Why Don't High SINAD Receivers Exist?

[Vincentponcet]

Here's what I'd like to see:

we are all dreaming of that.

It looks like if you want a true digital chain with no downsampling today, using the PC as source is the only way.

 

[rhollan]

Well, on the specs tab on the web page for the device it says "Processing resolution and sample rate 32 bit / Matched sample rate". Page 9 of the manual notes "Note that the internal sample rate is 48 kHz regardless. An asynchronous sample rate converter on the output channels is used to produce data at the sample rate on input detected on input channels 1 and 2."

 

[rhollan]

we are all dreaming of that.

It looks like if you want a true digital chain with no downsampling today, using the PC as source is the only way.

Yes, but then no Atmos and likely limited to Windows if wanting PC-based Dirac Live.

 

[Vincentponcet]

Well, on the specs tab on the web page for the device it says "Processing resolution and sample rate 32 bit / Matched sample rate". Page 9 of the manual notes "Note that the internal sample rate is 48 kHz regardless. An asynchronous sample rate converter on the output channels is used to produce data at the sample rate on input detected on input channels 1 and 2."

If you should read the 50 pages manual to understand that the product specs saying "Matched sample rate clocked to first input channel (e.g. 192kHz in, 192kHz out)" means there is an internal resampling to 48KHz, that's a bit of a scam IMO.
They could add in the product page a simple note like :"( Note that the internal sample rate is 48 kHz regardless)".

So it means that even standard CD 16/44.1 is resampled to 48KHz, that's not good.

 

[lashto]

Here's what I'd like to see:

A basic pre/pro with HDMI switching and decoding. It might have room correction, or not. It might have DACs or not. And it might have some channels of amplification (be an AVR) or not. But what it DOES have is this: an HDMI "loop" carrying up to 16 channels of LPCM audio for outboard processing, with either an HDMI input of same (to use it's DACs) or a pair of DB-25 (Yamaha pinout or similar) balanced analog inputs, to run directly to outputs (or VCAs for volume control). The most basic box just does switching, and you can license decoding software for it (TrueHD, DTS, etc. Maybe even room correction). Should cost $1000 or less for the basic model. Essentially, a "baby Trinnov". With this you can do outboard room correction, DACs, amps, to your heart's content, so long as digital audio paths are protected by HDCP on HDMI interfaces.
.

Would be nice but sounds like your "basic pre/pro" will do exactly what HDCP (and the whole industry) 'protects' against: decoding HDMI to lossless dígital out.

Oh nooooo.
I didn't see that in MiniDSP product page, that's they do upsampling. That would be like a scam if it is not said in their product page.
Dirac claims to support 192KHz but then I don't know which devices.
https://www.dirac.com/news/2016/2/1...ection-suite-with-192-khz-sample-rate-support

Pretty much all AVR/Ps with room correction use 48kHz DSP. Surprisingly, noone advertises that downsampling 'feature' . The only exceptions are some seriously pricey trinnov and datasat boxes. And if you search this thread, I linked a few AVP-like boxes that do 192kHz DSP (car sound processors).
There are a few legitimate reasons for processing at 48kHz (i.e. a max audio freq of ~24kHz):

• 48kHz is ~ the standard samplerate for the movie industry. Only DTS-HD and newer formats can go above 48kHz and not sure how many movies actually do that. Even google seems clueless, I was only able to find one: the Akira bluray. So (arguably), processing above 48kHz is not necessary for movies.
• AVR measurements mics are not exactly state of the art and are not reliable >20kHz. IIRC, the cheapo Denon/Marantz mics can barely do 18kHz.
• audio freqs over 20kHz are not exactly audible and many speakers/tweeters don't even play them. Processing any samples over 20kHz is (again arguably) a waste of DSP power.

In spite of all that, there is no need to downsample to 48kHz, the DSP algorithm can 'simply' ignore all audio samples over 20kHz. Also, DSP power to fully process 192kHz is not that expensive, those 192kHz car-processors don't cost an arm and a leg like trinnov/datasat.

Long story short: "why-dont-high-sinad-receivers-exist" is IMO just a matter of cost cutting and/or incompetence and the general consumer-unfriendly state of the movie/media industry.

 

[rhollan]

If you should read the 50 pages manual to understand that the product specs saying "Matched sample rate clocked to first input channel (e.g. 192kHz in, 192kHz out)" means there is an internal resampling to 48KHz, that's a bit of a scam IMO.
They could add in the product page a simple note like :"( Note that the internal sample rate is 48 kHz regardless)".

So it means that even standard CD 16/44.1 is resampled to 48KHz, that's not good.

I didn't find it misleading, just worded badly. It is questionable if it is audible: ASRC done right isn't. But it raises the movies/processed vs. stereo/unprocessed dichotomy. Why NOT apply room correction and digital bass filtering to stereo? One is basically trading one form of dustortion for another and it is not clear that this would not be an improvement. (Of course room treatment first is a better idea.)

 

[rhollan]

Would be nice but sounds like your "basic pre/pro" will do exactly what HDCP (and the whole industry) 'protects' against: decoding HDMI to lossless dígital out.

