Close in jitter?

[RayDunzl]

But, I am not clear on the other measurement.

The "clean" measurement is the output of the RCA jacks on Zone 2 of the preamp. The XLR of Zone 2 goes to the JBL speakers.

Zone 1 of the preamp feeds the Krells and MartinLogans.

The gear chain:

Clean trace:

PC ->REW Generator -> Optical Out at PC -> 25 foot Optical Cable -> Optical/Coax Switch -> Behringer DEQ2496 -> MiniDSP OpenDRC-DI with FIR filter active -> Benchmark DAC2 HGC -> Krell KTC preamp -> Zone 2 RCA 25 foot cable -> PC on-board sound ADC -> REW analysis

Dirty traces (speakers):

PC ->REW Generator -> Optical Out at PC -> 25 foot Optical Cable -> Optical/Coax Switch -> Behringer DEQ2496 -> MiniDSP OpenDRC-DI with FIR filter active -> Benchmark DAC2 HGC -> Krell KTC preamp -> Zone 1 XLR/CAST -> Krell amp -> MartinLogan

PC ->REW Generator -> Optical Out at PC -> 25 foot Optical Cable -> Optical/Coax Switch -> Behringer DEQ2496 -> MiniDSP OpenDRC-DI with FIR filter active -> Benchmark DAC2 HGC -> Krell KTC preamp -> Zone 2 XLR -> JBL LSR 308 (active speaker)

 

[DonH56]

The sidebands are ~60 dB down, about 0.1 % distortion. They appear for both speakers but not the preamp loop. I'm betting on the microphone's mic capsule, ADC, and/or buffer amp for the spurs and noise floor.

 

[Blumlein 88]

Your preamp measure looks maybe too clean. As if the ADC and DAC have locked clocks. The speakers too dirty. I'll see if I can do some measures like that later.

 

[RayDunzl]

Your preamp measure looks maybe too clean.

Maybe...



There's plenty of noise on the 25 foot RCA cable (analog preamp out to PC) to obscure really quiet stuff...

No Signal:



As for the preamp, with a MSRP of $8500 in 2003 or thereabouts (and not having much bling), I'd expect it to be at least competent. (Specs on last page).

(Sorry, can't get these images into thumbnails)

I bought it used on eBay. The flourescent display is a bit dim, and the tiniest power interruptions shut it down, other than that, it works.

As if the ADC and DAC have locked clocks.

No mechanism here for that.

Also, the Behringer (maybe) and the miniDSP and the DAC each have ASRC happening/available... the miniDSP turns everything into 48kHz, and the DAC2 does conversion to 211kHz.

miniDSP - https://www.minidsp.com/images/documents/Product Brief-OpenDRC-DI.pdf

"All digital inputs on the DAC2 are asynchronously upsampled to 211 kHz to remove all traces of interface jitter. Benchmark's UltraLock2™ upsampling system provides uniform jitter-free performance with all types of digital interfaces (coaxial, XLR, Optical, and USB). Cable length, cable type, and interface format will not cause an increase in jitter at the D/A chip in the DAC2. External jitter attenuation devices will have no positive or negative effect on the clock performance of the DAC2."

 

[RayDunzl]

The speakers too dirty.

I don't know... I do know the UMIK-1 is not an ultra-precision instrument, and that's all I have to work with for in-room measures.

I'll see if I can do some measures like that later.

I like that.

 

[amirm]

The measurement MIC I have has tons of distortion itself:

Other people with the same mic (UMM6) reported exactly the same distortion products.

 

[Blumlein 88]

Maybe...

View attachment 6939

There's plenty of noise on the 25 foot RCA cable (analog preamp out to PC) to obscure really quiet stuff...

No Signal:

View attachment 6940 View attachment 6941

As for the preamp, with a MSRP of $8500 in 2003 or thereabouts (and not having much bling), I'd expect it to be at least competent. (Specs on last page).

(Sorry, can't get these images into thumbnails)

I bought it used on eBay. The flourescent display is a bit dim, and the tiniest power interruptions shut it down, other than that, it works.





No mechanism here for that.

