MDAC first listen (part XVIII)

[Asphaltradler]

A large proportion of DG's DDD are indeed rubbish, their 4D series (a variant of DDD) if anyone could still remember is a laughing stock, but a few of their first generation DDD (early 80's made in West Germany) are actually very good albeit very expensive and difficult to find now. As to exemplars of good classical recording, I think most of them (>70%) are ADD, particularly recording of the 60's and 70's, and EMI and DECCA generally has done better than DG.

I always liked very much the recordings Denon made in the 80s. No compression, mostly done with only 2 excellent omni-characteristic microphones (Brüel&Kjär). Especially the Mahler Symphonies recorded in the Alte Oper Frankfurt, directed by Eliahu Inbal are a delight. Was there myself several times in the 2000s and the atmosphere of the location is captured really astonishing, especially when listened with headphones. Don't know why Denon seems to make no CDs any more.

There is one very serious problem, though, with them, as with some very good Telarc and Sheffield Lab productions of the time: the CDs are manufactures with Pre-Emphasis, and unfortunately the M-Dac as any DAC doesn't know about that and produces very harsh sound with these CDs.

It's possible to compute the Emphasis out with a tool called sox, so one can play them via PC at least over M-Dac.

John, would it be possible to detect the Emphasis directly from CD Player and do anything against it?

In more recent times quit good recordings came from MDG, Acousence (also High-Res downloads), and Telarc is not so bad, too, if perhaps a bit on the bassy and "spectacular" side.

Asphaltradler

 

[Asphaltradler]

Which is precisely what my ears told me.

Yes, you're right, and Mea culpa! But sometimes I want to SEE what I HEAR, even if that is nonsense... A better feeling nevertheless and quite interesting to measure oneself with amateur equipment and come to similar conclusions as the "Pros". Perhaps that is a Physicist disease...

 

[ridney]

I connected mine to a Synology NAS but the Synology music player was limited.

Now using Windows 7 with JRiver as music player which is much better.

You could route your NAS through a Squeezebox Touch using the NAS as a library- using Triode's app.

http://thewelltemperedcomputer.com is a good resource.

Thanks MD, currently I am also using Windows 7 with MediaMonkey streaming from my NAS and the sound is great but was looking for a cheap player that I could use so that I don't have to turn on my pc every time I want to listen to music. I thought the app from the NAS would be enough for basic playback but didn't get any sound at all. I'm also thinking of using an iPhone in a dock as a player as I have a spare and digital out to MDAC, same as what I am doing with my Zeppelin dock. I just forgot the name of that dock with digital out.

 

[Dominik]

There is one very serious problem, though, with them, as with some very good Telarc and Sheffield Lab productions of the time: the CDs are manufactures with Pre-Emphasis, and unfortunately the M-Dac as any DAC doesn't know about that and produces very harsh sound with these CDs.
...
John, would it be possible to detect the Emphasis directly from CD Player and do anything against it?

I'm only surprised it's taken this long - de-emphasis popping up! Yes, I can make the M-DAC apply de-emphasis, but this will only work when playing back an actual CD through a disc player. I'll include that in new firmware releases.

For computer playback, the only option is to grab and apply the filters offline. When doing so, make sure the output bitdepth is set to 24 bit and dithered properly.

Same goes for HDCD - if you have them, grab them, run them through software HDCD decoder and save as in 24 bit resolution. No dither needed there since most of the time, HDCD will show 17 bit resolution, going to 20 bit during peaks. When stored losslessly, there will be no difference in file size between 20 bit or 24 bit, but 20 bit is rather obscure format and may not be handled properly by some players.

 

[Dominik]

Hi John

Looks complete to me.

Only leaves, to quote you:-

"Later add USB to AC3 pass-though, and maybe SPDIF input to USB out (for SPDIF input to your PC).

Also SQ "optimisation" once a get a decent system going - with MPAX & MAMP...

This will be only a "Lakewest" release..."

I've got AC-3/DTS or any other bitstream pass-through working already. Mac supports it natively, apparently so does Windows but only from Vista up - haven't checked as I run XP only. Better still, M-DAC scans the data (any input, not just USB) for compressed sync patterns and can detect and show which codec is used - AC-3/DTS/AAC/WMA. It will also detect DTS-CD, mute the analog output and flag the digital output as non-PCM. On Windows XP, the USB Audio driver does not support bitstream output, but with DirectShow filter such as Spdifer, you can play the bitstream as if it was regular PCM audio. This in an itself would not do the trick due to the infamous KMixer molesting the data, but thanks to M-DAC's Lsb Restoration feature, the data gets - well, restored and is immediately identified, output muted and non-PCM flag on S/PDIF output set.

