Sampling rates and street sense. or when is close, close enough?

[Gregadd]

 

[The Smokester]

...As an aside, people keep talking about 96 KHZ/192 Khz as the better data rates but in reality, 88.4 Khz and 176.4 Khz are more optimal. Anyone know why?

This is just a guess, but 88.4 and 176.4 are multiples of 44.1--the redbook sampling rate--so I would think that it might make resampling between formats easier. ?

 

[amirm]

This is just a guess, but 88.4 and 176.4 are multiples of 44.1--the redbook sampling rate--so I would think that it might make resampling between formats easier. ?

You got it. Hopefully anyone targeting CD is using multiples of its sampling rate and therefore, when releasing the high resolution versions, would put out that. If they are creating at 96 and targeting the CD, then that is less optimal. And if they are creating at 88 and releasing at 96, that is really non-sense.

 

[JackD201]

It is true. Just using the pro data rate of 48Khz would have given us nearly perfect scenario. If we are going to change it, then anything north of 50 to 55 Khz will do the job. Building good digital systems would have been cheaper and easier if they had just given us a bit more headroom.

As an aside, people keep talking about 96 KHZ/192 Khz as the better data rates but in reality, 88.4 Khz and 176.4 Khz are more optimal. Anyone know why?

Less math to down convert to 44.1?

 

[JackD201]

Dang! Smoked by the Smokester!!!!

 

[amirm]

Don't worry. You can both split the prize.

 

[RBFC]

Wasn't the Colossus system, first popularized with Telarc's recording of Alexander Nevsky, using a 48 kHz sampling rate? There were a few other discs released on the Bainbridge label that used the system, and those offered very good sound. One had George Cleve conducting the Midsummer Mozart Festival Orchestra, and another was a cello/piano duet disc of Rachmaninov (IIRC..). I have all the mentioned discs, and will dig them out to confirm the repertoire.

Perhaps the Colossus system, and others like it, didn't catch on just because of using 48 kHz instead of redbook 44.1. Many movie soundtracks, however, are 24 bit/ 48 kHz sampling rate. Interesting how the two industries have not meshed in the use of their technologies.

Great contributions, Amir. Thank you.

Lee

 

[audioguy]

If I am converting analog to digital (e.g. my turntable through the TacT preamp), isn't (in theory) 24/192 better than, say, 24/48, since I am not creating or reading a CD?

 

[RBFC]

If I am converting analog to digital (e.g. my turntable through the TacT preamp), isn't (in theory) 24/192 better than, say, 24/48, since I am not creating or reading a CD?

I believe that the issue was the distortion level being higher for DACs running at 192kHz than at 96kHz or lower.

Lee

 

[RUR]

If I am converting analog to digital (e.g. my turntable through the TacT preamp), isn't (in theory) 24/192 better than, say, 24/48, since I am not creating or reading a CD?

Chuck, the TacT DSP's process in and convert everything to 96, so you may find that leaviing the output @ 96 provides the best AQ.

 

[Nicholas Bedworth]

Regarding the, uh, English in the Lavry white papers and documents, they definitely weren't edited by a fine hand. Dan is (as I recall) Russian, via Israel, and is a very smart (and engaging) fellow. So's his wife Priscilla. They're both music enthusiasts to the max. However, Dan's written English has its moments. He definitely presents things in a straightforward and graphical manner, which is quite helpful to the beginner, and useful to the more advanced reader.

And he prefers digitally-controlled analog attenuators to the all-software designs of, for example, Berkeley and Weiss (Weiss also makes an all-analog passive attenuator). Either approach can give very good results, although my personal bias is to do it all in software. Others think differently.

In addition to knowing his Bechsteins from his Boesendorfers, Daniel's favorite instrument is the accordian. Why? Vell, with accordian, one can play real Russian folk music!

