Head to Head JRMC19, Foobar+SACD (and HQ Player) doing Redbook to DSD and native DSD

[EuroDriver]

Thanks Edward, that makes perfect sense.

 

Oh, I might eventually get a Hypex DLCP processor or try HQP's crossover function, but for the former it's a *lot* of work to cut away the internal xovers and also to emulate my ATCs and Velodynes, then patch those speakers. My CPU is too strained for the latter. I'm looking at a much more muscular computer, but that's for later. Also, ATC's internal crossover is quite good and beautifully constructed, I inspected it when replacing a woofer. You know, it's *long* past time some of these top pro/consumer builders like ATC and B&W offered zero-crossover versions...

 

2 months ago, I was of the opinion that digital crossover was definitely the way to go a la Meridian. Today, after falling in love with DSD, I am not so sure. What the DSD experience has shown me is that the sampling rate is very important and that the bandwidth limiting requirement for A>D and D>A can be significantly damaging to the the sound. I think a major reason why DSD sounds good is that the sampling rates are much higher than the sample rates in PCM. My feeling is for A>D 64 fs 2.8 MHz is enough, but for playback you really want 256 / 11 MHz

 

T+A are of the opinion that reproduction of frequencies up to something like 70 KHz is beneficial which may be going too far. However it is an indication that perhaps one needs to be careful that the playback system is transparent at very much higher frequencies than human hearing of sine waves.

 

I did not realize that HQP had a crossover function, quite intriguing. Could I then go to a 4 channel set up ? I need to read the manual ! An E28 feeding 2 tweeters, 2 woofers and a 3 sub woofer swarm with its own 48 KHz DSP to handle the room modes

[Sam Lord]

2 months ago, I was of the opinion that digital crossover was definitely the way to go a la Meridian. Today, after falling in love with DSD, I am not so sure. What the DSD experience has shown me is that the sampling rate is very important and that the bandwidth limiting requirement for A>D and D>A can be significantly damaging to the the sound. I think a major reason why DSD sounds good is that the sampling rates are much higher than the sample rates in PCM. My feeling is for A>D 64 fs 2.8 MHz is enough, but for playback you really want 256 / 11 MHz

 

T+A are of the opinion that reproduction of frequencies up to something like 70 KHz is beneficial which may be going too far. However it is an indication that perhaps one needs to be careful that the playback system is transparent at very much higher frequencies than human hearing of sine waves.

 

I did not realize that HQP had a crossover function, quite intriguing. Could I then go to a 4 channel set up ? I need to read the manual ! An E28 feeding 2 tweeters, 2 woofers and a 3 sub woofer swarm with its own 48 KHz DSP to handle the room modes

 

Well, I have long thought that timing mattered more than level, i.e. timing errors should be dramatically less correlated than level errors. For a while it appeared that 24 bits was vastly more beneficial than a doubling of the 44.1kHz Redbook sampling rate. But that might have been only true for recording and reproduction with very good clocks. For years I saw no evidence save Julian Dunn's work (which I only vaguely understand), but the widespread adoption of sub-picosecond-jitter clocks is powerful testimony.

 

Another clue is the demand for higher sampling rates in DAWs for every need. I mean, 352kHz seems no longer enough... PeterSt and Jussi (Miska) have beaten this drum for a long time. Will 80-bit fixed processing at 700+kHz finally provide all the transparency of analog processing? I think the activity on output filtering is further evidence: high sampling rates and long words for both DSD-wide and PCM are now essential in processing. Once the spikes at harmonics of Fs are addressed, only the final HF output needs to be reduced by analog or digital filters, and are either or both still needed with DSDx4 or x8? I don't know.

 

Whoops, I can't confirm that HQP will cross over frequencies, but I think it should:

 

"Playback of stereo, 2.1, 3.0, quadrophonic, 3.1, 5.0, 5.1 and 7.1 channel material"

 

"For digital room correction and other equalization purposes, selectable convolution algorithms are included. Convolution engine supports RIFF (WAV) format FIR impulse responses, which can be produced with suitable software, such as Room EQ Wizard or DRC."

 

Can you figure that out? In my abode convolution is the sport of salad tossing.

