[Adz523]

Maybe Thomas asked this but here is my question - any plans for a built-in upsampler converting the sample rate of all incoming digital or analog DSP signals to 192 kHz including decoded Dolby Digital and DTS program material??

[nicolasb]

a short list of our most-asked-for features:

How many times do I have to ask for Dolby Headphone in order for it to be classed as "most-asked-for"?

Always producing a stereo downmix (optionally Dolby Headphone? ) at (say) the tape or VCR outputs, regardless of the nature of the source, would also be good.

any plans for a built-in upsampler converting the sample rate of all incoming digital or analog DSP signals to 192 kHz

While upsampling as such is not very interesting to me, and there is (so far as I am aware) nothing you can actually use as a source for 192kHz recordings, one thing which I think is important is the question of jitter removal, especially for PCM sources. As thomaspf mentions, upsampling a signal asynchronously means that (perforce) you have to buffer and reclock the signal. I gather the sonic benefits of doing this with music CD sources is are quite significant - as evidenced by the B60DA and BP25DA. Are we going to have this on the SP1.7?

Other issues that have been unresolved for some time:

- Can we work out some way whereby upgrading an SP1.7 in Britain doesn't cost (literally) double what it costs to upgrade it in america?

- Will there be any improvements in the SP1.7's pre-amp stage, to bring it in line with the performance of the BP25? (Or is its measured performance already just as good?) An external power supply, for example?

- Does "multiple cross-over frequencies" mean per input, or per speaker? If the latter, does this mean that it will finally eliminate the business of losing the top part of the LFE channel if you have a low cross-over frequency?

- Will the "new" processor require a new back-plane? (For example, do you propose to offer balanced outputs for the surround back speakers?)

- What about a "height channel"?

- What about a headphone stage (as, again, seen in the B60 and BP25)?

- And what about the possibility of parametric EQ to (for example) compensate for the differences in output vs frequency from one model of loudspeaker to another?

[antt]

...one thing which I think is important is the question of jitter removal, especially for PCM sources. As thomaspf mentions, upsampling a signal asynchronously means that (perforce) you have to buffer and reclock the signal. I gather the sonic benefits of doing this with music CD sources is are quite significant - as evidenced by the B60DA and BP25DA.

I have to agree with nicolasb on this one, this feature does seem to have some significance for digital sources (not just CD sources).  I can't see how this function could be used in a home theatre system unless it is incorporated into the processor itself.  Is this being looked at?

[stp1200]

Here are a bunch of the question compiled into one post - hopefully I won't get into trouble with James for saying too much!

· Is this an DSP update only or are also plans to upgrade the DACs and digital interfaces to support higher resolution formats? 24/192 on coax? DVD-A on 1394 or HDMI (2ch 24/192 and multichannel modes?) SACD on 1394?

We're sticking with the same DACs because we think they sound awesome.  The analog section on the DSP module has been tweeked a bit for slightly better performance.  For the digital input receiver, we're switching to the CS8416, which is 24/192 capable.  We're still waiting a little on the high-speed digital interface.  HDMI is looking good to me over 1394 right now.  For one thing, it's way easier to design with, and way cheaper to implement.  I'm encouraged by the number of new displays hitting the market with HDMI.  And the latest HDMI spec (version 1.1) that I have read includes this:
   
Support for Super Audio CD will be defined in a future version of this specification.
   ACP_Type = 3 : Reserved for Super Audio CD (SACD)
   
ACP stands for Audio Content Protection (for those who are interested).  We'll know if it has truly been adopted when it starts showing up as a standard feature on processors.  (I know the new Meridian has it)

· How many times do I have to ask for Dolby Headphone in order for it to be classed as "most-asked-for"?  What about a headphone stage (as, again, seen in the B60 and BP25)? Always producing a stereo downmix (optionally Dolby Headphone?) at (say) the tape or VCR outputs, regardless of the nature of the source, would also be good.

These both require some hardware modification outside of the pin-for-pin replaceable DSP module, and won't be available in the initial upgrade release.  Once again, the capability is there on the software side within the DSP.  We're looking at ways to include this in the future.

· Will there be any improvements in the SP1.7's pre-amp stage, to bring it in line with the performance of the BP25? (Or is its measured performance already just as good?) An external power supply, for example?

The BP25 is a specialized and optimized product for stereo reproduction.  We won't be able to duplicate BP25 performance in the SP1.7, just because there are so many other things going on in the box.  That being said, the bypass performance is very good based on the feedback we get and our own measurements.  There are no plans currently for an outboard power supply.

