Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Johnny B]

Although there may be some marketing hoopla to dissect and set aside, and there are several potential contenders in this competitive field of chip making, (Burr-Brown, AKM, etc) which ones are claiming the latest, greatest, strides for audio applications, specifically, for pro audio applications?

This should be fun to compare them.

[RedBus]

Don't forget the Wavefront/Alesis AL1101 A/D and AL1201 D/A!

If it's good enough for Behringer (ADA8000), it's got to be Pro-Audio.

RedBus.

[Jim Williams]

For A/D, you have the Crystal CD5381, look it up.
For D/A, it's the BurrBrown PCM/DSD1792. 130 db dynamic range ought to do.

Don't look for pro audio boxes to rush to use these, they seem to be stuck on low cost AKM stuff.

That's why they call it "prosumer".

Jim Williams
Audio Upgrades

[George_]

what's the sense of this thread?

Alesis vs. Behringer shows the results.. They both use the same chip in even the Alesis AI-3 and Behringer ADA8000. The AI-3 sound's more transparent, better linearity over frequency (this is from a magazinetest..) and better lowend..

they both use the same chips.

So it depends on the eng. behinde the chips..

[zetterstroem]

RedBus wrote on Mon, 09 May 2005 19:29

Don't forget the Wavefront/Alesis AL1101 A/D and AL1201 D/A! If it's good enough for Behringer (ADA8000), it's got to be Pro-Audio. RedBus.

                     

you're a funny guy!!

just fell off my chair!

[zetterstroem]

noGearslut wrote on Tue, 21 June 2005 05:08

what's the sense of this thread? Alesis vs. Behringer shows the results.. They both use the same chip in even the Alesis AI-3 and Behringer ADA8000. The AI-3 sound's more transparent, better linearity over frequency (this is from a magazinetest..) and better lowend.. they both use the same chips. So it depends on the eng. behinde the chips..

yeah and alesis even got 16bit resolution out of a 24bit chip!!

amazing!!    

[jfrigo]

Jim Williams wrote on Mon, 20 June 2005 07:32

Don't look for pro audio boxes to rush to use these, they seem to be stuck on low cost AKM stuff.

Nothing wrong with AKM 5394. The chip is just the tip of the iceberg. Put a high-end AKM chip in a well designed box and it can sound great.

[Johnny B]

Instead of making this contest of the OEM's who are only using the chips in their product/boxes and then selling them to the public, why don't we begin more simply and just list the companies that make A-to-D and D-to-A chips?

Once we have all the chip makers listed, we can then discuss the merits or de-merits of individual chips and the on-going R&D efforts and new product roll-outs of the various chip makers.

[Johnny B]

In no particular order:

1. Burr-Brown:

    http://focus.ti.com/analog/docs/analoghomepage.jsp?templateI d=1&familyId=2

2. Analog Devices:

http://www.analog.com/

3. AKM:

http://www.akm.com/index.asp

4. Crystal/Cirrus:

http://www.cirrus.com/en/products/

Please feel free to add others to this list of chip makers. Thanks.

[zetterstroem]

i like this one.... sounds wonderful

http://www.analog.com/en/prod/0,,765_796_AD1865,00.html

[dcollins]

Johnny B wrote on Sun, 26 June 2005 08:49

In no particular order: 1. Burr-Brown:      http://focus.ti.com/analog/docs/analoghomepage.jsp?templateI d=1&familyId=2 2. Analog Devices: http://www.analog.com/ 3. AKM: http://www.akm.com/index.asp 4. Crystal/Cirrus: http://www.cirrus.com/en/products/ Please feel free to add others to this list of chip makers. Thanks.

5. Wolfson

http://www.wolfson.co.uk/

DC

[Johnny B]

Thanks.

So the new list looks like this:

In no particular order:

1. Burr-Brown:

 http://focus.ti.com/analog/docs/analoghomepage.jsp?templateI d=1&familyId=2

2. Analog Devices:

http://www.analog.com/

3. AKM:

http://www.akm.com/index.asp

4. Crystal/Cirrus:

http://www.cirrus.com/en/products/


5. Wolfson

homepage:

http://www.wolfson.co.uk/

digital audio products page:

http://www.wolfsonmicro.com/products/digital_audio/


Please feel free to add others to this list of chip makers. Thanks.

