Dissection of Audio/Video Gears Components

[kapvin]

But this concept is what i already told as option 1.

Its not the option 2, because in the miniamp, the chip has an internal I2S to analog conversion before interfacing it with inbuilt class-D.

okay. I did not see that. I had assumed it was option 2. and there was some writeup about this too, that implied that it was more like option 2. thanks for correcting.

Don't you think that USB2.0 is more attractive, i don't know about how much I2S is popular in home audio.

Sure. But I was talking about 2 parts.
1. the pcm to pwm conversion - and onwards to the output stage
2. and then the front end, which could be spdif, usb, aes.

My point was each of the different kind of inputs you'd need to give would eventually need to be converted to a single protocol which would feed the sample rate convertor and then the pcm to PWM conversion (i guess the digital to pwm conversion would happen at a fixed sample rate, right?) typically this protocol (AFAIK) is i2s.

I was not suggesting an i2s input to the actual amp box (i don't think i2s support long cable lengths), but an i2s input internally which would be interfaced (through a switcher) which would have the various inputs usb/aes etc.. converted to i2s)

from a consumer point of view what I was trying to suggest was exactly as you have described .. a USB input / spdif/ possibly aes.

hope this clarifies what i was trying to say.

also in a situation such as this; the volume control would be digital, and where would it be? Post SRC? (i am assuming that you would upsample before converting to PWM).

please do help correct any misinterpretations

 

[kanwar]

Sure. But I was talking about 2 parts.
1. the pcm to pwm conversion - and onwards to the output stage
2. and then the front end, which could be spdif, usb, aes.

My point was each of the different kind of inputs you'd need to give would eventually need to be converted to a single protocol which would feed the sample rate convertor and then the pcm to PWM conversion (i guess the digital to pwm conversion would happen at a fixed sample rate, right?) typically this protocol (AFAIK) is i2s.

I was not suggesting an i2s input to the actual amp box (i don't think i2s support long cable lengths), but an i2s input internally which would be interfaced (through a switcher) which would have the various inputs usb/aes etc.. converted to i2s)
from a consumer point of view what I was trying to suggest was exactly as you have described .. a USB input / spdif/ possibly aes.
hope this clarifies what i was trying to say.

You are correct on saying that one needs to convert different formats to a single protocol which is I2S these days & this is the usual way, it is done everywhere. We do have various chips which performs exactly the same functions under discussion.

But as far the growing popularity of USB2.0 high speed is concerned its getting more famous and most flexible interface for the digital input.

also in a situation such as this; the volume control would be digital, and where would it be? Post SRC? (i am assuming that you would upsample before converting to PWM).
please do help correct any misinterpretations

I don't like going by the route of PCM to PWM conversion and driving the ouput stage directly because of poor performance associated with it[Even that TI chip is not doing this for the same reason.].

Also i don't like the digital volume control, i prefer the pure analog one for obvious reason, as it has zero resolution loss even at low volume setting.

 

[kapvin]

You are correct on saying that one needs to convert different formats to a single protocol which is I2S these days & this is the usual way, it is done everywhere. We do have various chips which performs exactly the same functions under discussion.

But as far the growing popularity of USB2.0 high speed is concerned its getting more famous and most flexible interface for the digital input.



I don't like going by the route of PCM to PWM conversion and driving the ouput stage directly because of poor performance associated with it[Even that TI chip is not doing this for the same reason.].

Also i don't like the digital volume control, i prefer the pure analog one for obvious reason, as it has zero resolution loss even at low volume setting.

see? i told you there should be a seperate thread for this!

resolution loss is something that really interests me. (because of the diy project I am working on--> active -digital crossover speakers) and it is absolutely germane when talking about the redbook standard. you really cannot afford to lose a single of those 16 bits.

but when you upsample I assume that word length is also increased (Say from 16-> 24 or 32 bits) now if it is left justified (0dbfs in 16bit = 0dbfs in 24 bit), then you get 8 free bits of dynamic range to play with.. = 48db.

plus any bits lost will be below the threshold of hearing (unless 0dbfs at digital output is amplified to a system spl of 144dB+, in which case you'll get deaf so soon that you really will not worry about resolution)

so how germane is resolution below hearing threshold?

Edit: this was only in the digital domain. in practice, i understand that 24 bit quantization offer more like 18db to 24 of additional dynamic range. but still the point holds

 

[kanwar]

see? i told you there should be a seperate thread for this!

resolution loss is something that really interests me. (because of the diy project I am working on--> active -digital crossover speakers) and it is absolutely germane when talking about the redbook standard. you really cannot afford to lose a single of those 16 bits.

but when you upsample I assume that word length is also increased (Say from 16-> 24 or 32 bits) now if it is left justified (0dbfs in 16bit = 0dbfs in 24 bit), then you get 8 free bits of dynamic range to play with.. = 48db.

plus any bits lost will be below the threshold of hearing (unless 0dbfs at digital output is amplified to a system spl of 144dB+, in which case you'll get deaf so soon that you really will not worry about resolution)

so how germane is resolution below hearing threshold?

