Keith_W system

[hvbias]

Thank you gents. The diagrams are necessary, because if I were to attempt to explain it in words, it might be rather difficult to read.

Right now I don't know how to take measurements of the system once the multi-channel DAC is installed. As I understand it:

1. Acourate is not able to use more than one ASIO driver at the same time, meaning that it can not send output through the NADAC and receive input from the mic preamp at the same time.

2. The clock on the ADC (i.e. mic input) needs to be synchronized with the clock on the DAC.

Not sure how to do this at the moment. Will hopefully be cleared up in the next few days. Maybe dallasjustice or Blizzard might like to say something about the design of the system or how I could do it better.

This is what I sketched up for my 4 channel setup (I may go for more subs, still need to read up about Welti config) that I was thinking about. I'm not entirely sure if it will work, and I'm still going through old group messages before I post this question to the group. It should be able to run off a single ASIO driver. Fortunately it seems like there are plenty of gear slutz that hang out there

The reason I chose Lynx is I've had one of their PCI cards in the past and they have long standing support even when products go legacy. What I need to find out is if the ADC and AES digital outputs are all synced to the same master clock. The Lynx LT-Thunderbird is a single connection from PC to Aurora. I have a single Yggy right now and love it.

Edit: this not a slight against the NADAC or any other pro DACs. I've had a couple of different Benchmark DACs (ancient rackmount DAC1 and one when they came out with a USB model). These are superbly well measuring DACs, but I never warmed to the sound of them. Since I'm not blind testing myself when comparing DACs, my subjective impressions are what matter to me. I wouldn't convince anyone else what I'm hearing is even right, this is just my personal justification for why I prefer to stick with 2-channel DACs I have personal experience with.

 

[TBone]

I bought a bunch of used CDs last year and a couple caused serious issues with dbPoweramp getting stuck forever and not able to rip them. I overrode the error checking/detection and ripped them that way.

EAC hangs the same depending on the error correction setting.

However, when it comes to the copy-protected cd's, EAC prebeta 0.95 (iirc) was the last version offering a go-around.

 

[Keith_W]

Keith, what is your subjective impression of the MPS-5? The only gear I'm familiar with are the Cary's, to some extent; did hear the Playback Design and that peaked my interest ...

I'm not easily impressed, and it would take a lot of impressing to make me open my wallet and shell out $15k. The MPS-5 replaced my old SACD player, a Cary CD306. The Cary had a nicer tone, the MPS-5 sounds more "dry". But it also retrieves more detail, and never sounds harsh. I love my MPS-5, and I will never sell it.

Keith, that's what's missing in your systems; an analog rig with tapes and LPs.

I did have an analog rig, right up to a month ago. I found that I wasn't listening to it very much, so I made a decision to let it go in favour of going fully digital.

 

[Keith_W]

This is what I sketched up for my 4 channel setup (I may go for more subs, still need to read up about Welti config) that I was thinking about. I'm not entirely sure if it will work, and I'm still going through old group messages before I post this question to the group. It should be able to run off a single ASIO driver. Fortunately it seems like there are plenty of gear slutz that hang out there

The reason I chose Lynx is I've had one of their PCI cards in the past and they have long standing support even when products go legacy. What I need to find out is if the ADC and AES digital outputs are all synced to the same master clock. The Lynx LT-Thunderbird is a single connection from PC to Aurora. I have a single Yggy right now and love it.

I could potentially go down this route as well. I have a RME HDSPe AES sound card in my PC, which is able to output AES/EBU in multiple channels. In fact, my original plan was to send the mid/high frequencies to a very good DAC, with cheaper DAC's for the low mids and bass. Someone pointed out that these DAC's are not clock sync'ed so it won't work.

Can your Yggy's can be clock synced? Have you checked?

 

[TBone]

The MPS-5 replaced my old SACD player, a Cary CD306.

Auditioned a 306 a long time ago, friend settled on a 303, after a long run of separates.

The Cary had a nicer tone, the MPS-5 sounds more "dry".

I get the dry comparison.

I love my MPS-5, and I will never sell it.

I get that also ...