[danlavry]

Jim Williams wrote on Mon, 20 June 2005 15:32

For A/D, you have the Crystal CD5381, look it up. For D/A, it's the BurrBrown PCM/DSD1792. 130 db dynamic range ought to do. Don't look for pro audio boxes to rush to use these, they seem to be stuck on low cost AKM stuff. That's why they call it "prosumer". Jim Williams Audio Upgrades

The AD world is much smaller than the DA world. There are a lot more people listning to music than recording music.

Therefore, some companies that do great job with DA's, and can probably do great on the AD side, choose to make DA's and not AD's.

When it comes to IC's for high quality AD's there are very few contenders. Yes, I agree with you that Crystal (Cirrus) is one of the best makers of AD IC's. But I disagree with your comments regarding AKM. In my view, AKM is also one of the best conversion IC companies.

While I disagree with the marketing departments of all of the IC makers tried to "jump on the wagon" of 192KHz, I do respect their capabilities and contributions.

Regards
Dan Lavry
www.lavryengineering.com

[Johnny B]

Dan,

I think right now we are just seeing who the various chip makers are.

Who knows, maybe there are some chip makers people have forgotten about, some chip makers that not everybody knows about, or, even some new companies.

In this day of Murders and Acquistions, LBO's, Management Buy-Outs, Spin-Offs, Spit-Ups, and Start-Ups, one needs to periodically review the landscape just to see what's out there.

Once we get a little list compiled of the chip makers, people can discuss the merits and de-merits of individual chips in the chip makers' product lines.

I do like your idea of maybe dividing up the two sides between A-to-D and D-to-A at some point. We could even expand it to include the other things like video stuff which may somehow be co-opted and re-deployed by the audio design community.

However, I would like to see this thread start simply by just compiling a list of the chip makers and their links for now.

How does that sound?  
 

 

[zmix]

ouch...

[danlavry]

Johnny B wrote on Mon, 27 June 2005 18:38

Dan, I think right now we are just seeing who the various chip makers are. How does that sound?

I posted my comment specifically regarding AKM for a reason. Someone put a negative comment about them, while in my view they are an excellent company.

Of course you can continue with your list...

Regards
Dan Lavry
www.lavryengineering.com

[Johnny B]

Do you like this chip or not?

Burr Brown PCM3052

[Johnny B]

I mention the Burr-Brown PCM3052 chip since that's the one that comes standard in Apple's wildly popular iMac G5 series.

[danlavry]

Johnny B wrote on Wed, 29 June 2005 16:19

Do you like this chip or not? Burr Brown PCM3052

Before I comment further, I would not judge everything by a chip alone.

Do I like it? The chip is not a dedicated DA. It is not a dedicated AD either. It is a codec - AD with DA and with Mic amp, Bias, MUX and PGA. It seems to be one of those all in one low cost deals that is sold to large computer makers in volume. The "rest of the world" does not get to see data sheets or specs! Go to TI web and see for yourself...

From what I know, the chip does 96KHz (not higher, which is good), has about 101dB SN (but I do not know if it is A weighted), so it is very far from being a candidate for high end conversion. It is a high volume low cost computer oriented all in one audio chip, including a micpre operating on a single 5V supply at 1/4 watt dissipation. It may be good value for the money and for what it is intended for...

Regards
Dan Lavry
www.lavryengineering.com

[Johnny B]

Spoken like a true "engineer type" with lots of diplomacy, qualifications, and mention of cost issues...

Funny, I was not aware that certain chip's specs are not freely available to the public-at-large and to people such as yourself...

Is that something that often occurs?

Do all the chip makers have such more or less "secret" chips?

If so, what does that say about the chip makers' communication methods with the audio engineering and design community, let alone the "ear people" community?  