Edit: this was only in the digital domain. in practice, i understand that 24 bit quantization offer more like 18db to 24 of additional dynamic range. but still the point holds

Yes the upsampling in theory does solves the problem with missing bits, but have you compared the same in real world tests.

The max you can do is to have high bit resolution to say 32-bit upsampling and source[16-bit wave file] interfacing will need to have 32-bit ready DAC supporting the internal digital volume control inbuilt.


Try a Test:
Take 2 samples of same format music, pass one from digital volume and another one from analog volume[using same DAC, but one at a time], then attenuate the same by 30dB with digital volume in one and thru analog volume with other and then record them. Now you have 2 recorded files of one with digital attenuation and one with analog attenuation. Now what you have to do is, listen to them by increasing the gain by 30dB or more through analog or digital volume chain but at max position, you will hear a BIG difference.


I have tried this on AKM 32 bit DACs on live real world programs where you have kilowatts of amplification and tons of speakers to show the difference in high profile version.

 

[greenhorn]

Posting the innards of my humble DAC. Bear in mind (in line with my signature) that I got it for a lot less than 100 USD




Source

 

[kanwar]

Posting the innards of my humble DAC. Bear in mind (in line with my signature) that I got it for a lot less than 100 USD
Source

Yours is ultra cheap VFM 16-bit 48khz sampling DAC using PCM2704 series chip from texas which we usually see in mass market USB players from Sinoland pumped into every country, very simple and good for the price.

 

[kanwar]

One more big bucks DAC, called the CAD 1543
Pics Source: 6moons audio reviews: CAD 1543 DAC

Acrylic Chassis[As per designer it helps in reducing chassis field, new unknown invention], Obsolete End of lifetime chip TDA1543 which is no longer in production, Horrible PCB layout with respect to high speed digital circuitry on board.

Designer boasts Parallel connected DAC chips with no reclocking, no upsampling, no jitter reduction because of NON OVER SAMPLING , just pure old is gold thing for 6900 Sterling Pounds.:clapping:

Surely after buying it the audiophile buyer will get a "doctrate" among his circles.

Who said people with "Golden ears " are not rich

The heart of the DAC is available right in our own back yard
TDA1543 Original supply, US $ 1-2 , [PHILIPS] Philips Electronics India Limited, TDA1543 Supplier - SeekIC.com

 

[denom]

Acrylic Chassis[As per designer it helps in reducing chassis field, new unknown invention], Obsolete End of lifetime chip TDA1543 which is no longer in production, Horrible PCB layout with respect to high speed digital circuitry on board.

Designer boasts Parallel connected DAC chips with no reclocking, no upsampling, no jitter reduction because of NON OVER SAMPLING , just pure old is gold thing for 6900 Sterling Pounds.:clapping:

Surely after buying it the audiophile buyer will get a "doctrate" among his circles.

Who said people with "Golden ears " are not rich

The heart of the DAC is available right in our own back yard
TDA1543 Original supply, US $ 1-2 , [PHILIPS] Philips Electronics India Limited, TDA1543 Supplier - SeekIC.com

:sad:

This is getting quite serious. I mean, does 1 have to look under the hood of all these high end gears to understand whats happening or what!
Reminds me of folks taking a mechanic along to see whether the car they are about to purchase is ship shape or not

I sure am learning lots from this thread, hope the others are too!
Apne hindustani kaan ki baje se bade phatke padh sakte hai

 

[reignofchaos]

Retarded price agreed, this should cost no more than a few hundred dollars at max. But Philips TDA1543 with no filters/upsampling/oversampling definitely does sound good - at least to my ears.

Though I don't own one right now, a TDA1543 design doesn't sound bad at all. It is an altogether different thing though that it measures horribly and is nonsense technically.

 

[kanwar]

:sad:

This is getting quite serious. I mean, does 1 have to look under the hood of all these high end gears to understand whats happening or what!
Reminds me of folks taking a mechanic along to see whether the car they are about to purchase is ship shape or not

I sure am learning lots from this thread, hope the others are too!
Apne hindustani kaan ki baje se bade phatke padh sakte hai

Well, its clearly evident now, a person with sane mind must know the actual worth of the equipment by looking what is under the hood.

Retarded price agreed, this should cost no more than a few hundred dollars at max. But Philips TDA1543 with no filters/upsampling/oversampling definitely does sound good - at least to my ears.

Though I don't own one right now, a TDA1543 design doesn't sound bad at all. It is an altogether different thing though that it measures horribly and is nonsense technically.