Regarding higher sampling rates, etc., remember that Nature sometimes fights back. Sure, well-recorded and mastered 88/24 sounds terrific, to my ear: it's really free from "effort" and "sluggishness". On the other hand, certain artifacts and distortions hidden by the lower SNR of 44 can emerge. It's sort of like the difference between 480P and 1080P: the lower resolution format hid a multitude of sins. Under the microscope of 1080P, all is revealed.

 

[Nicholas Bedworth]

DICE frequency conversion chipset

Yeah, the frequency conversion deal is a real nightmare (speaking of jitter, anyone?)... The DICE chipset in the Weiss INT 202 can do this quite well... and the users manual is 250+ pages. One doesn't get into arbitrary frequency conversion without a lot of processing power.

Look at those jitter measurements. This is accomplished by a hybrid PLL that includes both software and hardware components. Attached is an excellent dive through the DICE product.

 

[Nicholas Bedworth]

@ Audioguy... going to 192K in this scenario guarantees your files will be 2X the 96K and 4X the 48K That's about it. As RUR noted, the TacT is 96K anyway, so that's a good balance of performance and bulk.

 

[audioguy]

Chuck, the TacT DSP's process in and convert everything to 96, so you may find that leaviing the output @ 96 provides the best AQ.

I forgot about that. Thanks.

 

[JackD201]

Hey Peter, you lost me there. What do you mean?

 

[Lee]

Regarding the, uh, English in the Lavry white papers and documents, they definitely weren't edited by a fine hand. Dan is (as I recall) Russian, via Israel, and is a very smart (and engaging) fellow. So's his wife Priscilla. They're both music enthusiasts to the max. However, Dan's written English has its moments. He definitely presents things in a straightforward and graphical manner, which is quite helpful to the beginner, and useful to the more advanced reader.

And he prefers digitally-controlled analog attenuators to the all-software designs of, for example, Berkeley and Weiss (Weiss also makes an all-analog passive attenuator). Either approach can give very good results, although my personal bias is to do it all in software. Others think differently.

In addition to knowing his Bechsteins from his Boesendorfers, Daniel's favorite instrument is the accordian. Why? Vell, with accordian, one can play real Russian folk music!

Regarding higher sampling rates, etc., remember that Nature sometimes fights back. Sure, well-recorded and mastered 88/24 sounds terrific, to my ear: it's really free from "effort" and "sluggishness". On the other hand, certain artifacts and distortions hidden by the lower SNR of 44 can emerge. It's sort of like the difference between 480P and 1080P: the lower resolution format hid a multitude of sins. Under the microscope of 1080P, all is revealed.

I am not an engineer like the Lavrys but I can report on some recording experience and studio observations. When I started doing recordings here in Atlanta we were doing 24/88.2 and 24/96 recordings and we got pretty good results. When we starte doing recordings at 24/176 we finally were able to recreate the "event" as nice as I wanted to. Now sampling theory would suggest 16/44.1 is fine enough. But in practical recording implementation it is not. It has for sure gotten a lot better but 24/176 just seems more natural - smoother, more detail, better imaging, better timbre.

 

[Bruce B]

Hey Peter, you lost me there. What do you mean?

plugins and DSP have a native rate that they process everything in. Most of the time it's in 24/96 or whatever. So instead of making your machine work harder and doing a lot of up/down converting, keep it at one sample rate throughout.

 

[Orb]

Yeah, the frequency conversion deal is a real nightmare (speaking of jitter, anyone?)... The DICE chipset in the Weiss INT 202 can do this quite well... and the users manual is 250+ pages. One doesn't get into arbitrary frequency conversion without a lot of processing power.

Look at those jitter measurements. This is accomplished by a hybrid PLL that includes both software and hardware components. Attached is an excellent dive through the DICE product.

Doh somehow missed this. thanks for the link.
Also shows that they have implemented support for the firewire/IEEE 1394 isolation in the Link Layer Controller, which is part of the standard (missing from USB and this means proprietary/bespoke engineered solutions as seen by Ayre QB9).

Cheers
Orb