Also, the Behringer (maybe) and the miniDSP and the DAC each have ASRC happening/available... the miniDSP turns everything into 48kHz, and the DAC2 does conversion to 211kHz.

miniDSP - https://www.minidsp.com/images/documents/Product Brief-OpenDRC-DI.pdf

"All digital inputs on the DAC2 are asynchronously upsampled to 211 kHz to remove all traces of interface jitter. Benchmark's UltraLock2™ upsampling system provides uniform jitter-free performance with all types of digital interfaces (coaxial, XLR, Optical, and USB). Cable length, cable type, and interface format will not cause an increase in jitter at the D/A chip in the DAC2. External jitter attenuation devices will have no positive or negative effect on the clock performance of the DAC2."

I didn't realize you were using a Benchmark. Poor reading comprehension on my part. Yes they are able to put out super low jitter results like that. Jitter so low it is practically the same as having locked clocks. So no mystery there. We do have the remaining mystery via the speakers however.

 

[RayDunzl]

We do have the remaining mystery via the speakers however.

I'm thinking jitter in the USB powered USB Microphone ADC at this point... OMG!

I'll try a lower frequency for grins.

 

[RayDunzl]

1200 Hz:

I'm not sure what I'm looking for having reduced the frequency of the test tone, but I don't see much to see here.

+/- 1.8 Hz spreading at -60dB

120 Hz tone: 1.3Hz spread at -60dB or so

Not much for sidetones...

So, I might be back to "no idea".

 

[Blumlein 88]

Alright I repeated your measures with REW and a 12 khz tone. I have a very similar looking graph to show. I did look at a 1 khz tone which is pretty clean. I was measuring from 12 feet away. Repeated measurements using Audacity for playing the tone and Audacity recording the Umik1. Same looking result. This using the Revel F12 right speaker only. The DAC is what is built into an older H-K AVR receiver. Like 10 years old.

Here is Umik1 with 1 khz tone.

 

[Cosmik]

Hmm. Your PC is driving a DAC generating the test tone. This might be isochronous or asynchronous..? Either way, it is either the PC or the DAC that defines the sample rate.

At the same time, the PC is acquiring data from your microphone. According to the mic's blurb:

Forget about driver installation, OS compatibility and un-calibrated mics. The Umik-1 is a USB Audio class 1 device automatically recognized by all Operating Systems

Has the mic got its own sample rate, too? And it is generating packets of data with an embedded clock? Commonly a DAC receiving isochronous USB packets would incorporate a PLL to smooth out the jitter in this. When a PC acquires USB data from a mic, can it do this?

And what does it mean if the PC generates a tone at 48.0001 kHz and the mic acquires it at 47.99999 kHz? It is not obvious to me how this is resolved except through some sort of resampling or timing adjustment i.e. the appearance of JITTER.

For my setup, I ensured I used an analogue mic, being picked up from the sound card's own microphone input, so I knew the sample rates for playback and record were synchronised.

 

[RayDunzl]

Well, the common factor is the UMIK-1 at this point.

Has the mic got its own sample rate, too?

It's 48/16 I believe...

And what does it mean if the PC generates a tone at 48.0001 kHz and the mic acquires it at 47.99999 kHz?

In that case the measured frequency of the tone would be off a hair. I wouldn't think you can connect sample rates when there is analog in the middle.

For my setup, I ensured I used an analogue mic, being picked up from the sound card's own microphone input, so I knew the sample rates for playback and record were synchronised.

I don't have that luxury, here, at the moment. Maybe I have a CheezMic someplace that I can try. I have an analog measurement mic (cheap Behringer whatever) but no way to capture it with the PC.

 

[Cosmik]

In that case the measured frequency of the tone would be off a hair. I wouldn't think you can connect sample rates when there is analog in the middle.

But if you are using REW to generate the test tone and acquire the samples from the mic simultaneously, how does it work? REW in its typical usage is trying to correlate what it generates with what it picks up from the mic. If it has generated N samples, and receives N+37 in return, what does it do? It could drop/duplicate samples = appearance of jitter, or it could apply some sort of resampling algorithm = possible appearance of jitter depending on the algorithm.

Unless you're already doing it, could you use a separate application (say Audacity) to record the received tone and then analyse that?

Maybe I am on the wrong track, but isochronous USB from the mic would be my primary suspicion.

 

[fas42]

This is how audiophilia works...