[danlavry]

Johnny B wrote on Fri, 01 July 2005 05:06

Spoken like a true "engineer type" with lots of diplomacy, qualifications, and mention of cost issues... Funny, I was not aware that certain chip's specs are not freely available to the public-at-large and to people such as yourself... Is that something that often occurs? Do all the chip makers have such more or less "secret" chips? If so, what does that say about the chip makers' communication methods with the audio engineering and design community, let alone the "ear people" community?

Johnny,

I do not view "engineer types" as diplomatic, on the contrary.
I am referring to technical engineers such as EE's (not the recording engineers). We mostly deal with data, hard numbers and hard facts. In my view, engineers tend to be less socially skilled and rather un diplomatic. But that is only is my own personal opinion.
 
But yes, engineer types do speak with qualifications and cost issues are a real thing that needs mentioning.

I can not comment about TI decision not to publishing the specs. It is not the first time it happened. The specific IC is, after all, suitable for computer applications, and it is possible they decided that there is no market outside computers.

An "engineer type" has to qualify and quantify what they say. An "engineer type" looks at a list of issues and specifications, some of which are a trad-off. The IC you mentioned is, for example, a very low power device, certainly given the many functions it has. One expects the low power to be an advantage for say laptops, but low power also calls for trad-offs in performance. Cost is also a trad-off...

I do not think that the computer makers top priority is audio and ear people. And at least so far, I doubt it that any ear people expect much from a mic input on a computer mini jack - probably good enough for some "Internet phone" applications...

My concern is that given the size of the computer makers, if (or when) their marketing decides to hype people into some 384fs or more than 24 bits audio format, you may end up with a large crowed of un informed people ready to eat some BS, which will be followed by manufacturers selling it.

Regards
Dan Lavry
www.lavryengineering.com        

[zmix]

Johnny B wrote on Fri, 01 July 2005 00:06

Spoken like a true "engineer type" with lots of diplomacy, qualifications, and mention of cost issues.

ouch....

[Johnny B]

I was talking about Double E types, not recording engineers. Usu. they are very different animals.  

Translated, I think Dan is saying this BB chip is a mass consumer market chip, not a high audio quality chip even though it does 96k.

I find it interesting that some chip makers do not freely disclose all their data and specs. Even more troubling, it seems there are little or no communication channels existing between the chip makers and either the "audio design engineering community" or the "ear people" community. If true, that should be addressed and improved.  

[Johnny B]

Re: the chip makers and their most recent R&D efforts, budgets, and future product discussions.

Some of these companies may be listed on the Stock Exchange, as a result, they have to file disclosures and reports with the regulators such as the SEC. Often, market analysts and others read thru those reports to gauge the future direction of a company and closely examine things like pending patents, R&D budgets, and diclosures about future products. Some of these listed-company disclosure reports can be viewed on-line by the public for free.

[zmix]

Johnny B wrote on Sat, 02 July 2005 09:29

I was talking about Double E types, not recording engineers. Usu. they are very different animals.

Oh? I happen to be both, - and a musician, too...

[Johnny B]

Cool, maybe you should be one of the people that the chip makers consult with.

I guess we might need some links to the SEC and the USPTO to see what kinds of claims and disclosures the chip makers are filing and get a handle on the future direction they will be taking.

Perhaps, there are some industry segment experts who have met with management, the engineers, and have published what they found out or what they believe will be new and exciting products in our future.

128-bit chips clocking along at 4 zillion Khz!  
(probably an oxy in there, aye?  "zillion Khz?")

In any event, by investigating some of these alternative sources of on-line information, we might get some clues as to what direction the R&D efforts are heading and what the Next Gen chips will be like.

[danickstr]

I like to reflect on what the future might hold in the world of tech, and I would guess that you can be sure that the "next gen" chips will have the qualities of the current set with improvements that lower the cost and raise the s/n so that what is offered can be cheaper and better.  Then it becomes the responsibility of the box maker to craft a legend.  But I think the current set of offerings are quite formidable, and need little in the area of improvement.  I would like to see optical on the Lavry for easier integration into McPT rigs, but that is minor.
I wouldn't rule out "Lavry 2005 sound" and super PCM quality playback in the 2010 MOTU Traveler, available for 895 at musician's friend.  But if you want to work today and make it sound good, better get out a juicy credit card and head to Mercenary, cause the G5 freebee chip won't cut it.  Not that you don't have good gear, and not that it isn't fun to muse about what will be.

cheers

[Johnny B]

I dunno...I see a lot of format changes on the horizon...clearly the CD format was a big mistake (too slow and too few bits among other problems) so the old, obsolete CD format leaves a lot to be desired in terms of its lack of that "high sound quality."