We are here discussing about the blatant lies propagated by the manufacturers not the sonic chit-chat. I know you can get the NOS dac under 100USD very easily but the point here is that someone is marketing the DAC using same chip for a price which is very insane and non-practical, its clear way of robbing the customers.

 

[ghatgepatil]

Very relevant thread for someone intending to buy real hi-fi! That person is suggested first to open up the top cover of the prospective buy, click a picture, post it here and take opinions of experts here.

Seriously, its an eye-opener! Marketing ploys leading to equipment branding is nothing but creating a bigger hole in our already famished pockets. Thanks Kanwar for your no-nonsensical observations. It is quite obvious that you are 'in it' in quite a depth to give out so. :clapping:

 

[kapvin]

Very relevant thread for someone intending to buy real hi-fi! That person is suggested first to open up the top cover of the prospective buy, click a picture, post it here and take opinions of experts here.

Seriously, its an eye-opener! Marketing ploys leading to equipment branding is nothing but creating a bigger hole in our already famished pockets. Thanks Kanwar for your no-nonsensical observations. It is quite obvious that you are 'in it' in quite a depth to give out so. :clapping:

Well, to now give contrarian point of view. it is also about how well it sounds together. In a DAC there are many components. the Clock, the actual DAC chip, the analog output stages, the Sample rate convertor / upsampling alogrithm. each of them have their own set of trade offs.

Controversial point: you do not need the best of technology to get great results.

to take an example of Bentley and Rolls Royce. the top of the line bentley uses a old gen pushrod engine based on pre - 1980 technology. the transmission is a american "old iron" automatic which has also been around long.

Roll royce used similar engines.

In terms of smoothness, power delivery, ride comfort, overall silence, those old gen rolls cannot be faulted. but it was not the latest tech.

then again, int eh audio domain tubes are hardly latest tech yet many swear by them.

the latest tech would be digital volume controls and direct coupling to pwm amplifiers. which is what I discussed with Kanwar a few posts ago. and as kanwar himself argued, while theoretically, this cutting edge tech should minimise losses, the fact remains that it does not sound as good (yet).

so my point of view is thus: feel free to pop the hood and look inside. but do not use that as a way of evaluating whether it is worth it. sure, if it is previous gen tech, be a bit more rigourous. But finally you need to evaluate how it sounds to You in your system. of course, there are some complete hoodwinks such as the celebrated oppo-in- alumunium casing story. but as long as you do not accept the superiority of any difference you cannot hear, you will never be the "emperor in new clothes"

 

[ghatgepatil]

Well, your point heeded.

But, there exists two kinds of individuals. One has a technical bent of mind and the other is a pure layman, as far as the content related aspect is concerned. The one who is a 'technician' would always try and evaluate an approx. cost of built with a certain amount of premium thrown in too, either due to its established name or its actual performance. IF, the content is not to his liking, no matter what the performance, then it is a few sleepless nights for him .

The above would not hold any water for the other variety, of course. That's what is 'branding' or 'buying a brand', I am afraid.

 

[Thad E Ginathom]

Indeed one does not need the latest technology, one needs the most appropriate technology. It's age is really irrelevant.

Rolls Royce engine technology might have been not the latest, but would anyone have looked at it and said that there were ordinary, cheap parts under a skin being sold for crores? Hardly likely.

Controversial though his ideas are here, NwAvGuy makes some interesting points, and describes some interesting experiences. Two things that I recall from recent reading are...

--- pretty is not always good. He describes PCB layout as being of paramount importance for a DAC, and that does not mean looking pretty.

--- improvements can be counter-intuitive. In testing hundreds of parts and combinations, he found that it was not always the expensive or audiophile-approved components that actually gave the best results.

As a layman, both of those points are counter-intuitive to me. For starters, I like innards to look good. If I could tell one from another, I'd also be comforted by expensive components.

Although it may be true that a low-price product can be good, I don't think it is possible to use NwAvGuy's work as a "proof" (If one accepts the products) of this, because the economic model is absolutely different. At least the intellectual, if not the physical, part of the R&D, a major cost of most products, was free, and the marketing costs must have been absolutely zero. It's an interesting model, and one it would be good to see more of, but simply impossible for the established industry.

 

[kanwar]

IMHO, There are 2 main things to evaluate:

1. Sonics , if you like or not
2. Do the internals worth the money you are spending on it . Here i am not saying the technology if old or new but the implementation + workmanship + ofcourse the component quality [A moron selling you TDA NOS DAC for 6500 Sterling is an ideal case which i have investigated in above posts.]

 

[kapvin]

Indeed one does not need the latest technology, one needs the most appropriate technology. It's age is really irrelevant.

Rolls Royce engine technology might have been not the latest, but would anyone have looked at it and said that there were ordinary, cheap parts under a skin being sold for crores? Hardly likely.