I meant decoding formats like Atmos from bitstream to LPCM, with digital outputs still over HDMI, protected with HDCP.

 

[lashto]

I meant decoding formats like Atmos from bitstream to LPCM, with digital outputs still over HDMI, protected with HDCP.

Not 100% sure but I would bet that any form of hdmi-bitstream-to-lpcm-lossless conversion is forbidden because you will be able to 'freely' extract/access that LPCM (e.g. with a PC). And that is what you have to do in order to use your own DSP/DAC.

I can only see two possibilities:

• someone hacks the new formats (Atmos & co) and provides 'open' codecs, just like we have for all older formats.
• the AVR/P industry gets its act together and builds proper highres units for reasonable prices. Would not bet a single cent on this one

 

[rhollan]

Not 100% sure but I would bet that any form of hdmi-bitstream-to-lpcm-lossless conversion is forbidden. Otherwise you will be able to 'freely' extract/access that LPCM (e.g. with a PC). And that is what you have to do in order to use your own DSP/DAC.

I can only see two possibilities:

• someone hacks the new formats (Atmos & co) and provides 'open' codecs, just like we have for all older formats.
• the AVR/P industry gets its act together and builds proper highres units for reasonable prices. Would not bet a single cent on this one

Even protected by HDCP? HDMI 2.0 allows for 32 channels of LPCM audio. The idea is to allow for arbitrarily good LPCM DACs to be used with any bitstream decoder and not tying them together. I would think HDCP would be sufficient to protect the digital audio stream.

 

[Vincentponcet]

Not 100% sure but I would bet that any form of hdmi-bitstream-to-lpcm-lossless conversion is forbidden because you will be able to 'freely' extract/access that LPCM (e.g. with a PC). And that is what you have to do in order to use your own DSP/DAC.

I can only see two possibilities:

• someone hacks the new formats (Atmos & co) and provides 'open' codecs, just like we have for all older formats.
• the AVR/P industry gets its act together and builds proper highres units for reasonable prices. Would not bet a single cent on this one

Object based codecs like Atmos or DTS:X are not just about decoding, it is calculating the sound to send to a channel depending on loudspeakers position and their numbers.

 

[lashto]

Even protected by HDCP? HDMI 2.0 allows for 32 channels of LPCM audio. The idea is to allow for arbitrarily good LPCM DACs to be used with any bitstream decoder and not tying them together. I would think HDCP would be sufficient to.protect the digital audio stream.

Nothing is 100% "sufficient". Even with a HDCP chain, you can (in theory) tap the digital inputs of your DAC and get the LPCM.
Not 100% but quite sure that "even with HDCP" is forbidden. HDFury has (or used to have) such boxes and they were/are constantly getting sued by "the industry". IIRC, miniDSP had to deal with similar headaches.

And if it was possible, what exactly will you gain? You'll still need a DSP/DAC box with HDCP-compliant inputs and guess who will build that one: the same "cartel" that builds AVR/Ps. Same people, same tech, same as bad, same as overpriced...

Object based codecs like Atmos or DTS:X are not just about decoding, it is calculating the sound to send to a channel depending on loudspeakers position and their numbers.

true, but not sure how is that relevant. 'Anyone' can do that .. as long as they pay the media-cartel tax

 

[rhollan]

Object based codecs like Atmos or DTS:X are not just about decoding, it is calculating the sound to send to a channel depending on loudspeakers position and their numbers.

Sure, and the decoder/processor can still do that: the decoded output will be tied to the room. Even more reason to not try and copy and store it.

 

[rhollan]

Nothing is 100% "sufficient". Even with a HDCP chain, you can (in theory) tap the digital inputs of your DAC and get the LPCM.
Not 100% but quite sure that "even with HDCP" is forbidden. HDFury has (or used to have) such boxes and they were/are constantly getting sued by "the industry". IIRC, miniDSP had to deal with similar headaches.

And if it was possible, what exactly will you gain? You'll still need a DSP/DAC box with HDCP-compliant inputs and guess who will build that one: the same "cartel" that builds AVR/Ps. Same people, same tech, same as bad, same as overpriced...

true, but not sure how is that relevant. 'Anyone' can do that .. as long as they pay the media-cartel tax

Well, HDCP is deemed "sufficient" to protect video and audio today. I2S audio could be picked off from inside modern AVRs too.

I just think there is an argument for separating D/A from decoding/processing, and HDMI with HDCP should be adequate protection for a "raw" digital audio signal.

HDFury got in trouble for HDMI to analog component video decoding (and they won!). MiniDSP requires HDCP on their nanoAVRs so I can't see how they ran afoul of anything. They probably discontinued tge line because they were not selling enough to cover the HDCP licencing costs.

 

[mdsimon2]

Would be nice but sounds like your "basic pre/pro" will do exactly what HDCP (and the whole industry) 'protects' against: decoding HDMI to lossless dígital out.

