manisandher wrote:

And I can indeed replicate this effect... but only with my RME FF800 and not with my Weiss AFI1. With the FF800 (slaved to the Pacific), changing the Pacific's wordclock does indeed change the speed of any audio/video being played. (Going further, using the FF800's internal Direct Digital Synthesizer (DDS) clock, I have 'total' control of the speed of audio/video.) This suggests to me that software can indeed be slaved to a firewire device. However, I cannot replicate this effect with the Weiss AFI1. Here, the player/recorder software seems to dictate the frequency. If the frequency down the chain (with the AFI1 or Pacific as Master) is different to that of the software, then there is simply no sound produced.

Interesting. If so then I would say that Weiss AFI1 defeat the purpose of running the Pacific as Master Clock. What would be even more interesting to see if the sonic benefits of having Pacific as Master will manifest itself while Weiss screw up the software slaving. It is possible that despise the AFI1 does not do slaving-transparency but it still would sound very reasonable. Why knows…

manisandher wrote:

For now, I have tried to hear a difference in the sound by simply unplugging the second AES/EBU cable whilst listening to a 44.1/88.2 file in single-wire mode. I'm not sure how closely this replicates step #10 of your procedure, but I cannot hear a difference (in either level or sound quality) with the second cable in or out... Have I totally misunderstood the issue here?

Actually you did not misunderstand the issue the issue but it turned out that I was an idiot. As soon I got your email I decided to do it again. This time I very much as you do not hear any difference. I did it very many times and there was no difference. I absolutely remember as I heard a huge worsening of sound as soon I plug the second digital line in. Upon the further experiments I learned that this problem manifest itself ONLY if I do not mute this second digital line on my Lynx mixer. I was under impression that switching the Lynx and Pacific in single-wire mode shall make them doing not even feel the second line, not it is not how it works. Anyhow, explicitly muting the second line via the Lynx mixer addresses all problems. I thought I did it but I did not realize that it was a very critical step. I wish the Pacific would be smarter with not-used line rejection and I wish the Lynx mixer lines would be more isolated.

Anyhow, some I figure out how to use it the Pacific is rehabilitated, in fact it become a relatively comfy multi-format processor. I just need to find for it a location what I would be sit higher in my rack as it is troublesome to  bend my body to read it display and to push the button with my big toe finger.

Whoever follows the Pacific Microsonics  threads, or whoever use Pacific know that both Models one and two have an interesting idiosyncrasy: at maximum sampling rate, when Pacific runs in dual-wire mode, the clock runs at half speed. It does not make since to have A/D running at 176K but clock at 88K. So, it would be responsible to ask: what is the purpose to slave DAW if the processor does not care the right sampling reference?

I was under this presumption for a couple year but last might I suddenly come to me and I concluders that it does not matter. Let me to explain my rational. Be advised that I have no specific knowledge in the field or consulted with knowledgeable people, this is just my take that is based strictly upon common sense, at least the way how I understand the common sense.

There are two arguments that I would like to bring up:

1)  Pacific Microsonics was designed and built by very knowledgeable people who very much knew what they did. Furthermore Pacific was made in “no expense spare mode”, so it is not such a thing the some functionality was not implemented because they wanted to keep it cheaper. So, if Pacific teal decided to output a half-speed clock at in double-wire mode then they feel that it was how it needs to be done.

2)  What is double-wire mode? It is what one wire cares a half sampling rate and another wire cares another half. The double-wire is not stereo and not dual mono.  In stereo we have two different signals that fluctuate by own rules cared in two wire. In dual-wire AES/EBU each wire cares a half of the very same signal and a half of sampling frequency. In dual-wire digital both wires are running from the very same clock and if the signal in dual-wire divided not identically (I do not know if it is possible) then the fluctuation clock marks across the wires are absolutely identically as there is just one clock for both wires. So, what is the clock slaving? It is some king PLL driven phase-locking of one clock to another clock.  The key in here is to get reference and to trace the rate of change. Since the master clock is only one for both wires then the slave clock might look for the sampling marks in any of the wires ion case it reads samples from AES/EBU. In case the clock reference delivered to slave clock by own 75-ohm BNC word clock cable this cable can care ANY sampling frequency as long phase-wise it reflects the masterclock.