 

[Bazz]

I terribly miss not being able to stream BBC iPlayer by way of Apple TV3 due to the sound dropping (yes, I tried all the tricks).

What I don't understand is why Apple TV would work perfectly with e.g. my old Marantz receiver why not with better DACs such as the MDAC (I read that some Cambridge owners have the same problem).

Isn't it possible to program some option in a coming MDAC update that copes with the AppleTV fault?? John? ExtraExtra high jitter rejection, is must at the cost of SQ?

 

[Plutox]

I've got AC-3/DTS or any other bitstream pass-through working already.

Hi Dominik -

while you're working in that area, as someone mentioned a few posts ago, would it be possible to indicate on the front panel when a pre-emphasis flag is detected in the incoming stream?

You won't be able to do anything about it, but it might be useful to indicate that the audio output will be over-bright!

 

[Dominik]

I terribly miss not being able to stream BBC iPlayer by way of Apple TV3 due to the sound dropping (yes, I tried all the tricks).

What I don't understand is why Apple TV would work perfectly with e.g. my old Marantz receiver why not with better DACs such as the MDAC (I read that some Cambridge owners have the same problem).

Isn't it possible to program some option in a coming MDAC update that copes with the AppleTV fault?? John? ExtraExtra high jitter rejection, is must at the cost of SQ?

If there was something we could do about it, we would. The High bandwidth setting is already the most forgiving there is in the ESS Sabre DAC.

Traditionally S/PDIF sources all used a fixed crystal oscillators to drive the S/PDIF timing. There may be jitter, even high levels of it, but there's no wander and certainly no sudden changes to the nominal frequency on signals like that. The Sabre's S/PDIF receiver as well as the Wolfson S/PDIF receivers in M-DAC preceding it can tolerate obscene levels of period jitter, data correlated jitter and attenuate it to levels that simply cannot be distinguished from a clean signal using professional jitter analyzer test equipment. The problem lies with the DPLL implementation of the Sabre which cannot tolerate sudden discrete jumps in incoming frequency without going tits up and having to re-lock to the new rate. It's because there's no fail-safe mechanism for situations like that. It's not very reasonable to expect such things going on either as most sources either run a fixed or continuously/smoothly varying clock. It's a Apple's crude way of synchronizing playback rate with the source device in Airplay system that is outside of the general practice that has these unforeseen consequences.

As to why other devices have no problem with such sources - well they're using a very simple S/PDIF receivers with PLLs that have no attenuation in the audio band and can only attenuate jitter at RF, enough to successfully decode the data carried in the S/PDIF. With such a wide open loop bandwidth they are able to contain the discrete jumps in clock frequency without getting unlocked.

If ATV3 is ever to work with M-DAC and the others that are having issues, Apple would have to tweak their software to do a better job at rate adaptation with smoother transitions, not their ON/OFF control between slower and faster rate. I'm not even sure the WiFi router application processor can provide fine granularity in it's clock generator.

 

[Dominik]

Hi Dominik -

while you're working in that area, as someone mentioned a few posts ago, would it be possible to indicate on the front panel when a pre-emphasis flag is detected in the incoming stream?

You won't be able to do anything about it, but it might be useful to indicate that the audio output will be over-bright!

Hi Plutox,

No need, the M-DAC will apply proper de-emphasis when it detects the flag. Only there's no flag when playing via computer.

 

[Bazz]

If there was something we could do about it, we would. The High bandwidth setting is already the most forgiving there is in the ESS Sabre DAC.

Traditionally S/PDIF sources all used a fixed crystal oscillators to drive the S/PDIF timing. There may be jitter, even high levels of it, but there's no wander and certainly no sudden changes to the nominal frequency on signals like that. The Sabre's S/PDIF receiver as well as the Wolfson S/PDIF receivers in M-DAC preceding it can tolerate obscene levels of period jitter, data correlated jitter and attenuate it to levels that simply cannot be distinguished from a clean signal using professional jitter analyzer test equipment. The problem lies with the DPLL implementation of the Sabre which cannot tolerate sudden discrete jumps in incoming frequency without going tits up and having to re-lock to the new rate. It's because there's no fail-safe mechanism for situations like that. It's not very reasonable to expect such things going on either as most sources either run a fixed or continuously/smoothly varying clock. It's a Apple's crude way of synchronizing playback rate with the source device in Airplay system that is outside of the general practice that has these unforeseen consequences.

