Print

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

[Arf! Mastering]

Comparing rafts of various DA converters one thing most can agree on is that they all sound different, yet the specs are all but identical.  For example, the LE4496 (Blue) DAC manual lists a flatness response of +/- .05dB (10Hz - 22kHz) and 2 degrees phase linearity.  The Lavry Engineering DA924 (Gold) DA manual lists exactly the same values, yet they are very different sounding devices - I'm sure Dan himself will attest to that.  Distortion for the LE4496 is .002% (1kHz @ -1dBFs) while distortion for the DA924 is .0009% ( virtually = .001) but is the ear really that sensitive to a difference of .001% vs .002%?  The same can be said for any flat-response ultra-low distortion amplifier.   What is it that is not being specified the gives same-spec devices such obviously different texture, character, and tonal balance?

[Rob Darling]

No test can simulate dynamic performance.  All response curves become moot when amplifiers have to begin behaving dynamically.  Overshoot and undershoot of different frequencies under varying conditions are going to give the characteristic sound of any given piece of equipment.

[Bob Olhsson]

Power supply, board layouts, grounding, RFI... and that's not to mention the gear that happens to be feeding it and the load it happens to be feeding.

Different implementations are subject to different interface issues. A digi 888 sounds amazingly decent feeding the 2 megohm balanced input of a Mark Levinson preamp and amazingly bad feeding most consoles.

I'm getting a headache!

[danlavry]

AlanS wrote on Sun, 09 July 2006 15:55

Comparing rafts of various DA converters one thing most can agree on is that they all sound different, yet the specs are all but identical.  For example, the LE4496 (Blue) DAC manual lists a flatness response of +/- .05dB (10Hz - 22kHz) and 2 degrees phase linearity.  The Lavry Engineering DA924 (Gold) DA manual lists exactly the same values, yet they are very different sounding devices - I'm sure Dan himself will attest to that.  Distortion for the LE4496 is .002% (1kHz @ -1dBFs) while distortion for the DA924 is .0009% ( virtually = .001) but is the ear really that sensitive to a difference of .001% vs .002%?  The same can be said for any flat-response ultra-low distortion amplifier.   What is it that is not being specified the gives same-spec devices such obviously different texture, character, and tonal balance?

Before I even start to say anything about the vast area of testing, lets think about the very basics: Take a simple 1KHz tone (sine wave). Say the reading is 1% THD+N. What does it mean? mostly 1% noise? Mostly 1% distortions? If so, is the distortion energy concentrated at the hearing sensitive 3KHz region? Is it concentrated at the not so hearing sensitive 20KHz? Is it spread around the audio range? Does it change with signal amplitude?

When you buy a car, you look at the specs, and it tells you a dozen or so "numbers" such as engine size, torque at some RPM, wheel base... Does it really tell you much about the performance, the quality, the comfort, the endurance....???

Regards
Dan Lavry

[Jim Williams]

It's the analog.

Since most makers are using the same converter chips, the same if not similar VCXO and phase lock loop, it's
the analog and psu offering the most sonic
differences.

Running a great converter chip through a 5532 op-
amp will reduce the sonics to the level of the analog.

[Jim Williams]

AlanS wrote on Sun, 09 July 2006 07:55

Comparing rafts of various DA converters one thing most can agree on is that they all sound different, yet the specs are all but identical.  For example, the LE4496 (Blue) DAC manual lists a flatness response of +/- .05dB (10Hz - 22kHz) and 2 degrees phase linearity.  The Lavry Engineering DA924 (Gold) DA manual lists exactly the same values, yet they are very different sounding devices - I'm sure Dan himself will attest to that.  Distortion for the LE4496 is .002% (1kHz @ -1dBFs) while distortion for the DA924 is .0009% ( virtually = .001) but is the ear really that sensitive to a difference of .001% vs .002%?  The same can be said for any flat-response ultra-low distortion amplifier.   What is it that is not being specified the gives same-spec devices such obviously different texture, character, and tonal balance?

Some converter chips offer much better performance than what you're quoting. Compare a AKM5394A to the BurrBrown PCM1792, big difference in measurement, noise, dynamic range and sound.
Then again, the BB parts are more expensive, but not that much more when figured into the cost of production. Following the dac with a Bessel linear phase filter is also important for dynamics and to control ringing and overshoot. Many converters use a butterworth filter designs with steep slopes which adds harmonics to the square waves, not too musical.

For me, it's the PCM1704 or the 1792 with Bessel filters set to 65k hz followed with class A biased TI THS opamps with no coupling caps. Add a great psu and it's real good.

[danlavry]

Jim Williams wrote on Thu, 20 July 2006 17:28

AlanS wrote on Sun, 09 July 2006 07:55 Comparing rafts of various DA converters one thing most can agree on is that they all sound different, yet the specs are all but identical.  For example, the LE4496 (Blue) DAC manual lists a flatness response of +/- .05dB (10Hz - 22kHz) and 2 degrees phase linearity.  The Lavry Engineering DA924 (Gold) DA manual lists exactly the same values, yet they are very different sounding devices - I'm sure Dan himself will attest to that.  Distortion for the LE4496 is .002% (1kHz @ -1dBFs) while distortion for the DA924 is .0009% ( virtually = .001) but is the ear really that sensitive to a difference of .001% vs .002%?  The same can be said for any flat-response ultra-low distortion amplifier.   What is it that is not being specified the gives same-spec devices such obviously different texture, character, and tonal balance? Some converter chips offer much better performance than what you're quoting. Compare a AKM5394A to the BurrBrown PCM1792, big difference in measurement, noise, dynamic range and sound. Then again, the BB parts are more expensive, but not that much more when figured into the cost of production. Following the dac with a Bessel linear phase filter is also important for dynamics and to control ringing and overshoot. Many converters use a butterworth filter designs with steep slopes which adds harmonics to the square waves, not too musical. For me, it's the PCM1704 or the 1792 with Bessel filters set to 65k hz followed with class A biased TI THS opamps with no coupling caps. Add a great psu and it's real good.

Hi,

You are comparing an AKM5394A which is an ANALOG TO DIGITAL CONVERTER integrated circuit, to a PCM1704 (and PCM1792) which is a DIGITAL TO ANALOG CONVERTER IC...

Regards
Dan Lavry