 

[hvbias]

I could potentially go down this route as well. I have a RME HDSPe AES sound card in my PC, which is able to output AES/EBU in multiple channels. In fact, my original plan was to send the mid/high frequencies to a very good DAC, with cheaper DAC's for the low mids and bass. Someone pointed out that these DAC's are not clock sync'ed so it won't work.

Can your Yggy's can be clock synced? Have you checked?

I hadn't considered that. I assumed since they were both the same DAC model there wouldn't be issues slaving them to the Aurora's master clock. I'll have to find out with Schiit.

Edit: I just saw the latest posts from today in the group. You're right I'm not sure my solution will work since there is no word clock in/out on the Schiit.

Emm Labs also make multichannel converters with word clock in/out. I like the sound of their DACs I've heard.

 

[Keith_W]

Well, it's up and running. There are a few problems, however.

1. Computer is connected to the DAC via USB, and not via AES/EBU. This is because the pro sound card I am using (RME HDSPe AES) uses a D25 sub output for AES/EBU, whereas the Playback Designs MPS-5 SACD player (being used as a DAC) uses the conventional AES/EBU (XLR) socket. I will need to get a breakout cable for the above.

2. With no convolution, the PC as a source sounds awful, compared to playing a CD in the MPS-5. On the MPS-5, notes have a nice, clear, defined edge. On the PC, it sounds rounded off. And this is with using HQPlayer.

3. For some reason, the Playback MPS-5's USB input maxes out at PCM 48kHz. It should be capable of much higher than this. I need to read the manual or update the firmware - something to do later.

4. It's noisy! There is a background hiss going through the system, and doing things on the computer, even switching windows, sends a click through the system.

5. The aesthetics of the playback software (HQPlayer) on the PC leave much to be desired. I know I can use either Roon or Muso as the front end instead, but I hate both of them!

Anyway, I fired up Acourate. This is the result of my very first measurement sweep:

You can see that the high frequencies are -10dB down compared to the midrange. Also a downward curve in the upper bass, above 100Hz.

I then designed a target curve. Since it's my first, I decided to make it flat. I told Acourate to invert the frequency response with respect to the target curve, and used it to generate a convolution filter. I then redid the measurement:

WOW! It's almost flat, apart from the dip in the bass which is a room issue (I decided not to touch this).

I loaded the convolution filter into the HQPlayer convolver engine and started playing music. Jaw dropping. Close to the best I have ever heard my system.

Right now, only the frequency response and room correction has been fixed. I will not be able to perform any of the advanced functions of Acourate (time alignment of the drivers, individual driver correction) until I obtain more DAC channels and thus be able to individually measure each driver.

I will have to slowly work through all the issues above. But for now, it all bodes well.

 

[amirm]

Interesting. Play some music with a lot of high frequencies and see if the response there is acceptable. I usually find flat response like that too bright.

 

[Keith_W]

Yes Amir, I think it does sound a little too bright. Or rather in my case, the mids and bass sound recessed since I turn the volume down. I will redo the target curve when I get home tonight. Probably go with Dr. Olive's recommendation of starting a gentle downward slope from 1kHz until it is -6dB at 20kHz.

If anyone else is running DSP, I would love to hear how you designed your target curve.

For a first effort, I am pretty impressed by what it can do.

 

[RayDunzl]

If anyone else is running DSP, I would love to hear how you designed your target curve.

I tend to use a flat curve myself, not that I can hear those high frequencies...

My speakers don't spray the walls with HF, I hear direct, dipole bounce, and whole room length bounce, verified by recording an impulsive 'click' and slowing it way down to see what's hitting me.

https://soundcloud.com/ray-880875693/pulse-echo

And, I compare what is 'heard' at the listening position (corrected) with the uncorrected source. When I look at the curves on the source material - they're all over the place, so my excuse is that I go with the "artist intent".

If I ever get a complaint, it can be easily changed.

Picture below: Some Arvo Part Organ Music playing. Top - room RTA. Middle - left source RTA. Bottom - right source RTA.

And for most listening, I'm set to mimic the CD output level with the room SPL, hence the SPL meters on the left.

 

[Fitzcaraldo215]

Yes Amir, I think it does sound a little too bright. Or rather in my case, the mids and bass sound recessed since I turn the volume down. I will redo the target curve when I get home tonight. Probably go with Dr. Olive's recommendation of starting a gentle downward slope from 1kHz until it is -6dB at 20kHz.

