Philips LHH1000 CD player Measurements
Sidebar 2: Measurements
Looking first at the analog stages, the output impedance of the unbalanced sockets was pretty much to specification at 57 ohms, while 0dB at 1kHz corresponded to an RMS voltage of 2.018V. Frequency response was flat throughout the audio band, with a faint hint of lift (0.1dB) at 20kHz and a lower roll-off point of 2dB at 4Hz. De-emphasis was within ±0.05dB from 125Hz to 8kHz, but reached a maximum error of 0.26dB at 16kHz. Crosstalk was below my measurement threshold at all frequencies.
The output stage was non-inverting; ie, absolute-polarity correct, while fig.1 shows the LHH1000's response to a 1kHz squarewave. The symmetrical ringing on the wave tops reveals the linear-phase characteristic of the digital filter and analog post filter. Moving on to the LHH1000's deviation from perfect linearity, using the dithered test tones used from the CBS CD-1 disc, this measured 0dB at 70.31dB (both channels), 0dB (left) and 0.8dB (right) at 80.77dB, and 1.2dB (left) and 2.2dB (right) at 90.31dB.
Fig.1 Philips LHHH1000, 1khz squarewave at 0dBFS.
Fig.2 shows the 1/3-octave-analyzed spectrum of the player's left output while playing the dithered 90.31dB track. Power-supply components can be discerned at 60Hz (111dB) and 120Hz (109dB), while the 1kHz tone is reasonably clean, with an absolute level of 91.5dB. Apart from the rising level of noise in the top two audio octaves, the main harmonics present are the third, at approximately 101dB, and fourth and fifth at 102dB. The monotonicity step waveform from the CBS test disc for the right channel of the LHH1000 is shown in fig.3. This can be seen to have evenly spaced steps implying good differential linearity, though the very lowest level is asymmetrical about the time axis, revealing a very-lowlevel DC shift as well as a level error.
Fig.2 Philips LHHH1000, 1/3-octave spectrum with noise and spuriae of dithered 1kHz tone at 90dBFS.
Fig.3 Philips LHHH1000, monotocity waveform.
The actual shape of the undithered 1kHz tone at 80.77dB can be seen in fig.4. This appears to have evenly spaced bit levels, as would be expected from fig.3. At 90.31dB, however, the undithered tone (fig.5) appears as a squarewave with intense switching noise apparent at the points where in fact the signal should linger on digital zero. For comparison, fig.6 shows the same waveform when reproduced by Sony's expensive DAS-R1 decoder, which also uses the TDA1541A dual-DAC chip, but this time doubled-up, each DAC running at a normal 4x rate but handling alternate samples so that the whole can run at an 8x rate. (The DAC cannot reliably be run at more than 176kHz due to its settling time requirement.) Here, despite a spike appearing every time the input word changes from 1111111111111 to 0000000000000000 and from 0000000000000000 to 0000000000000001, the actual waveshape is exquisitely formed.
Fig.4 Philips LHHH1000, undithered 1kHz tone at 80.77dBFS.
Fig.5 Philips LHHH1000, undithered 1kHz tone at 90.31dBFS.
Fig.6 Sony DAS-R1, undithered 1kHz tone at 80.77dBFS.
The use of record dither implies data on the disc below the apparent 96dB dynamic-range floor of a 16-bit linear system. The dithered tracks on the CBS disc therefore allow the reviewer to see whether, indeed, a player can resolve this information. Looking at the waveform of the dithered 90dB 1kHz tone, when 32 samples triggered at the same point in the waveform are averaged, the noise level is considerably reduced and the result (fig.7) can be seen to be a reasonable if still noisy sinewave rather than the stepped squarewave in fig.6. For comparison, fig.8 is the same dithered 90.31dB tone when reproduced by the Sony DAS-R1, again time-averaged over 32 samples to reduce the noise level.
Fig.7 Philips LHHH1000, dithered 1kHz tone at 90dBFS.
Fig.8 Sony DAS-R1 dithered 1kHz tone at 90dBFS.
The fade to silence of the 500Hz tone on the CBS disc revealed the Philips to produce a relatively pure-sounding tone that faded cleanly away. Yes, HF noise could be heard (as well as a faint amount of mains hum well below the 60dB starting point of the tone), but this sounded very even in nature, if "white" rather than "pink," and didn't change its nature as the tone dropped away to inaudibility.
Using Stereophile's standard test disc for determining a player's error-correction ability (Pierre Verany PV.788031&2), it appeared that this Philips player was not quite in the top league when it came to handling severe data abuse. It had no problem coping with gaps in the data up to 0.75mm in length, but above that an increasing number of glitches were apparent. When the dropout reached 3mm and above, the player kept "losing its place," and ultimately stopped the disc. Things were much the same when the missing data were combined with minimum track pitch, while two dropouts following in quick successiontwo gaps in the data each more than 0.5mm longproduced glitches and mutes. This performance is actually quite good. It just isn't in the same class as the very best machines.John Atkinson