9z-4-1, buffer sizes and clock resolution

[boleary]

So last night I had an opportunity for a nice 4 hour listening session and was blown away by the latest change in the sound of my NOS1. I am now in week four of the 24/7 break-in period. After about a half hour I started to hear a bit of glare I hadn't noticed before. The glare didn't suddenly start; rather it was there from the beginning of the session, but I didn't immediately recognize it cause of the other positive changes I was hearing. After reading the tool tip on the clock resolution button and remembering what my "best" buffer setting was with my OS DAC, I changed the buffer size from 512 to 48 and the clock resolution from "Max for this System" to "Min for this System." The glare disappeared and a more natural, and I think as Mani would say, "woody" sound emerged. The difference is a bit similar to the difference in sound that came with changing the SFS in 9z-2.

Anyone else notice this?

[PeterSt]

I didn't try this combination, but maybe I should.
So, Device Buffer at 48 (I suppose you set that in the Soundcard Driver settings as well !!?), but with how many samples of latency ?

Thanks,
Peter


PS: This seems rather low to me, and I merely have the idea that you didn't change the Soundcard Driver setting at all, and without knowing only forced Q1 to a lower latency by means of setting the maximum to 48 in the Settings that causing Q1 to max out at 48 ???

[PeterSt]

Ok, I did what I expect you did : Just change the Device Buffer Size in the Settings (nothing else), which makes the latency 48 samples (not 32 as I suggested); this while leaving the soundcard to be at 512.
Set the Clock Resolution to mimimum as well.

This is 3 minutes in now, and although I'm not listening to the same tracks, I'd say the very first thing I notice is more hall (as a positive).

But maybe you did differently afterall ?

Peter


PS: Listening to Muddy Waters - Muddy Mississipy Waters Live. This piano sounded rather dry ... not so anymore ... at all.

[manisandher]

Well I've just set the NOS1 to 48 samples, XX device buffer to 48 and Q1 to 16 samples (Q1=3)... and I'm getting a very, very nice sound indeed.

BrianG was over to my place yesterday to listen to the Magtech amp. We played around with a few things, including comparing my new AC mains vs. the PurePower. I really, really like the sound of my AC mains, but my main criticism has been a lack of 'sparkle'. The above settings bring the sparkle back... in spades.

I will keep the settings for a while and see if I can live with them.

Thanks for the tip boleary!

Mani.

[boleary]

Ok, I did what I expect you did : Just change the Device Buffer Size in the Settings (nothing else), which makes the latency 48 samples (not 32 as I suggested); this while leaving the soundcard to be at 512.

Yes, that is exactly what I did. I forgot about resetting the sound card buffer, but it sounds like that's what Mani did. I have a bit of time right now so I'll give that a try too.

[PeterSt]

Thanks for the tip boleary!

LOL!

Well, after Muddy Waters was finished (for the three remaining tracks) and I put on another album, after a first track of that I set it back. But, I deliberately used that album, knowing it was a more "difficult" one. Then, 3 tracks further again, I reverted to the 48/512 situation from Boleary (as it now indeed turns out). 2 minutes into that track my wife said "hey, you changed something eh ?!" and "much better !".
And next we were talking about how nice / convenient / helpful it is for us and all of us now others using the NOS1, so not being alone anymore on THAT.

Never mind users dialing in settings not intended at all, which in this case is a combination I would not have chosen myself easily !

Highs seem more clear, bass is more defined, and indeed all seems to be more natural.
Fun.

Thanks,
Peter

[boleary]

So glad to be helpful. 

[PeterSt]

Maybe a small hint, at least for NOS1 users ...

After dialing in Boleary's accidental crazy combination - and listening to it since (the first post in this topic), it came to me that all sounded maybe a bit "tight". Maybe short. Also I recognized the bass to be better before, but that would be a tough call.
What triggered me to change it yesterday, was the observation of these 10 days or so that the Clock Resolution of 0.5 seems to hog the system in "circumstances". Not sure, but I didn't have it before. So I changed it to 5ms to see whether it will bring some news.

Well, it did right away, but in an unexpected fashion. BASSSSS.
I never changed my signature as fast as I did now ...

Key values :
- KS Special;
- Device Buffer at 512 (as in the Device itself);
- Q1=48 (samples);
- Clock Resolution = 5ms.
Of course this is with "NOS1" 24/384 (8x Arc Prediction), but this may not be related much.

Peter

[boleary]

Key values :
- KS Special;
- Device Buffer at 512 (as in the Device itself);
- Q1=48 (samples);
- Clock Resolution = 5ms.

Sorry for being slow, but what is your actual XX setting for "Device Buffer Size" with this set up so that your Q1=48? Am assuming that your "Device Buffer at 512 (as in the Device itself)" means that the latency setting in the Juli@ card is 512?

[Gerard]

Key values :
- KS Special;
- Device Buffer at 512 (as in the Device itself);
- Q1=48 (samples);
- Clock Resolution = 5ms.

Sorry for being slow, but what is your actual XX setting for "Device Buffer Size" with this set up so that your Q1=48? Am assuming that your "Device Buffer at 512 (as in the Device itself)" means that the latency setting in the Juli@ card is 512?

Yes that i do not understand too. What is the setting in the settingstab (XX) That could not be 512. Since there is a line written underneath Device Buffer Size that says that both Juli@ (Soundcard) and DBS should be the same.


 


 

[PeterSt]

Ehh ..

- Device Buffer at 512 (as in the Device itself);

This should have said

- Device Buffer Size at 512 (as in the Device itself);

Is it now more clear ?
So both in XX as in the "soundcard" 512.
Next Q1 to 48.

[Gerard]

Ehh ..

- Device Buffer at 512 (as in the Device itself);

This should have said

- Device Buffer Size at 512 (as in the Device itself);

Is it now more clear ?
So both in XX as in the "soundcard" 512.
Next Q1 to 48.

Wel not really.

There is no 48 than in my Q1 only 32 and 64.

[PeterSt]

I am totally wrong !!
Oh boy, there's so few logic in this that I can't even remember it, but I *am* using it. So again :

Key values :
- KS Special;
- Device Buffer Size (in XX) at 48
- Buffer Size in the Device at 512
- Q1=48 (samples);
- Clock Resolution = 5ms.

Apologizes ...

[Gerard]

Apologizes ...

That's ok.... 

But a what about the settings in the Juli@ than? Did you try to set it on 48?

 

[boleary]

Am pretty certain that "Buffer Size in the Device at 512" means the Juli@ latency setting. So the only thing Peter changed from what was first described in this thread is setting the clock res to 5ms.

I tried setting the Juli@ to 48 but got weird popping noises, but that might be due to my not so up to date system. I think Mani changed his latency setting in the Juli@ to 48 with good results. See his post above. On the other hand, I recall Peter saying that setting the latency too low will cause problems.