People espouse science and objectivity, and acknowledge the existence of the 'placebo effect'. But in parallel they clearly believe that if audiophiles love vinyl, 'there must be something in it'. And so a science-based forum still finds itself discussing vinyl year after year...

If people really believed in science, vinyl would not warrant any more attention than 78s.

The 'something in it' for vinyl is that up to fairly recently there were aspects of the perceived sound that LP replay got right far more easily than digital. It's trivially easy to get measurements that "prove" digital is superior to conventional analogue - but digital has always struggled to shake off disturbing anomalies in the playback which made it unpleasant, boring or fatiguing in the long term - LP typically doesn't suffer from these issues. And because the science world finds it difficult to measure what's going on here the solution has been to pretend it doesn't exist.

The situation has definitely improved over the last few years - good digital can now tick all the boxes, and more people are discovering that they don't need to resort to vinyl to get long term pleasure in listening. But, the case won't really "be closed" until full understanding of the situation is achieved.

 

[Blumlein 88]

But if you are using REW to generate the test tone and acquire the samples from the mic simultaneously, how does it work? REW in its typical usage is trying to correlate what it generates with what it picks up from the mic. If it has generated N samples, and receives N+37 in return, what does it do? It could drop/duplicate samples = appearance of jitter, or it could apply some sort of resampling algorithm = possible appearance of jitter depending on the algorithm.

Unless you're already doing it, could you use a separate application (say Audacity) to record the received tone and then analyse that?

Maybe I am on the wrong track, but isochronous USB from the mic would be my primary suspicion.

I don't happen to know which type of USB connection the Umik is using other than 48/16. In my case, I sent over an asynch USB SPDIF converter the digital signal to the coax input of an AVR. The Umik is being used for recording the result and is running on its own clock and over its own USB connection. So the Umik is a microphone that feeds an analog signal to the built in ADC which sends that to be recorded digitally. So the Umik is not looking for any particular number of samples. It does a delay with an acoustic pulse to align timing to maybe the millisecond level. In this use however the DAC clock and ADC clock are not directly related. Such clocks differ by some few dozen ppm in clock rate anyway.

I have other mics and interfaces that are USB ASIO devices I can use for this here later on this evening. I'll post the results.

 

[RayDunzl]

[

Unless you're already doing it, could you use a separate application (say Audacity) to record the received tone and then analyse that?

Audacity sending to system and measuring with UMIK-1 in REW.



Audacity playing 12kHz, UMIK-1 recording via Audacity, and spectrum display in Audacity:

 

[Cosmik]

[

Audacity sending to system and measuring with UMIK-1 in REW.

View attachment 6950

Audacity playing 12kHz, UMIK-1 recording via Audacity, and spectrum display in Audacity:

View attachment 6951

I was thinking of one application generating the tone, and a separate one receiving it. Or a different DAC, or even separate PCs as a comparison - will the 'jitter' appear different? Do we know how an application with its own recording sample rate handles packets of data from the mic that are at a different sample rate, or being received with an element of real time jitter?

As I say, it is not obvious to me, at least.

 

[RayDunzl]

Sync between software on PC:

Top: Audacity sending 12kHz tone at 48kHz and receiving.
Middle: REW sending 12kHz at 48kHz and Audactiy receiving at 48kHz.
Bottom: 12001Hz tone (as example of what would be seen with poor clock lock)

I was thinking of one application generating the tone, and a separate one receiving it.

Yes, did that above:

Audacity sending to system and measuring with UMIK-1 in REW.

Or a different DAC

DAC isn't showing a problem when it is measured electrically.

 

[amirm]

It's 48/16 I believe...

You can check that in Preferences for REW:

 

[Blumlein 88]

Okay, I used REW to send the 11025 hz tone out via Vlink USB to SPDIF converter to the AVR DAC. I acquired the speaker signal via an Avantone CK1 omni mic for which I have a cal curve. It fed a Focusrite Forte which was the input to REW. So separate clocks for playback and record. The result is much nicer, and not terribly far off of what I know other good DACs and the Forte to be. So the Umik is the problem for this sort of measurement. I have not calibrated the level of this mic which is why it seems low. The FR is good, but it actually was about 20 db louder than what is shown. Still around 85 db between signal and noise floor.

The three spikes each side of the main signal would come and go over a few seconds. Either some noise or a side effect of clocks not in synch.