I sense that certain hardware engineers, software engineers (including the micro-coders among the chip designers) as well as some math and theory folks are in large part responsible for the deplorable sound of digital.

Analogue still beats digital all to hell, I wish it were different, but it's not.

Since it's not, that means a lot more work to get digital up to "World Class Analogue Standards."

I know there are those that believe that digital sound is just perfect the way it is, that no improvements can be made...ever...but I happen to disagree with those people. I feel there is a lot of room for improvement in digital sound.

Maybe with the new formats that will be comming and the Next Gen Chips, digital sound will improve but I'm not holding my breath.

What I am doing is trying to keep my eye on what the chip makers are doing with the Next Gen Chips, hence, this thread.

[cyrilmartin]

Hello all,

I've always wanted to respond to that trend but was kept to secrecy. It doesn't matter anymore, so I will give you more information on the PCM3052 (and it's new revision the PCM3052A).

It was a really wrong guess to think that this little beast was a low cost, mass market chip. It was actually designed as a high(er) end product, and had nothing to do in a computer- that's why it was so fun to squeeze it in. I have seen MANY pro-audio gears that didn't sport a DAC as good as the one you are talking about. What makes it mass market (and cheap) was the high volume at which Apple buys it, NOT the quality of the part. For information, the DAC supports 192KHz, while the ADC is limited to 96KHz.

Regarding the simple topic DAC, ADC or codec, it really doesn't matter- all those devices are built using a 0.35uM process, so the die is pretty small. The only time where integrating both ADC and DAC on the same die can be a problem is when asynchronous operation is required beat tones can appear. In the case of the PCM3052, ADC and DAC are synchronous so it really doesn't matter. For the note, there are countless chips on the market that sports both ADC and DAC, but only one section is bounded out, and available to the user. It just outlines that die area is not what you pay: IP is what you pay! It is also very common with single opamps: it is actually a dual opamp but one of the chips is disabled,,,

Analysing the chip, you will notice that the DAC section is of much better facture than the ADC section. The integration of a few features in front of the ADC allowed for a better integration in the Mac, especially the ability to switch between the stereo analog input and the microphone input. The mic input is only an inverting opamp with high gain, then switch, then PGA. The integration of the PGA allow for a good matching of the gain range. The ADC is a single bit, oversampled ADC with pretty good quality FIR downsamplers. It is the classic Burr Brown ADC.

The DAC is an 8 level quantizer with a 4th order modulator. SNR is around 104db(A) and THD is at -93db. Not bad for a 2001 product. Not looking at numbers, it is the sound of Burr Brown converters that made them famous- and it shows. Listening to the PCM3052, the "slam" is really good, and it kept a soft but detailled sound. I defy anyone to find a chip used in a computer that sounds that good. For the note, the same exact DAC structure was used in NAD CD players, with a very similar sound... NAD is usually pretty critical in the choice of their chips.

Of course the DAC is not directly connected to the output jack- a stereo headphone amplifier is used. The selection varies depending on the machine considered, but all of them are on a Bi-Cmos process, with bipolar input stage. Bipolar input stages are much better at maintaining low noise at low power. All of them are also using a "charge pump" to remove the need for coupling capacitors- the only coupling cap is BEFORE the amplifier, where it is small and can be chosen of better quality. Using coupling cap on the output would force the use of high CV capacitors (at least 100uF) that have much more impact on sound...

So there- the PCM3052 will disappear of the mac now that the system is Intel based. And it is a shame- what replaces it is certainly not as good....

For the note, TI made the datasheet of the PCM3052A available on their websites if you want to have a look. Please note that quality of the filters and such- we've seen way worse, and very often!

