Communication
Contacts
Offering
Investors
Careers
Sesame
Embedded memories
Logic virtual components
Analog virtual components
Test structures
Virtual test & diagnostic
Hardware/Software Codesign
Hardware/Software Codesign
Layout verification
Quadrant of skills
SoC Integration
Custom Fabless Supplier
 
 

Search dolphin:

silicium estimator

Benchmark

 

Considering the importance of a truthful competitive evaluation, diverse benchmarking techniques had to be invented. We are here at the meeting point of two contradictory cultures: PCB-based product design and SoC Integration. The former has traditionally been reluctant to perform thorough simulations, while the latter has thrived with the practice of pin-to-pin simulation. The former has promoted experimenting and twiddling, while the latter has developed Virtual test and diagnostics enabling high-yield designs right-on-first-pass…

And the challenge is about Mixed Signal SoC performance, versus separating pure logic IC's and Codecs! Whatever the choice by the equipment integrator, due to the specific emphasis which prospective users wish to apply, the relevance of the benchmark for their key performance, namely SNR, must be pointed out.

Until now, product-oriented developers tended to resorted bizarrely to evaluators meaningful at the Specification stage, but not applicable through Test, which explains a number of misunderstandings when prospective users had expected similar figures from existing products referring to irrelevant or too approximate benchmarks and had been deceived by their comparisons.

There is a hidden law in business: traditional providers do not how to address benchmarking for innovative products… but we stick to the DOLPHIN tradition of striving to promote fair "Master Benchmarks"!

To begin with, thanks to the World Wide Web, a constructive proposal for the JAZZ product line is offered as our explicitly recommended twin Master Benchmarks: the rough and the smart! It can be seen as an invitation to our prospects and customers to shed a new look with a reasonable but thorough yardstick for the preferred characteristics of their profession.

This Benchmark could be replaced and thus shall be replaced by any other that shall be deemed more relevant, therefore all prospective users should not hesitate channeling their best suggestions to us, lest they would nor reap the benefit from their own know-how.

Have a look at our other Product Line specifics!

As each has its own style for setting its benchmarks…

 

 

Audio CODEC Benchmark

 

audio codec benchmark

 

 

Benchmark formulas

Resolution benchmarks accounting for:

  • Noise on time scale (Clock jitter)
  • Noise on amplitude (power supply)
  • Instrumentation Noise
  • External components characteristics

Power-consumption benchmarks accounting for:

  • Input signal characteristics
  • Device configurations (gain settings, activaed features)
  • Load conditions

 

 

SNR Definitions

SNR : practical definition
Ratio of output level with 1-kHz full-scale input, to the output level with the input short circuited, measured A-weighted over a 20-Hz to 20-kHz bandwidth using an audio analyzer.
SNR : purist definition
Ratio of fundamental frequency output level with 1-kHz full-scale input, to the noise level measured A-weighted over a 20-Hz to 20-kHz bandwidth using an audio analyzer.
Potentially hide noise floor increase with signal level (ΣΔ modulator overload)
Efficient for industrial testing: smaller test time, loosen signal generator requirement
Accounts for all noise contributions with full scale signal ΣΔ operation
Accurate for device characterization

 

 

Audio CODEC Benchmark

Parameters
Benchmark for (SNR,THD)
Benchmark for Power
Rough
Smart
Playback only
Record only
VDDA (analog supply)
Typ
Noise-less
Min
w/ bumble bee
w/ noise
typ
typ
Input Signal
1 kHz sine
Noise-less
1 kHz sine
Noise-less
1 kHz sine, Full scale
Or AES2 standard (pink noise CF=6 dB)
same
CLK
44.1 kHz (Fs)
Jitter-free
44.1 kHz (Fs)
Jittered
44.1 kHz (Fs)
16 kHz (Fs)
External Components
Ideal
Real
-
-
Device operation / configuration
FS (SNR)
FS –3 dB (THD)
FS (SNR)
x mW output power (THD)
DAC path only
Typical gain settings
ADC only
Measurement Bandwidth
A-weighted
20 Hz-20 kHz
A-weighted 20 Hz-20 kHz
Or wide band 100 kHz
-
-
Audio load
10 kOhm Resistive load
Low Ohmic [8/16/32] Ohm
Low ohmic speaker
(load power is excluded)
-