 

Cheers

[EuroDriver]

Whoops, I can't confirm that HQP will cross over frequencies, but I think it should:

 

"Playback of stereo, 2.1, 3.0, quadrophonic, 3.1, 5.0, 5.1 and 7.1 channel material"

 

"For digital room correction and other equalization purposes, selectable convolution algorithms are included. Convolution engine supports RIFF (WAV) format FIR impulse responses, which can be produced with suitable software, such as Room EQ Wizard or DRC."

 

Can you figure that out? In my abode convolution is the sport of salad tossing.

 

Cheers

 

Hi Sam,

 

my understanding about convolution is that you can load a filter for room correction or tone control into the software convolution engine, and then when you pass a signal data stream through it, it will produced a modified signal stream with the desired filter attributes. So technically one should be able to take a stereo data stream split it in to 2 left streams, 2 right streams and a mono stream to feed 5 Chanel DAC and then 5 amps

 

I have only played with REW a bit in stereo, but I suspect it gets quite complicated in 5 channel !

 

My miniDSP Dirac systems works great because I have restricted correction to frequencies below 2khz

[Miska]

Whoops, I can't confirm that HQP will cross over frequencies, but I think it should:

 

You could run X-over filters with the HQPlayer's convolution engine. The item still missing is routing configuration to allow routing single input channel to multiple output channels. This has been on my TODO-list for quite a while and was planned to be ready already, but got delayed because I shifted focus to OS X support and spent some time designing DSC1 DAC...

 

Once it's done, it'll work the same for both PCM and DSD.

[EuroDriver]

You could run X-over filters with the HQPlayer's convolution engine. The item still missing is routing configuration to allow routing single input channel to multiple output channels. This has been on my TODO-list for quite a while and was planned to be ready already, but got delayed because I shifted focus to OS X support and spent some time designing DSC1 DAC...

 

Once it's done, it'll work the same for both PCM and DSD.

 

This has been my ultimate dream for more than a decade, I can wait for a few more months

 

Would the NAA be able to handle 5 Channels of DSD256 going over USB to a 5 channel DAC ?

 

HQ P for OSX and the trail blazing DSC1 take priority ;-)

[Sam Lord]

...you can load a filter for room correction or tone control into the software convolution engine, and then when you pass a signal data stream through it, it will produced a modified signal stream with the desired filter attributes.

Ah thanks, easier than tossing salad (for Jussi)

 

...The item still missing is routing configuration to allow routing single input channel to multiple output channels... ...Once it's done, it'll work the same for both PCM and DSD.

 

This is *very* exciting. That should take another couple of days; heck I'll even give you the weekend so you can fly to Brazil for the quarterfinals. But don't rush, we can wait till Tuesday. Oh wait, your SACD RIPPING software will make it Wednesday.

 

This has been my ultimate dream for more than a decade, I can wait for a few more days.

FIFY

 

Would the NAA be able to handle 5 Channels of DSD256 going over USB to a 5 channel DAC?

Geez that's about 100Mbps one way, for USB2 that's a tall order I think. Maybe you could survive hearing mid-fi DSD128 through the rear Bose stick-on units

 

HQ P for OSX and the trail blazing DSC1 take priority ;-)

 

Well sure if ********* loses, it is.

[paul3000h]

I have been using Foobar and Exasound E20, asio sdm type d to dsd 256 with definite improved sound all round.

 

I would like to know if the constant 50/65% cpu usage over 3 to 4 hours at a time is shortening the life of the cpu or computer.

 

Would be grateful for any insight here.

 

pic attached.

[bogi]

I would like to know if the constant 50/65% cpu usage over 3 to 4 hours at a time is shortening the life of the cpu or computer.

 

No. Many computer games are more demanding on computer resources.

[liquidsmoke]

What was disastrous was 24/192 material which had been sourced from original DSD 64. The conversion of this material to DSD128 or DSD 256 was strident and unlistenable. It appears that the record company's effort to produce an attractive sound in 24/192 PCM from a DSD64 original including adding "stuff" to make it sound sharper more etched. This "stuff" then becomes poisonous when it's converted to DSD.

 

Hi, Thanks for this really informative post This point is very interesting. So are you saying that if you have say a flac file, which you have originally sourced from SACD DSD64 rip, and converted to 24/88.2; then converting this 24/88 flac file back to DSD at play time is not a good idea?