[Bob Olhsson]

I don't look at chips very often but the last time I did, it sure looked like they were all in a race to develop a lowest common denominator, one size fits all solution at the lowest possible price. Is anybody developing new high-end chips?

[mpd]

OK, I am an engineer with a decent background in hardware design.

If you are an designer, then you know where to get specs from.  Most of the time they are in the datasheets, if not, then you call up your rep and get them.  If you are a good enough customer or seasoned enough to have good connections, then you can call someone inside the company and get the info you need straight from the horse's mouth.

There are "secret" chips, but they are really pre-production samples.  Depending on the stage of chip, the specs can change, or they maybe pretty much set.  However, the availability of these chips is limited to people with NDAs, and they typically aren't available in large enough quantaties for real production.

It is also possible to fab your own chips.

That being said, there is a lot more to analog design that just specs, especially with converters.  And even when you know the specs, you have to be very careful with what you are reading and make sure you understand exactly what is being tested, how the test was performed, and under what conditions.

[ammitsboel]

Any discreet ones?

[mike chafee]

Bob Olhsson wrote on Mon, 15 May 2006 13:31

I don't look at chips very often but the last time I did, it sure looked like they were all in a race to develop a lowest common denominator, one size fits all solution at the lowest possible price. Is anybody developing new high-end chips?

Wolfson

Mike Chafee

[Daniel Asti]

Johnny B wrote on Sat, 02 July 2005 09:29

I find it interesting that some chip makers do not freely disclose all their data and specs. Even more troubling, it seems there are little or no communication channels existing between the chip makers and either the "audio design engineering community" or the "ear people" community. If true, that should be addressed and improved.

They do communicate with the community in regard to their product(s) designed for the community. I'm almost positive if you called them up they would "disclose" its specs. Point being I don't think Mr. Lavry will be designing his latest gizmo around it and they certainly realize this fact.

[Chuck]

Hi Johnny and all,

as for the AD chips, all mentioned by now are of the sigma-delta type.

There is however another type of AD-converter chips progressing rapidly in performance, although mostly used for industry precision application, namely successive approximation ADCs.

The difference is not only the higher price of the latter but also the fact, that they are not averaging, but taking a sample at a discrete point in time, and doing a true 16..18 bit precision measurement on it. No calculations or let's say mathematical interpretations involved of any kind.

Just pure measurement of the analog input right down to DC...

Charles

[Chuck]

... when it comes to ultimate D to A precision COMBINED with great achievable sound-quality there is only one:

the one and only  PCM1704.

Up to 1Mhz of pure 24 bit precision.

Nothing can toucht it.   Nothing.

That's why Dan uses it in his Gold DAC.

Also very nice are (as zetterstroem pointed out) the R2R's from Analog Devices.

Also very nice: our good old 16bit ladder TDA1543, yeah you can get an oh so great sound from that one too.

All the rest is switchers and IME just toys.

You can fool around with them, but its no more than fooling around... Very limited usefulness for accessing the sound quality of a recording.

Charles

[mike derrick]

Johnny B wrote on Tue, 12 July 2005 19:42

I dunno...I see a lot of format changes on the horizon...clearly the CD format was a big mistake (too slow and too few bits among other problems) so the old, obsolete CD format leaves a lot to be desired in terms of its lack of that "high sound quality." I sense that certain hardware engineers, software engineers (including the micro-coders among the chip designers) as well as some math and theory folks are in large part responsible for the deplorable sound of digital. Analogue still beats digital all to hell, I wish it were different, but it's not. Since it's not, that means a lot more work to get digital up to "World Class Analogue Standards." I know there are those that believe that digital sound is just perfect the way it is, that no improvements can be made...ever...but I happen to disagree with those people. I feel there is a lot of room for improvement in digital sound. Maybe with the new formats that will be comming and the Next Gen Chips, digital sound will improve but I'm not holding my breath. What I am doing is trying to keep my eye on what the chip makers are doing with the Next Gen Chips, hence, this thread.

This electronic engineering speak is beyond me, as I am just an end user, but interesting none the less.

It's got me curious though,

Any idea when the next generation of chips are coming and what kinds of improvements we can expect sonically?