It is important to differentiate between technology and implementation of the technology. (better quality components, better tolerances) which is my basic point.

Controversial though his ideas are here, NwAvGuy makes some interesting points, and describes some interesting experiences. Two things that I recall from recent reading are...

--- pretty is not always good. He describes PCB layout as being of paramount importance for a DAC, and that does not mean looking pretty.

--- improvements can be counter-intuitive. In testing hundreds of parts and combinations, he found that it was not always the expensive or audiophile-approved components that actually gave the best results.

have not read NwAvGuy but this is precisely my point. well executed components gelling together makes for excellence. (Also every choice, whether component or topology has trade-offs)

As a layman, both of those points are counter-intuitive to me. For starters, I like innards to look good. If I could tell one from another, I'd also be comforted by expensive components.

Although it may be true that a low-price product can be good, I don't think it is possible to use NwAvGuy's work as a "proof" (If one accepts the products) of this, because the economic model is absolutely different. At least the intellectual, if not the physical, part of the R&D, a major cost of most products, was free, and the marketing costs must have been absolutely zero. It's an interesting model, and one it would be good to see more of, but simply impossible for the established industry.

my final point was that price and performance, or component exotica and performance are less correlated than we make them out to be. At some level we all realise that. And that is why we audition, or rely on trusted people's reviews or our past experience with a brand to make purchase decisions. (though personally, for expensive purchases, I'd tend to use the latter only for shortlisting.)

 

[denom]

Kapvin,

your analogy to compare the Rolls Royce & Bentley is little incorrect in todays times

1) Both ain't owned by same co today
2) See how far they have come under new ownership, technically & product performance wise.
3) When Bentley was under Rolls Royce Ownership, it was deliberately shortchanged otherwise it would have cannibalized the sales of the Rolls Royce. That is a well known fact.

Coming back to electronics, whose 'dissection' we are carrying out in this thread:
whatever items have been dissected so far have been done to highlight the actual design & components implemented, keeping the selling price in mind & NOT for the way its sounds, which even Kanwar has rightly touched upon in earlier post of another FM. The way it sounds & how you can justify its cost is another topic all together, which am afraid will be a 'HOTLY' debated one

Having an Indian mindset of trying to achieve paramount value perception to the cost of the gear being discussed is the main objective.

Of course your POV & that of the other FM's may vary, am sure. Just letting you & the others know mine, since I started this thread.

 

[greenhorn]

Interesting story about pretty PCB layouts. Steve Jobs wasn't always right you know
Folklore.org: Macintosh Stories: PC Board Esthetics

 

[neutral]

There is no way one can say how a dac will sound by looking under its hood the proof is in the pudding !

One can decipher how much money has gone into the components by taking a parts count. Also one can decipher how good the engineering and design / layout is provided the guy who is looking has many years of dac design under his belt and a proven record.

All this makes no sense under practical conditions. I once remember a DIY amp design which was handed to me by a so called capable designer containing MIL spec parts of exorbitant cost. It sounded like a broken record compared to a naim olive amp I used to own at that point of time. The naim had much lesser parts , spec, prestige . rank whatever. But that thing could make music.

 

[kapvin]

Kapvin,

your analogy to compare the Rolls Royce & Bentley is little incorrect in todays times - I was not talking just about today.. these engines were outdated even before the brands were sold. note what I said about old gen

1) Both ain't owned by same co today - i never said they are
2) See how far they have come under new ownership, technically & product performance wise. - the top of the line bentley of today uses the same old pushrod engine (duly updated)
Bentley Motors Website : Models : Mulsanne : Detailed Specification

3) When Bentley was under Rolls Royce Ownership, it was deliberately shortchanged otherwise it would have cannibalized the sales of the Rolls Royce. That is a well known fact. - no debate - but how is that germane to what i was saying?

Coming back to electronics, whose 'dissection' we are carrying out in this thread:
whatever items have been dissected so far have been done to highlight the actual design & components implemented, keeping the selling price in mind & NOT for the way its sounds, which even Kanwar has rightly touched upon in earlier post of another FM. The way it sounds & how you can justify its cost is another topic all together, which am afraid will be a 'HOTLY' debated one

Having an Indian mindset of trying to achieve paramount value perception to the cost of the gear being discussed is the main objective.

Of course your POV & that of the other FM's may vary, am sure. Just letting you & the others know mine, since I started this thread.

Nothing wrong with the thread. I am certainly not questioning it's existence. I agree with the intention and I love kanwar's explanation. i just gave an alternate view because to draw correlation between technology and sound quality can be a bit inappropriate.

also to make purchases based on the BOM (bill of materials) is as wrong (or right) as making the purchase based on looks.

there was no "between the lines intention"- at least not on this one.