Aren't they just chips? (1 Viewer)

[Wayde_R]

This forum is great btw! Thanks for all the great dialogue, I've been reading past discussions and a lot of Vince's guides, they're very well written for the layman.

I've always had a bias toward heavy components with real transformers in their power supplies made from genuine minerals from mother earth. I loved my modest separates stereo stuff for years and years. But now I'm starting to enter the digital/multi channel age (little late, yes). I have a natural tendency to believe in the 'old ways' heavy components, wide but short circuit paths unobstructed by extras, effects, processors or whatever. So all this talk of DD and DTS processing as well as DAC all sort of make me sweat, I see it as a necessary evil at best.

My question is, are there quality differences in these 'chips'? Are some dacs/processors etc made to be 'better' hi-fi quality (higher conductive materials greater current capacity etc) than others outside the sum of their firmware or specifications? Or is it possible that a 16/44.1 PCM chip is just a 16/44.1 PCM chip? Aren't all IC's going to be basically the same? Is it possible some manufactures get their processors from the same factories churning out generic chips?

What I'm getting at is my 'real world' application of collecting hi-fi equipment (I laugh when I hear of a $6K component called "pretty-good-for-the-money"). I'm always looking for the 'cheaper' way out. I feel that the playback device might be permitted to be the weakest link. If your DVD player for instance is just using a digital output and the receiver is doing the processing, isn't the only thing the player doing is reading the code off the media and porting a flat (digital) signal out to be processed in a higher quality component? Thus giving it less opportunity to muddy your acoustic waters?

On the same token, if a dolby surround processor is only passing a flat analogue signal through a series of copyrighted processes found on ICs isn't the acoustic damage already being done to your precious audio signal whether it's a $100 add-on processor or a $2K pre/processor? Can you get away with spending less on a cheaper add-on surround processor (without getting carried away and buying complete garbage) if you're only going to use it for movies anyway and it won't be heard for music?

I hope I've expressed myself clearly. I have friends who collect hi-fi that don't feel any urge to make the playback device or processors nearly as high grade as the rest of their components.

Thanks
Wayde

 

[AaronBatiuk]

are there quality differences in these 'chips'? ... Aren't all IC's going to be basically the same?

Yes, their are differences, and no, they are not all the same.

Take two DSP chips for example. A DSP chip will process digital audio to perform things like EQ or tone control, crossover filters, and those infamous love-'em or hate-'em "DSP modes", like Concert Hall or Jazz Club. To do this, the DSP performs a great many mathematical calculations on the audio samples. Cheap DSPs use integer math (no fractional values) which have poor precision and result in added noise in the signal by the time that the DSP is done with it. Better ones use floating point. And all have a certain pre-determined precision measured in bits. The more bits used to store calculation results throughout the process, the more accurate output samples are in the end. This can be 32 bits (bad), 48 bits (poor), 56 bits (good), 64 bits (really good), or as high as 72 bits or more. Because all filters (EQ, tone control, crossovers) are implemented as IIR filters (Infinite Impulse Response), the result of each calculation is used in the next calculation. Each sample effectively goes through thousands of calculations. So even a small amount of error on each calcuation can propagate through many calculations and result in a relatively large amount of noise (error) in the output. The DSP quality definitely has bearing on the quality of the audio.

DACs (Digital to Analogue Convertors) definitely are not all created equal. Notwithstanding the performance of the analogue portion of the circuit, the performance of the digital section itself is subject to many conditions. To start with, all modern DACs oversample the input signal. So a DAC which is converting a digital signal at a 48 kHz sampling rate will not convert it at 48 kHz. It will generally oversample (upsample) to 8 times that rate, or 384 kHz. The reason for this is an effect known as imaging (or aliasing). The analog output of a DAC will have a mirror image of the spectrum of the signal at above 1/2 the sampling frequency. Let me illustrate: a 21 kHz tone encoded in the digital (48 kHz sampling rate) signal will produce an image (duplicate of itself) at 27 kHz at the output of the DAC. (48 kHz/2 is 24 kHz; 21 kHz is 3 kHz below that; image is at 24+3 = 27 kHz). So a filter is necessary to remove all images: an anti-imaging or anti-aliasing filter. Implemented as an analog filter, this would be a bad sounding filter. It would have to be as flat as possible up to 22 kHz (the typical 'passband' for a 48 kHz sampling rate), but filter out everything above 24 kHz (1/2 sampling freq.). Such a steep filter has a very poor impulse response, which really just means that it will sound harsh and nasty. So what modern DACs do is upsample the signal to (e.g.) 384 KHz, then apply a digital filter (with a managable and favourable-sounding impulse reponse) to remove all the aliases above 22 kHz. Then when the DAC produces the analog signal, the first alias appears at 192 + (192-22) = 362 kHz! That is very easily removed by a gentle analog filter that does not sound harsh or nasty. The point to all this is that cheaper DACs oversample to a lower frequency (therefore requiring more aggressive analog filtering), and will implement a poorer performance digital filter in the oversampling part. And since modern receivers will convert an analog input (like a tuner, turntable, tape, TV, etc.) into digital in order to process it, the ADC is just as important as well (and has very similar filtering issues to the DAC).

Virtually all modern audio equipment use op-amps in at least some of their circuit paths. You'll find a few exceptions in the highest of the high end stuff, which may use an all-discrete circuit path (not to be confused with "discrete output stage", etc.). Modern op-amps can perform very well, but the better performance ones cost quite a bit. With 7.1 (8) channels, and with op-amps required in each of several circuit stages, there can be quite a few - I counted 22 dual op-amps listed in the service manual for my receiver. With costs ranging from over $50 each for premium grade ones to less than $.50 for the cheapies, you can see that there can be significant parts cost in this area. Higher priced equipment is far more likely to use higher quality op-amps than cheap equipment.

Note also that even two identical ICs coming off the same fabrication line, even the same batch, will perform differently. Small varations in conditions present during manufacture (defects in the silicon crystal, impurities, etc.) will result in variations in the performance. Some DACs will be more accurate. Some will have less noise. Often, the manufacturers test every part and grade them into different performance grades, with the best performing parts of a specific type commanding a premium price.

 

[Wayde_R]

Thanks. This is a lot of information. I've copied it into a text file so I can refer back to it, it's a lot to digest. But I suspected some might be made better than others, but didn't know they actually have different calculations and sample differently. This is interesting. Thanks for the spending the time to post this.

Wayde