· Does "multiple cross-over frequencies" mean per input, or per speaker? If the latter, does this mean that it will finally eliminate the business of losing the top part of the LFE channel if you have a low cross-over frequency?

The crossover frequencies are done in groups.  One setting for the Front (LR) speakers.  One setting for the Center speaker.  One setting for the Surround Speakers.  And one setting for the Back speakers.  The Sub will adopt the highest of the previous settings.  And you will be able to save different values for each input source.

· Will the "new" processor require a new back-plane? (For example, do you propose to offer balanced outputs for the surround back speakers?)

The new processor will not require a new backplane for those who will be upgrading.  We are looking at a new backplane as an option.

· What about a "height channel"?

No plans at this time to move beyond 7.1. We've all heard of the 10.2 systems, but there doesn't seem to be a lot of demand for them right now.  At AES, we saw a 22.2 system, and someone from Fraunhofer told me that they're working on a 304.8 system! (I had a hard time comprehending that number, so I hope that I heard her correctly). That system is over-the-top and I think is initially targeting the theme-park market.  It's pretty cool technology though, check it out at http://www.iosono-sound.com

· And what about the possibility of parametric EQ to (for example) compensate for the differences in output vs frequency from one model of loudspeaker to another?

This is on the roadmap, as the capability is there within the DSP.  I'm not sure it will make it into the first version, but it will definitely be available as a downloadable software upgrade.

· While upsampling as such is not very interesting to me, and there is (so far as I am aware) nothing you can actually use as a source for 192kHz recordings, one thing which I think is important is the question of jitter removal, especially for PCM sources. As thomaspf mentions, upsampling a signal asynchronously means that (perforce) you have to buffer and reclock the signal. I gather the sonic benefits of doing this with music CD sources is are quite significant - as evidenced by the B60DA and BP25DA. Are we going to have this on the SP1.7? Any plans for a built-in upsampler converting the sample rate of all incoming digital or analog DSP signals to 192 kHz including decoded Dolby Digital and DTS program material? Any plans around better decoupling the DAC clock from the input signal or in lack of that maybe a master clock input?

We will be featuring a dsp-based synchronous upsample function with the new DSP.  It's operation will depend on the other modes that are active, especially post-processing modes.  It will have the ability to upsample at double-rate and quadruple-rate to a maximum of 192kHz. We'll probably do something like have a minimum and maximum samplerate setting, and the software will choose the upsample rate based on available resources.

The jitter-reduction issue is a tough one for us.  There are some hard issues about doing it without destroying a Dolby or DTS bitstream.  It would certainly add to the cost and complexity of the unit.  It will take some hardware modification, and we hope that we can find a reasonable way to address this issue in the future.  Alas, we won't see it in the initial upgrade release.

[jethro]

Shane,

Thanks for all the info on proposed changes.

If you go with a new backplane at some point in the future, have you considered going with hardware handshaking on the RS-232 port to make it more reliable for use with software updates via the serial port ?

I bring this up because the OSD on my SPV can do some pretty strange things and I can't help but guess at least some of the problem lies with using software handshaking on the serial port.

Thanks.

[nicolasb]

All good stuff, Shane, thank you. Inevitably I have a "few" requests for clarification.

How many times do I have to ask for Dolby Headphone in order for it to be classed as "most-asked-for"?
...

These require some hardware modification outside of the pin-for-pin replaceable DSP module, and won't be available in the initial upgrade release. Once again, the capability is there on the software side within the DSP. We're looking at ways to include this in the future.

I can see how my more fanciful requests (such as a headphone jack!) would require additional hardware, but would Dolby Headphone output alone require hardware modification? In particular, suppose I have a 5.1 (rather than 7.1) speaker/amp set-up - could a Dolby Headphone signal not be sent to the otherwise-unused surround-back channel outputs?

In some ways this would be more useful than having an actual headphone jack, as it allows 'phonephiles to use their own choice of dedicated headphone amp.

The crossover frequencies are done in groups... Front (LR) speakers... Center speaker... Surround Speakers... Back speakers. The Sub will adopt the highest of the previous settings.

That's good, but you didn't answer my question about losing the LFE channel.

As it stands, the subwoofer output is generated by mixing the LFE channel with all the signals for which the corresponding speaker is set to "Small", then applying a low-pass filter to the mix, and then forwarding what's left to the subwoofer. So, if you have a cross-over set to (say) 30Hz and you watch a 5.1 movie, then everything above 30Hz in the LFE channel is being filtered out at 24 dB/octave.