Pretty much all AVR/Ps with room correction use 48kHz DSP. Surprisingly, noone advertises that downsampling 'feature' . The only exceptions are some seriously pricey trinnov and datasat boxes. And if you search this thread, I linked a few AVP-like boxes that do 192kHz DSP (car sound processors).
There are a few legitimate reasons for processing at 48kHz (i.e. a max audio freq of ~24kHz):

• 48kHz is ~ the standard samplerate for the movie industry. Only DTS-HD and newer formats can go above 48kHz and not sure how many movies actually do that. Even google seems clueless, I was only able to find one: the Akira bluray. So (arguably), processing above 48kHz is not necessary for movies.
• AVR measurements mics are not exactly state of the art and are not reliable >20kHz. IIRC, the cheapo Denon/Marantz mics can barely do 18kHz.
• audio freqs over 20kHz are not exactly audible and many speakers/tweeters don't even play them. Processing any samples over 20kHz is (again arguably) a waste of DSP power.

In spite of all that, there is no need to downsample to 48kHz, the DSP algorithm can 'simply' ignore all audio samples over 20kHz. Also, DSP power to fully process 192kHz is not that expensive, those 192kHz car-processors don't cost an arm and a leg like trinnov/datasat.

Long story short: "why-dont-high-sinad-receivers-exist" is IMO just a matter of cost cutting and/or incompetence and the general consumer-unfriendly state of the movie/media industry.

I think you may be over simplifying here. Doing FIR filtering at higher sample rates drastically increases the number of taps required, this is especially an issue with lower frequencies. Admittedly I am not familiar with all room correction technology but I imagine that most of them are using some amount of FIR filtering like Dirac.

Michael

 

[rhollan]

I think you may be over simplifying here. Doing FIR filtering at higher sample rates drastically increases the number of taps required, this is especially an issue with lower frequencies. Admittedly I am not familiar with all room correction technology but I imagine that most of them are using some amount of FIR filtering like Dirac.

Michael

Yes, but only as whatever the increase in sampling rate is: 192 kHz requires four times as many taps at 48 kHz for the same frequency resolution. It's an O(n) increase.

And, Dirac Live uses both (short) FIR and IIR filters (the FIR mostly to correct speaker impulse response, IIRC).

 

[mdsimon2]

Yes, but only as whatever the increase in sampling rate is: 192 kHz requires four times as many taps at 48 kHz for the same frequency resolution. It's an O(n) increase.

And, Dirac Live uses both (short) FIR and IIR filters (the FIR mostly to correct speaker impulse response, IIRC).

Agree with everything you say here but I do think it is a bit worse than you describe. I do not claim to be an expert here but I have played around quite a bit with FIR filters using a miniSHARC. If you double the sampling rate you effectively need 4 times the processing power because not only do you require double the amount of taps but at the higher rate the processor will also halve the number of taps.

See link below for a comment by pos (author of rephase) that also supports this.

https://www.minidsp.com/forum/opend...e-minisharc-not-yet-released?limit=6&start=84

Michael

 

[rhollan]

Agree with everything you say here but I do think it is a bit worse than you describe. I do not claim to be an expert here but I have played around quite a bit with FIR filters using a miniSHARC. If you double the sampling rate you effectively need 4 times the processing power because not only do you require double the amount of taps but at the higher rate the processor will also halve the number of taps.

See link below for a comment by pos (author of rephase) that also supports this.

https://www.minidsp.com/forum/opend...e-minisharc-not-yet-released?limit=6&start=84

Michael

That's a limit on the total number of taps imposed by the hardware.

I consider a factor of four on a parallelizable process reasonable: just add more DSP chips. Others may find the cost prohibitive.

 

[Kustomize]

I’ve a Denon AVR X2400H which connects to my TV with HDMI Arc. Will I benefit from using the spdif optical connection to my tv instead? Will i benefit from adding some of the under 100 USD Dac tested here?

 

[lashto]

I think you may be over simplifying here. Doing FIR filtering at higher sample rates drastically increases the number of taps required, this is especially an issue with lower frequencies. Admittedly I am not familiar with all room correction technology but I imagine that most of them are using some amount of FIR filtering like Dirac.

Michael

I surely am, that's why I wrote 'simply' in quotes
Not a DSP expert but I can see car-processors doing 8-16 channels DSP at 192kHz for about $4000.
The $4k AVR/Ps can barely do 48kHz DSP, use much cheaper DACs and measure much worse. No need for a DSP doctorate to see the problem here.

 

[markus]

Regarding "taps", long FIR filters can't be used in an AVR/P because it would introduce too much lag. Low frequency resolution can be increased even with a lower tap count using some tricks. Look at what Audyssey did with XT32.

I surely am, that's why I wrote 'simply' in quotes
Not a DSP expert but I can see car-processors doing 8-16 channels DSP at 192kHz for about $4000.
The $4k AVR/Ps can barely do 48kHz DSP, use much cheaper DACs and measure much worse. No need for a DSP doctorate to see the problem here.

An AVR/P does more than a "car-processor" and more development effort is required. There's also additional licensing fees that need to be covered (Dolby, DTS, Auro, HDMI). There's a reason why Trinnov charges a bit more than 4k