And what other areas of A/D and D/A can we expect changes and improvments?

(if this is going too far off topic for this thread, let me know, and I'll start another thread)

[Jon Hodgson]

Did someone just ask Johnny B a technical question?  

Please tell me I'm dreaming!!  

[Jon Hodgson]

mike derrick wrote on Wed, 24 May 2006 09:36

This electronic engineering speak is beyond me, as I am just an end user, but interesting none the less. It's got me curious though, Any idea when the next generation of chips are coming and what kinds of improvements we can expect sonically? And what other areas of A/D and D/A can we expect changes and improvments? (if this is going too far off topic for this thread, let me know, and I'll start another thread)

Hi Mike,

Firstly a disclaimer, because I'm not involved with any converter designs what follows are guesses... I would say educated guesses, but guesses none the less.

Firstly it's probably not really a question of generations of chips in this field. We've already been through three generations of audio ADC converters, starting with Successive Approximation (I guess Chuck must have been asleep for those years), followed by single bit sigma delta (DSD was an offshoot of that), and now multibit sigma-delta seems pretty well established. I haven't heard of anything being about to knock those off... though who knows, things can change.

If we assume that we are going to stay with multi bit sigma delta then the obvious things to improve have already been done. 24 bit converters mean that your quantization steps are below the noise floor of any signal you can actually feed into them, and 96kHz conversion gives the potential to keep all filter effects way out of the audible range. Increasing either is at best a waste of time, at worst detrimental (doesn't stop some people from throwing tantrums and demanding both though).

Also this is a fairly slow moving market compared to things like CPUs and (even more so) Graphics processors, where if you don't have a new device that's way better every few months your competitors bury you.

So we're looking at incremental improvements probably, either getting better quality, or achieving the same quality at lower prices. Now perhaps one way to look at how to improve quality is to look at where your money goes on high end converters.

It isn't the actual converter stage, there may be a few esoteric exceptions out there that implement stuff discretely, but generally speaking they're using the same converter ICs.

It's in
High quality analogue audio stages on the input and output, including handling overs in a "nice" way.
Super stable power supplies
Seperate DSPs to perform the filtering and downconversion.

So you might look at improving your converter IC design by...
1) improving the quality of your analogue stages
2) Moving more of those stages on to the chip itself (especially if you're going for a lower end market and trying to get better quality overall for a given budget)
3) Improving power supply rejection
4) Increasing the onboard DSP power and so implementing better digital filters

[Jon Hodgson]

Oh, and there's clocking too, especially for slaving multiple units to a master clock.

[Chuck]

Jon Hodgson wrote on Wed, 24 May 2006 15:07

Firstly it's probably not really a question of generations of chips in this field. We've already been through three generations of audio ADC converters, starting with Successive Approximation (I guess Chuck must have been asleep for those years), followed by single bit sigma delta (DSD was an offshoot of that), and now multibit sigma-delta seems pretty well established.

That's correct all the way, including me dreaming (of a great sound)

I have experimented quite a bit on this, using  Analog Devices newest PulSAR converters (1 per channel) which I drive up to 192kHz (yeah, I could also twice that speed).

The cool thing: No filter altogether. No analog filter, no digital filter. I can do that as the sampling frequency 192kHz is high enough for not expecting any alias effects.

Whereas modern sigma delta converter just half the decimation rate when outputting 192kHz (which does not improve resolution at all), the SAR machine doubles the resolution finally giving the great sound I was dreaming of.

Charles

[mike derrick]

Chuck

Re: 192 not having any aliasing frequencies.

I was under the impression that 96 was high enough to achieve the same thing.

Why 192 if all it does is add more data and work for your converters and processor?