It seems likely that, if you can have different cross-over frequencies for different speakers, then the low-pass filtering must be being done before the "Small" channels are mixed with the LFE signal rather than after, which would solve this problem. But I just wanted to check that explicitly.

What about a "height channel"?

No plans at this time to move beyond 7.1.

I was thinking of something more like the Tag McLaren solution. Any actual piece of present-day source material will be, at most, 6.1. If you have a 7.1 speaker setup then the surround back speakers are both producing the same signal, so, in principle, you could drive both of them from the same output (this obviously assumes that you don't require different channel delays or volume settings).

So, given an actual 7.1 source (with a height channel), it's played back by sending the surround back signal to one surround back output, and the height channel to the other. You can then either have just one surround back speaker, or use a splitter cable if you want two.

And what about the possibility of parametric EQ

This is on the roadmap, as the capability is there within the DSP. I'm not sure it will make it into the first version, but it will definitely be available as a downloadable software upgrade.

Hooray!

Will this be based on signal frequency only, or could it also depend on signal amplitude?

No chance of TacT-style time-domain correction I suppose?  

We will be featuring a dsp-based synchronous upsample function with the new DSP.

But not asynchronous (as in the BP25DA)? That's a shame.

The jitter-reduction issue is a tough one for us. There are some hard issues about doing it without destroying a Dolby or DTS bitstream.

While DD and DTS streams are certainly affected by jitter, I get the distinct impression that it is more significant for PCM. This is precisely what you get from upsampling a 44.1kHz signal to 96 kHz asynchronously, of course - if the upsampling is asynchronous then, perforce, the signal is being reclocked.

Maybe you could market a stand-alone pure-digital jitter-removal device that sits in between the DVD player output and the processor input? Or is S/PDIF so rubbish that this wouldn't actually help anyway...?  


And of course the final question is: do we have any idea when all this might be happening?

[thomaspf]

Thanks Shane,

I think you are right about HDMI. This will likely take off for the masss market but it will be a while before SACD will be made to work. I am encouraged that you have a hires multichannel digital input on your radar.

Since the current DACs which are indeed very nice do not support DSD I assume you will have to do DSD->PCM conversion?

I like the idea of a synchronous sample rate converter. The asynchronous solutions basically convert different samples every time you play a track. They do estimate very well but even knowing it makes me feel uneasy.

If you follow the design patterns of the Apogee DACs or Weiss DACs you can build a pretty low jitter design by a small buffer before the DACs. You continue to process everything the way you do now but write the PCM streams into the buffer insteads of the DACs directly.

Then you implement a very low jitter and stable local clock that you need to carefully and slowly regulate to match the clock reconstructed from the incoming signal. The DACs are driven by that clock and synchronously read samples out of the buffer.

Even for the less preferred solution with an asynchronous sample rate converter you would put these right before the DACs and so they will not interfere with AC3/DTS processing.

Here is another thought for you. If you consider a USB input you can make the whole processing chain driven by the local clock. USB audio has an asynchronous mode where you basically extract all your data at the local clock speed and there is an asynchronous bidirectional protocol back to the PC to make sure you never run out of data. Pioneer implemented something very similar over their 1394 interfaces to improve the jitter peformance over the standard A&M 2.1 profile which mandates isochronous transfer.
 
For movies this obviously a bit tricky because of the lip A/V synchronization but for music this makes for a very simple design and implicit low jitter. If you ever revisit the 2 channel solution this could be a nice improvement.


I hope you don't mind me telling you how to build your processor

Cheers

    Thomas

[thomaspf]

While I am in whishful thinking mode let me throw in one feature that I really would like to see. Parametric eq and room correction is all nice but since I am using the SP1.7 at home and not in the studio my top issue is loudness restoration.

I don't want to suffer from hearing damage so I never watch movies at reference level. I think the loudest scenes rarely get over 85db combined level at the seating position. To music I often listen even lower than that.

Naturally Fletcher-Munson kicks in and everything just sounds wrong if not played at the recording level. I have seen that some units now offer digital loudness restoratio.

Are you so much focused on the studio market where reference level is probably more common that you don't feel this is a priority feature?

Cheers

   Thomas

[antt]

While I am in whishful thinking mode let me throw in one feature that I really would like to see. Parametric eq and room correction is all nice but since I am using the SP1.7 at home and not in the studio my top issue is loudness restoration.

