k over optical is tricky, and for a consumer device it is unlikely it would be anything but a compressed multichannel stream over optical).
I think HDMI licensing prevents a clean HDMI to spdif conversion for 8ch (4 x stereo spdif signals) of uncompressed spdif output.
Last edited:
Me too😉
I don't think this is a limitation, it should be possible to have as many outputs as needed if you use the analog capability creatively.
Just need to consider each output not as an analog version of a HDMI channel but as an independent "way".
For instance, a separate NanoAVR dedicated to each of the HDMI channels would allow 8-way active speakers for 8 channels.
To be more realistic, I have considered 1 nanoAVR to drive the 3-way DSP L + 3-way R and 2-way C.
Second nanoAVR for L and R surrounds and rears, 2-way each.
Make sense?
Best wishes
David
Last edited:
There are some posts about a setup like you describe on minidsp forum on their website, I think it was called 'full gonzo' or similar. Would be worth looking over there I think if you're considering options like that.
To me, the challenge of that type of setup is input switching and volume control and locking yourself to hdmi sources only.
Chris
Yes, number of channels isn't a problem but there are some complications.
The nanoAVR has master volume control so I don't think volume control is a problem.
Probably even an improvement, because it can do all the calculations on full scale data and then attenuate.
That should lead to better accuracy than attenuation prior to an A to D and D to A reconversion so I assume that's what they do.
Non HDMI sources are the catch, I'm new to this bit, what's the easiest way to AD convert to HDMI?
For me that would be only radio anyway, not critical.
Best wishes
David
Thanks for the title of the miniDSP thread, I see there that volume control is indeed a problem, hmm.
Last edited:
Dave Zan,
For some horn system ideas, check out John Inlow's site. (USA). If you use good quality, wide band compression drivers like Tad, Jbl, BMS etc .and add a nice Fostex horn tweeter you can easily use a 6 db passive crossover down to 400-500 hz with excellent results. Just physically line up the voice coils. DSP has its place, but try to avoid making things more complicated than they have to be. Just my .02 cents.
For some horn system ideas, check out John Inlow's site. (USA). If you use good quality, wide band compression drivers like Tad, Jbl, BMS etc .and add a nice Fostex horn tweeter you can easily use a 6 db passive crossover down to 400-500 hz with excellent results. Just physically line up the voice coils. DSP has its place, but try to avoid making things more complicated than they have to be. Just my .02 cents.
I have JBL 2453s compression drivers so no problem there, but I think 6dB/octave is too slow.
However, the main issue is physical layout, the horns are reasonably deep and so the VCs do not line up.
Major investment of time and effort into the cabinets so this is not easy to alter.
I planned an active DSP crossover anyway so it seemed like no extra complication to compensate for the offset.
Now I realize that my plan has a few problems that I didn't anticipate.
It has made me reconsider the GedLee Summa. Is the horn profile shallower and does this make it easier to line up the drivers?
Has this been discussed?
Best wishes
David
I have seen John Inlow's site previously, thanks for the reminder.
Not exactly my style but a fun read.
Last edited:
It's now the "Geddes NS-15" and yes the shallower waveguide does make for an easier time alignment. With the low pass filter delay they are almost perfectly aligned as they stand. I now use a DE500 which has the diaphragm more forward than the older DE250, which helps as well.
This is the usual audiophile dilemma--looks vs size/performance!!!
A proper horn system is usually big and "ugly" to sound up to its full potential. NO box system can match the sound quality. The Summa is a very good sounding speaker system for its size but can't compete with a "real" 3 or 4 way proper horn system sound wise IMHO.
If you can't / don't want to have a system that looks similar to John Inlow's horns, then go for a Summa type system as a compromised (size wise) horn system will not sound as good.
Choices, choices--there is no free lunch.
A proper horn system is usually big and "ugly" to sound up to its full potential. NO box system can match the sound quality. The Summa is a very good sounding speaker system for its size but can't compete with a "real" 3 or 4 way proper horn system sound wise IMHO.
If you can't / don't want to have a system that looks similar to John Inlow's horns, then go for a Summa type system as a compromised (size wise) horn system will not sound as good.
Choices, choices--there is no free lunch.
I does seem like CD manu's should provide the exit angle now that 3D printing, CNC are hobbyist tools and we can actually do something with the information
Do you have any data to support this claim? Or is it just a guess? My guess is that the opposite is true because the fewer the crossover points the better.
- Status
- Not open for further replies.