As to why other devices have no problem with such sources - well they're using a very simple S/PDIF receivers with PLLs that have no attenuation in the audio band and can only attenuate jitter at RF, enough to successfully decode the data carried in the S/PDIF. With such a wide open loop bandwidth they are able to contain the discrete jumps in clock frequency without getting unlocked.

If ATV3 is ever to work with M-DAC and the others that are having issues, Apple would have to tweak their software to do a better job at rate adaptation with smoother transitions, not their ON/OFF control between slower and faster rate. I'm not even sure the WiFi router application processor can provide fine granularity in it's clock generator.

Thanks Dominik for your clear reply. At least now I understand. I hope Apple is aware of the problem and sets it right. There's lots of people complaining about this but (at usual) no comments from Apple. If only they would have the same dedication to their customers as John and you!

 

[EIffel]

Hi Plutox,

No need, the M-DAC will apply proper de-emphasis when it detects the flag. Only there's no flag when playing via computer.

Dominik, it might still be helpful to show that deemphasis is applied on a particular CD, as this may help diagnose why rips don't sound right, and would lead me to take the appropriate corrective action (sox, etc.)

 

[Asphaltradler]

Dominik, it might still be helpful to show that deemphasis is applied on a particular CD, as this may help diagnose why rips don't sound right, and would lead me to take the appropriate corrective action (sox, etc.)

Well it's unexpectedly good news that De-emphasis can even be applied, thanks very much to Dominik! Hadn't dreamed that was possible. For me correction when playing from CD is perfect, when I ripped CDs to PC I already de-emphasized them with sox, and new CDs with Pre-Emphasis are very rare.

@Dominik, why should I rip with 24bits? You think of quality loss by the sox filter alogorithm?

@Eiffel, detection of Pre-Emphasis is quite easy - just switch between CD-Player analog out and M-Dac and the difference is obvious to the ear . But you're right, an optical indicator would sure be nice, too.

Asphaltradler

 

[Plutox]

the M-DAC will apply proper de-emphasis when it detects the flag

Thanks for that - I didn't appreciate that the M-DAC had enough DSP clout to make audio changes. You'll have people asking for tone controls next

 

[KAP]

Thanks Dominik for your clear reply. At least now I understand. I hope Apple is aware of the problem and sets it right. There's lots of people complaining about this but (at usual) no comments from Apple. If only they would have the same dedication to their customers as John and you!

Since a month, I have the M-DAC. And I am very satisfied. I use an Apple TV 3 and so far I have absolutely no problem of drop-outs.

 

[JohnW]

Thanks for that - I didn't appreciate that the M-DAC had enough DSP clout to make audio changes. You'll have people asking for tone controls next

Erh - much simpler then that - the ESS has an de-emphasis filter - just needs to be set by registers.

Lets see if we can add a de-emphasis indicator on the MDAC...

The only CD I can recall that I own with de-emphasis was a Human League CD, now long lost during my many moves around the world...

John

 

[Plutox]

..a Human League CD, now long lost...

Perhaps that's best, really

 

[Dominik]

Since a month, I have the M-DAC. And I am very satisfied. I use an Apple TV 3 and so far I have absolutely no problem of drop-outs.

Are you sending audio to it via Airplay and to M-DAC with no issues? And I mean audio only - not as part of a video. Two very different beasts.

Since Airplay is source-driven timing system, maybe you can strike it lucky with a particular Airplay source device you'd use that's not making the ATV3 jump about as much as with another.. So far there wasn't any happy story reported by ATV3 owners though..

 

[JohnW]

Thanks Dominik for your clear reply. At least now I understand. I hope Apple is aware of the problem and sets it right. There's lots of people complaining about this but (at usual) no comments from Apple. If only they would have the same dedication to their customers as John and you!

In case you missed my earlier post with the link to Apple Support Communities...

https://discussions.apple.com/thread/4104819?start=75&tstart=0

Appears ATV3 works better then AE but outputs only at 48KHz (therefore upsamples CD to 48KHz which is not good news).

 

[KAP]

Are you sending audio to it via Airplay and to M-DAC with no issues? And I mean audio only - not as part of a video. Two very different beasts.

Since Airplay is source-driven timing system, maybe you can strike it lucky with a particular Airplay source device you'd use that's not making the ATV3 jump about as much as with another.. So far there wasn't any happy story reported by ATV3 owners though..

I send my audio from my Mac mini (10.8.2, iTunes 11.0.1) over Ethernet via modem/router (TG789) to the Apple TV (software 5.1.1) and then via optical to M-DAC.

 

[JohnW]

Perhaps that's best, really

I admit I was hardly in tears..