I agree.  I remember back in the stereo days when many products had loudness compensation knobs.  Most of the high end products did not, including Bryston.  I understand the philosophy behind that, but I don't think it applies to home Theatre.

  Please correct me if I'm wrong, but DVD's take their sound tracks from the motion picture tracks created for movie theatres.  And these tracks are recorded for (what I consider) deafening sound levels.  Like thomaspf, i prefer to listen to movies at low volume levels, not to mention I would probably get sued by other suites in my building if I played em at reference level.  

I consider the current state for low volume playback technology to be woefully inadequate.  The dynamic range compression built into (some?) DVD's really leaves alot to be desired.  And until studio's start recording tracks specifically for home theatre, I expect this to remain the status quo.

 Is Bryston looking at further options in this area?  I think one of the most effective solutions currently available would be a Loudness compensation.

[jimmyp58]

Geez, I must be odd as I enjoy not only listening to movies, but feeling them as well.  If I wanted to watch a movie in a quiet manner, I'd turn on a movie that was playing on the tv and watch it from that source instead.

[James Tanner]

HI All,

A large portion of DVD;s are now specifically mixed for home theater application. It is true that in the past most were just copies of the original large venu mix but that is no longer the case.

james

[thomaspf]

Hi James,

help me understand this statement. I have my signal and after decompression I end up with 24 bits. A full scale signal is still being full scale and when I calibrate my SP1.7 to reference levels this will be really loud. Now, I don't think the studios are lowering the resolution or the average loudness of their sound tracks. Dolby has an interesting  paper on their WEB site defending the encoding standards in face of the studios attempts to raise the average loudness with every movie. I have not seen a change in dialnorm either.

http://www.dolby.com/assets/pdf/tech_library/54_Moviestooloud.pdf


So in summary, I don't know what a home theater mix is and how that reduces the average SPL level on a system calibrated to reference levels. When I use a SPL meter I find newer movies are much louder on average. I am looking forward to understanding this better and keeping my hearing intact for a few more years.

Cheers

   Thomas

P.S. If you want an extreme view that I would of course never mention one could argue that regular movie watching on a calibrated SP1.7 is a THX certified health hazard.

[James Tanner]

Hi Thomas,

I was not commenting so much on the loudness issue as much as making the point that the frequency balance and perpective of the DVD for home applictaions was being done in a small room rather than simply taking the large theater venu mix and putting it on a DVD.

james

[thomaspf]

Hi James,

I did not know that but it is absolutely great to hear. So, the AC3 tracks on a DVD are actually eq'ed to sound correct at lower than the reference level?

Do you know what the target SPL is for these mixes for lets assume a 1Khz sine wave at the dialog level 31db down from full scale and with dialnorm set to -27 found on most DVDs.


Edited: Let's try this more practical. I assume the test tones in SP1.7 are -20db down from full scale. Do I need to change the target level from 85db to make these new home mixes sound correct in my theater or are the absolut levels in the tracks lowered together wit the eq. Is there some logo on these DVDs that tells me about this?

Cheers

   Thomas

[James Tanner]

Hi Thomas,

No I don't know but I will forward to one of the recording engineers I know in Holywood and see what they have to say.

james

[nicolasb]

I'd be interested to hear the answer, James.

The fact that DVDs are "mixed for home theatre" could mean almost anything. There's certainly no reason to assume that they are mixed to be listened to at lower-than-reference-level volumes.

Even if some of them are, as Thomas says, how do we know which ones?

And what if, occasionally, we want to listen to a film late at night? Even if it is balanced to sound okay at -10dB, it still won't sound right at -20dB.

I have to agree with thomas' original point, namely that it would be very useful if the SP1.7 could apply loudness correction based both on a specified (maybe variable) reference volume, and on whatever the current playback volume is.

And, as a side issue, I'd also quite like to see Dynamic Range Compression made available for non-DD sources - but I'm probably the only person who would.

[jimmyp58]

When I want to listen to a movie late at night, I throw on my headphones so that I don't disturb the family.

[nicolasb]

When I want to listen to a movie late at night, I throw on my headphones so that I don't disturb the family.

Yeah, but that's a bit of a waste of a 5.1 soundtrack, isn't it?

'Course, it would be less of a waste if the SP1.7 had Dolby Headphone.

(What do you mean "obsessed"? )

[nicolasb]

Ooh! I forgot a very important question.

Shane, are we going to see HDCD support in the SP1.7 any time soon?

[thomaspf]

HDCD would be pretty nice and is pretty low hanging fruit and could be easily implemented on your fast DSPs.

Cheers

    Thomas