I think this is why Pacific Microsonics  Model 2 does not output 176K clock and Model 1 does not output 88K clock. In fact I do not think that Model 2 even has 176K clock and Model 1 has 88K clock. They most likely have 44K and 48K (1X) clocks and multipliers. So, it is very possible that my concern about the half/clocking in dual-wire mode with Pacific was just my own self-inflicted worry.

Romy the Cat wrote:

Whoever follows the Pacific Microsonics  threads, or whoever use Pacific know that both Models one and two have an interesting idiosyncrasy: at maximum sampling rate, when Pacific runs in dual-wire mode, the clock runs at half speed. It does not make since to have A/D running at 176K but clock at 88K. So, it would be responsible to ask: what is the purpose to slave DAW if the processor does not care the right sampling reference? I was under this presumption for a couple year but last might I suddenly come to me and I concluders that it does not matter. Let me to explain my rational. Be advised that I have no specific knowledge in the field or consulted with knowledgeable people, this is just my take that is based strictly upon common sense, at least the way how I understand the common sense. There are two arguments that I would like to bring up: 1)  Pacific Microsonics was designed and built by very knowledgeable people who very much knew what they did. Furthermore Pacific was made in “no expense spare mode”, so it is not such a thing the some functionality was not implemented because they wanted to keep it cheaper. So, if Pacific teal decided to output a half-speed clock at in double-wire mode then they feel that it was how it needs to be done. 2)  What is double-wire mode? It is what one wire cares a half sampling rate and another wire cares another half. The double-wire is not stereo and not dual mono.  In stereo we have two different signals that fluctuate by own rules cared in two wire. In dual-wire AES/EBU each wire cares a half of the very same signal and a half of sampling frequency. In dual-wire digital both wires are running from the very same clock and if the signal in dual-wire divided not identically (I do not know if it is possible) then the fluctuation clock marks across the wires are absolutely identically as there is just one clock for both wires. So, what is the clock slaving? It is some king PLL driven phase-locking of one clock to another clock.  The key in here is to get reference and to trace the rate of change. Since the master clock is only one for both wires then the slave clock might look for the sampling marks in any of the wires ion case it reads samples from AES/EBU. In case the clock reference delivered to slave clock by own 75-ohm BNC word clock cable this cable can care ANY sampling frequency as long phase-wise it reflects the masterclock. I think this is why Pacific Microsonics  Model 2 does not output 176K clock and Model 1 does not output 88K clock. In fact I do not think that Model 2 even has 176K clock and Model 1 has 88K clock. They most likely have 44K and 48K (1X) clocks and multipliers. So, it is very possible that my concern about the half/clocking in dual-wire mode with Pacific was just my own self-inflicted worry.

Romy the Cat wrote:

Lynx has internal dabbler that activated as soon dual-wire inputs activated. So, when Pacific outputs in dual-wire configuration a half of the sampling speed then is done in accordance to the dual-wire specification.

Romy the Cat wrote:

It is possible that it does it, you might not know how to activate it or it defeated on your unit by some kind of hardware switch. Or perhaps it does it but you just do not know it.

Romy the Cat wrote:

Anyhow, I am sure Daniel would help you.

manisandher wrote:

The last I heard from him, he was looking into how to get the clock frequency doubled, maybe by an external device (I think).

The size, I can understand.

But can't you simply take the tape out from your Placette and plug it into the Pacific's analog inputs? I'm assuming that the Placette has, in addition to its variable output, a fixed output...

Surely this would help with the Pacific's low input impedance also, no?

Mani.

manisandher wrote:

But can't you simply take the tape out from your Placette and plug it into the Pacific's analog inputs? I'm assuming that the Placette has, in addition to its variable output, a fixed output...

Romy the Cat wrote:

Yes, I have thought about it. My Placette has no tape loop but I think it needs to be added. It is shame that I did not think about it before – it would save me some cable not to mention an ability to run A/D from anything… I kind of mildly contemplating what would be better: a full tape loop or a full active tape buffer, but it is not the subject of this thread.

OK, I decided to do it and my beloved Placette is packed up and will be riding tomorrow back to the birth place for an installation of most likely a tape loop, I am still contemplating if a tape buffer will be better. The idea will be to have Pacific permanently bout to the Placette’s tape loop (tape buffer) and then I will be able to record any of my souse with max quietly and with no cable switching of any kind. Even the installation of the tape loop is expensive I anticipate economic feasibly from the project and it will allow freeing up an extra pair of those mean PAD cables….