If anyone else is running DSP, I would love to hear how you designed your target curve.

For a first effort, I am pretty impressed by what it can do.

I really did not "design" my target curve. But, I am strongly in agreement with the proposition that a downward sloping with increasing frequency curve sounds most "natural" to me. I currently use Dirac Live, whose curve has that characteristic but with a somewhat accelerated downward slope above about 15k. I have left it alone and I am quite happy with it.

Commentators over the years have mentioned this downward sloping in-room response concept. I nearly fell off my chair when Gordon Holt first mentioned it in Stereophile decades ago. I think acousticians widely accept that this downward slope is one thing that must be done in order to create a perceptually flat sound. I think it is a compensation for the difference between how ears hear at higher frequencies vs. omni calibration mikes. Unfortunately, while there is general acceptance of the concept, actual empirical, perceptually flat target curves differ slightly due to the expected variations from perceptual experiment to experiment. But, I agree, that "measures flat" is far too bright. It is hard for audiophiles to accept that, given the nice flat curves they have seen for electronics over the years. But, in-room mike acoustic calibration is a different animal.

One can get carried away if one has the tools to do so, like target curve design for EQ. I have a good friend who also uses Dirac. But, he is a very different kind of listener than I am. He listens to many more CDs than do I. I listen mostly in hi rez. For whatever reason, he constantly complains of "digititis", and he has developed a series of target curves meticulously fine-tuned by ear that attenuate the HF to an even greater degree. I keep telling him that he is trying to compensate for digital filter issues, which are in the time domain, by using EQ in the frequency domain via the target curve. My pleadings are in vain. But, he just likes to fiddle. He has spent hours and hours on it. Hey, it's his system. What can I say? I would rather spend all that time just listening to the music, not the target curves.

 

[Keith_W]

A friend came over last night and we did more listening tests in various different configurations. The outcome is pretty disappointing.

Bear in mind that ALL the comparisons were done with the same DAC, with volume matching via an SPL meter. The only difference was the digital source!

Files on the computer were ripped using Exact Audio Copy and stored in FLAC. This was compared against the original CD.

1. CD Player --> Preamp.
Easily sounds the best. As I mentioned earlier, notes sound "whole". You can easily tell the start and stop of a note, whereas on the computer it sounds as if the edges are rounded off. When instruments are massed, it is easy to pick out individual instruments. The computer tends to smear them.


2. Computer optical drive --> HQPlayer with NO convolution --> USB --> DAC --> Preamp
The purpose of this configuration is to test the adequacy of the optical drive on the PC to the CD player transport. No contest, CD Player wins hands down. So much for "bits is bits" and the marketing of Playback Designs which claims to "completely eliminate" jitter.


3. HQPlayer with Acourate convolved filters @44.1kHz --> USB --> DAC --> Preamp
(Note, I have to use 44.1kHz because the DAC only accepts a maximum of 48kHz via USB input)

Improves on the CD player in some ways - the system sounds more coherent, bass in particular is very good. However - notes are smeared, and in some passages I can hear some kind of digital noise. Hard to describe. Notes are smeared, and the tone sounds off and unnatural.

I can accept that this is only my third attempt at using Acourate, and I may not have done the measurements or filters correctly. So let's remove the convolution and see what happens:


4. HQPlayer with NO convolution @ 44.1kHz --> USB --> DAC --> Preamp
Sounds worse than HQPlayer with convolution! Some of the faults of the system are back (poor integration of bass, poor coherence). On top of that, digital noise is STILL there, notes are STILL smeared, and the tone still sounds off.

Since I have an RME sound card, capable of 192kHz PCM output via AES/EBU, let's try this:


5. HQPlayer with Acourate convolved filters @44.1kHz --> AES/EBU --> DAC --> Preamp
Note that this is exactly the same configuration as (3), the only difference is that I am using the RME card and outputting via AES/EBU.

Surprisingly, this configuration drops the volume by about 3dB when compared to USB. I looked carefully at the RME driver and the manual of the DAC and could find no reason why the volume should drop. Anyway, the overall presentation seemed more "polite". Dynamics weren't quite as good - transients didn't have the same attack. And this was after level matching.