~ mike

[mike derrick]

Jon Hodgson wrote on Wed, 24 May 2006 08:07

Hi Mike, Firstly it's probably not really a question of generations of chips in this field. We've already been through three generations of audio ADC converters, starting with Successive Approximation (I guess Chuck must have been asleep for those years), followed by single bit sigma delta (DSD was an offshoot of that), and now multibit sigma-delta seems pretty well established. I haven't heard of anything being about to knock those off... though who knows, things can change. If we assume that we are going to stay with multi bit sigma delta then the obvious things to improve have already been done. 24 bit converters mean that your quantization steps are below the noise floor of any signal you can actually feed into them, and 96kHz conversion gives the potential to keep all filter effects way out of the audible range. Increasing either is at best a waste of time, at worst detrimental (doesn't stop some people from throwing tantrums and demanding both though). Also this is a fairly slow moving market compared to things like CPUs and (even more so) Graphics processors, where if you don't have a new device that's way better every few months your competitors bury you. So we're looking at incremental improvements probably, either getting better quality, or achieving the same quality at lower prices. Now perhaps one way to look at how to improve quality is to look at where your money goes on high end converters. It isn't the actual converter stage, there may be a few esoteric exceptions out there that implement stuff discretely, but generally speaking they're using the same converter ICs. It's in High quality analogue audio stages on the input and output, including handling overs in a "nice" way. Super stable power supplies Seperate DSPs to perform the filtering and downconversion. So you might look at improving your converter IC design by... 1) improving the quality of your analogue stages 2) Moving more of those stages on to the chip itself (especially if you're going for a lower end market and trying to get better quality overall for a given budget) 3) Improving power supply rejection 4) Increasing the onboard DSP power and so implementing better digital filters

Thanks Jon,

Hmmmm, if conversion is as good as it gets, then I am very disappointed.  If the chips are as good as they can get, perhaps improvements can be made in the other aspects in the A/D/A chain?

If converter quality is maxed out then Dan Lavry can shut down his R&D department!  But something tells me that we're not there yet.  (Perhaps it's my 1" 8 track talking...)

~ mike

[Jon Hodgson]

mike derrick wrote on Thu, 25 May 2006 12:23

Thanks Jon, Hmmmm, if conversion is as good as it gets, then I am very disappointed.  If the chips are as good as they can get, perhaps improvements can be made in the other aspects in the A/D/A chain? If converter quality is maxed out then Dan Lavry can shut down his R&D department!  But something tells me that we're not there yet.  (Perhaps it's my 1" 8 track talking...) ~ mike

You misunderstand me, I didn't say that conversion is as good as it gets, I said that it seems unlikely we'll be seeing any leaps generated by some radical change in architecture, like going from successive approximation to sigma delta, or specifically in increasing word size or sample rate.

Dan doesn't design converter ICs, he designs converter units, which use ICs within them. My suggestion is that the improvements you are likely to see in converter ICs are likely to be associated with the kinds of things he does to maximize quality... so for example Dan has to build very stable (and expensive) power supplies, therefore a sensible move on the part of a converter IC designer would be to improve power supply rejection, thus allowing the same or better quality with a cheaper power supply.

As for your 8 track, in what way would you say it is superior to a high end ADC/DAC combination... taking into consideration that the aim is accurate reproduction of the incoming sugnal, not subjectively beneficial colouration?

[Chuck]

mike derrick wrote on Thu, 25 May 2006 13:04

Chuck Re: 192 not having any aliasing frequencies. I was under the impression that 96 was high enough to achieve the same thing. ~ mike

Yes, 96 is fast enough to do without filters.

But hey, those sigma-delta converters cannot work without filters, by principle.

My SAR design works without filters.

The fun thing is, it works pretty well on 44.1 without filters.

mike derrick wrote on Thu, 25 May 2006 13:04

Why 192 if all it does is add more data and work for your converters and processor?

IME it is not all it does. The difference is not as brutally bold as going from 44.1 to 88.2 / 96, but there is a difference.

Charles

[Daniel Asti]

mike derrick wrote on Wed, 24 May 2006 04:36

This electronic engineering speak is beyond me, as I am just an end user, but interesting none the less. It's got me curious though, Any idea when the next generation of chips are coming and what kinds of improvements we can expect sonically? And what other areas of A/D and D/A can we expect changes and improvments? (if this is going too far off topic for this thread, let me know, and I'll start another thread)

I have a limited understanding of this also but I think your question is great because that is exactly why I keep coming to check out this thread.