I just wish Pacific had more friendly access to A/D input sensitivity attenuator. It would make my tape loop vs. tape buffer debate much simpler…

Mani, I was pointed to your thread at about Pacific:

http://www.computeraudiophile.com/content/Pacific-Microsonics-Model-Two

I do not want to join the thread and extend credit to the computer audiophile site. The idiot that runs that site is industry ass liking whore – I truly hate that type. I also do not truly know why you need to play those games.  You have the unit, you like what it does or not – why do you care about anybody else recognition of endorsement?  I truly do not understand it.

A few words from my experience. Figuratively speaking 50% of the Pacific’s quality is A/D converter, 35% is rate converter, the rest 15% is DAC.  The D/A section of this processor is juts for control and monitor what is A/D converter does. It is not bad but it is not what Pacific was made for.

Anyhow, I hope you find a lot of compassion from the computer audiophile Morons.

Romy the Cat wrote:

It would be interesting if Pacific would have an option to slave it to the source’s clock not only vice-versa.

The Weiss AFI1 interface is measuring the following clock rates (in KHz) from the Model Two:

- 44.122
- 88.244
- 176.488

AND

- 48.000
- 96.000
- 192.000

Does this imply that the 48x clock is more accurate than the 44.1x clock in the Model Two? I'm assuming that the 44.1x clock is the same as that in the Model One, so perhaps an 'older' and less accurate design than the 48x clock.

How important is a 0.05% inaccuracy in the 44.1x clock rate anyway? Audible? Anything to be worried about, e.g. is this a sign that the clock at the end of its life?

Mani.

manisandher wrote:

The Weiss AFI1 interface is measuring the following clock rates (in KHz) from the Model Two: - 44.122 - 88.244 - 176.488 AND - 48.000 - 96.000 - 192.000 Does this imply that the 48x clock is more accurate than the 44.1x clock in the Model Two? I'm assuming that the 44.1x clock is the same as that in the Model One, so perhaps an 'older' and less accurate design than the 48x clock. How important is a 0.05% inaccuracy in the 44.1x clock rate anyway? Audible? Anything to be worried about, e.g. is this a sign that the clock at the end of its life?

Mani, I do not understand the argument that older clock has less accurate design. What do you mean the “older clock”? The model one has the same 44x and 48x clocks, so there is no older and newer clocks I guess. I presume that in Pacific as in any other processors there are just two base clocks and multipliers. So, one of them is a bit off the calibration. Presumably that it is not a mis-calibration Weiss AFI1 interface the question would be what difference it makes?

If you record and play on the same Pacific then it would make no different of any kind. If you use file from another A/D processors then… how do you know that they were calibrated to the exact sapling rate? I am sure none of them were. So, the question would be: how auditable it is?

If you record at one sapling rate and play at slightly other then you have the absolutely the same effect as tape played at wrong speed. You need 44.100K but you have instead 22Hz away. This will change the tube of your A440 note. Different orchestras, for different music tune the A note to different center frequency, from 420Hz to 475Hz, sometime even more. A few Hertz here and there are auditable but how much the 22Hz of sampling offset (from 44.1K!!!) would relate to a whole note? I did not do circulations but I think it would be in vicinity of 1/131072 (2 power 8) of tone. You might hear 1/64 of note, you might hear 1/128, you might distinct even more but I do not think that you will be able to acknowledge anything more refund. I think anything behind 1/64 of full note is way burden behind the other influences to tine in audio and I do think that might have any practical meaning to look for absolute reference. At least I would not worry about it and would search for other thing to worry.

Romy the Cat wrote:

The model one has the same 44x and 48x clocks...

Romy the Cat wrote:

At least I would not worry about it and would search for other thing to worry.

Romy the Cat wrote:

I did not do circulations but I think it would be in vicinity of 1/131072 (2 power 8) of tone.

manisandher wrote:

Ah, OK. I was of the belief that the Model One could only operate at 44.1 or 88.2.

manisandher wrote:

Oh yeah, there are plenty of other things to worry about... provided the clock isn't drifting and accelerating away (if indeed it started out at exactly 44.100Khz in the first place).

manisandher wrote:

Yes, it's pretty safe to bet that I wouldn't be able to hear this. Although the other day, my wife walked in wearing a platinum necklace, and I could have sworn that I heard an improvement in the sound. But it doesn't work for anything cheaper than gold.