Very disappointing considering the cost of the RME card - I would have hoped for better. But, since it is capable of passing through 192kHz PCM, let's try this:


6. HQPlayer with Acourate convolved filters @192kHz --> AES/EBU --> DAC --> Preamp
By this stage we were getting a little tired, but we could hear very little difference between this configuration and (5). We have to go back and listen again.

To round off the night, we went back to listening to the CD player direct. We both agreed that it is much superior to ANY of the computer configurations.


CONCLUSIONS
1. CD Player is superior to the computer in any configuration. Not sure why, or whether this can be addressed by tweaking the computer more (e.g. linear power supply, etc).

2. HQPlayer with Acourate convolved filters sounds better than no convolution. Even my early attempts at generating filters with Acourate results in a better sounding system than without.

3. AES/EBU sounds worse than USB in my system with a 3dB volume drop. I am not sure why, might be a setting in the RME card. I will have to look closer.

4. Upsampling to 192kHz made relatively little difference compared to all the above. Further listening may reveal more differences, but right now I would not put money on passing a blind test. All of the above, I could easily pass a blind test with flying colours.

 

[hvbias]

What USB driver are you using? Hearing clicks when using Windows with no convolution is not normal (that might apply to the hiss as well). My suggestion is try a generic ASIO or WASAPI driver with foobar2k and see how it sounds. foobar2k with either of those drivers should be straight bit perfect output. HQPlayer has a ton of options to configure and IMHO best left until you have satisfactory stock sound through USB.

 

[Keith_W]

USB driver = manufacturer supplied ASIO driver.

Erm, forgive my newbieness but how do you use a generic ASIO/WASAPI driver? If I choose another driver in the drop down box (which is not the manufacturer supplied driver) ... I get no sound.

Good suggestion re: Foobar. I'll download it and try.

 

[hvbias]

USB driver = manufacturer supplied ASIO driver.

Erm, forgive my newbieness but how do you use a generic ASIO/WASAPI driver? If I choose another driver in the drop down box (which is not the manufacturer supplied driver) ... I get no sound.

Good suggestion re: Foobar. I'll download it and try.

I'm not sure how to do it in HQPlayer, it's been a while since I tried it. Like you I didn't like the GUI and at the time didn't have any need for its primary selling point (PCM to DSD or DSD64 to higher speeds) so I dropped it.

In foobar2k you place the various drivers in the program files/foobar2000/components folder. Then you can select which driver to use in the app's output settings.

WASAPI: http://www.foobar2000.org/components/view/foo_out_wasapi
ASIO: http://www.foobar2000.org/components/view/foo_out_asio

There is also ASIO4ALL though I've never tried it. If you still get no sound with any of them, the player may well require only the manufacturer's driver, they might need to offer assistance wrt to the clicks/hiss/poor sound quality.

You mention needing to upgrade the player's firmware, that might be it as well. 48 khz max over USB is quite dated. It should be able to do 24/192 with USB 2.o.

 

[RayDunzl]

There is also ASIO4ALL though I've never tried it.

I have ASIO4All.

I can select my DAC's proprietary USB driver with it (if that was a question above).

For some reason, AcourateDRC wants to see ASIO4All in its input/output path for measurement purposes..

But, I rarely play music from the PC, preferring to spin discs on a cheap 'player' and send the digits via S/PDIF to the DAC.

I don't know if I don't like the sound coming from the PC, but do know I don't care for the convenience of it.

 

[Sal1950]

don't know if I don't like the sound coming from the PC, but do know I don't care for the convenience of it.

At your age you need the exercise.

 

[Keith_W]

Exciting times ahead. I have ordered my 8 channel NADAC, and my new subwoofers will arrive in a few weeks.

I happened to bump into the Australian importer / distributor of Merging in a hifi show a few weeks back. Pretty interesting guy, he founded a speaker company with Neville Thiele (he of Thiele-Small fame).

 

[Purité Audio]

All of your listening comparisons were carried out unsighted I presume?
Keith

 

[Keith_W]

No, sighted. I don't need a double blind test to hear a "thump" go through the speakers or a low level "weeeee-ooooo-weeee-oooo-weeee" just to prove to some random person on the internet that what I hear actually exists. If you think I can't tell the difference between a thump going through the system and not, then you need to check yourself in to some conspiracy theory clinic.