My questions are: Will there ever really be digital audio? Will we ever be able to record and playback in digital? I understand that we can't build a speaker with 0s and 1s.

Does anyone see another format arising in digital beyond the converter chip? Is the DA converter comparable to the stylus and cartrige of a turntable? Something infinetely linked to that medium?

Is the main separtion between pro and hobby AD/DA mostly the analog signal path? Assuming they are using the same converters and clock of course.

I notice an emu interface is using well regarded converters and claiming to achieve what is a very good s/n ratio and other impressive specs. Where are the shortcomings in this peice? How did they achieve those specs? When looking for quality have people been essentially duped by them(emu)? Can something with 120db of dynamic range (AD/DA) actually not be a professional unit.? (I haven't heard this unit and I'm not bashing it. I just want to understand the justification in price and performance to other AD/DA). I know that opamp, pcb and craftmanship would contribute to cost but wouldn't they affect the dynamic range if it was cheap junk?

I own an ad8000se now, which I think sounds good. I'm planning on upgrading the DA with 6 Lavry Blue which I think sounds better. I do notice an upgrade in sound quality on every speaker I own. To me it's a total mystery as to why though, even at 24/48 it is obvious to my ears which I prefer.. But the recording industry is no mystery; having something that sounds even a little bit better or worse is how you win and lose.

[mike chafee]

Ceramic transducer technology was pioneered by New england Audio Research (NEar0 mant years ago.

The main advantage is a high Youngs modulus, (stiffness to weight ratio) and break up modes that occur at a higher frequency than the passband of the transducer.

Revel currently uses organic ceramic in their Performa line, which although exceptional, hasve no profile in the mastering community.

Mike Chafee

[Chuck]

Daniel Asti wrote on Fri, 26 May 2006 00:54

Is the DA converter comparable to the stylus and cartrige of a turntable? Something infinetely linked to that medium?

Hi Daniel,

I think this is a very good analogy. The DAC is the stylus, cartridge, tonearm, phono preamp and when it comes to wow, flutter, rumble (jitter) the DAC also can take over the function of turntable motor, platter, turntable housing, bearing simply everything.

Digital is digital only for storing. In digital we have perfect storing- and copying-capabilities (which we do not have in analog).

In analog, on the other hand, we have 'perfect' processing capabilities (earwise), which we do not have in digital.

This means that there always is a great deal of interpretation involved as every DAC (as well as ADC) sounds different.

There is also a great deal of interpretation involved in any digital processing algorithm (digital filter) which changes the recorded numbers, including sample-rate-conversion.

I read from time to time of people claiming to have a recording / reproduction chain, that is 100% transparent and indistinguishable from their analog bypass.

I think a statement like this can only be founded on poor or unsuited playback equipment and / or limited or unevolved listening awareness as there are no 2 recording / playback chains that can ever sound the same and they never sound exactly like the analog bypass.

So IME digital IS perfect for storing and copying. Everything else is interpretational to a high degree and digital devices sound boldy differently, really comparable to different TTs, cartridges, tonearms, etc...

Btw. here's a fun turntable-link:
http://www.altmann.haan.de/turntable/

Charles

[Daniel Asti]

Chuck wrote on Sun, 28 May 2006 06:35

So IME digital IS perfect for storing and copying. Everything else is interpretational to a high degree and digital devices sound boldy differently, really comparable to different TTs, cartridges, tonearms, etc... Btw. here's a fun turntable-link: http://www.altmann.haan.de/turntable/ Charles

Thanks Charles,

There obviously was a point when other breakthough components hit there maximum potential. Such as transformer based components like a mic pre amp or power amplifier. My opinion is that there really hasn't been that much innovation in these products for some time.

I would be aweful hard to say that you couldn't achieve the same results with the best preamp from the 1970's as you could with the best from today. I'm not sure that from an engineering standpoint, this level hasn't been reached with the ADC/DAC.?

Either I'm correct and we'll find a different medium or system within digital or my children will be recording at 192bit 1536k.?
Improving a converter might be accepted at finding a workable operational frequency and making it more efficient on other levels.