Benchtest TASCAM HD-P2 Linear PCM Recorder

The TASCAM HD-P2 is another fascinating two-channel solid state memory recorder of the new generation of machines utilizing solid state memory for storage unstead of a HDD, tape, or optical disc. This unit came with a 1 GB card installed which provides about 14.5 minutes of recording at the highest sample rate of 192 kHz and 94 minutes at 44.1 kHz.
Author:
Publish date:

ANALOG/ANALOG I/O

HEADPHONE OUTPUT LEVEL

At –0.1 dBFS digital level, just shy of onset of digital clipping, headphone output
High impedance loading2.98V, 11.7 dBu
Both channels loaded with
50 ohm @ 1% THD+N 63.0 mW, 1.77V, 7.2 dBu

INPUT SENSITIVITY

Line level input, for digital out 0 dB FS,
Line input volume full up,
44.1 kHz – 192 kHz Fs 27.4 mV, -29.0 dBu
Mic input for digital Out 0 dB FS, mic pad at 0 dB 4.4 mV, -44.9 dBu

INPUT IMPEDANCE

Line Inputs7.5 kilohm
Mic inputs1.17 kilohm

OUTPUT IMPEDANCE

Line Outputs105 ohm

FREQUENCY RESPONSE

44.1 kHz Fs+0.0, -0.1 dB 20 Hz – 15.8 kHz
-3.0 dB @ < 10 Hz, 22 kHz

96 kHz Fs+0.0, -0.1 dB 20 Hz – 14.1 kHz
-3.0 dB @ < 10 Hz, 40.9 kHz

192 kHz Fs+0.0, -0.1 dB 20 Hz – 14.4 kHz
-3.0 dB @ < 10 Hz, 40 kHz

TOTAL HARMONIC DISTORTION

22 kHz measurement filter
44.1 kHz – 192kHz Fs< 0.1%, 20 Hz – 20 kHz

LINEARITY ERROR

44.1, 96 kHz Fs+/- 1 dB 0 to -95 dBFS
< +10 dB @ -120 dBFS

SIGNAL TO NOISE RATIO

44.1 kHz Fs
Wideband49.2 dB
A-weighted84.7 dB
22 kHz BW96 dB

96 kHz, 192 kHz Fs
Wideband50.6 dB
A-weighted88.5 dB
22 kHz BW96.8 dB

DYNAMIC RANGE

44.1 kHz - 192 kHz99.6 dB

QUANTIZATION NOISE

44.1 kHz – 192 kHz Fs-94 dB FS

CHANNEL SEPARATION

44.1 kHz - 192 kHz Fs

Ch1 > Ch2, Ch2 > Ch 175 dB 20 Hz – 4.0 kHz > 65 dB @ 20.0 kHz

ANALOG/DIGITAL I/O

FREQUENCY RESPONSE

44.1 kHz Fs+0.0, -0.05 dB 20 Hz - 19.7 kHz
-3.0 dB @ < 10 Hz, 22.1 kHz

96 kHz Fs+0.0, -0.05 dB 20 Hz – 17.3 kHz
-3.0 dB @ < 10 Hz, 47.6 kHz

192 kHz Fs+0.0, -0.05 dB 20 Hz – 17 kHz
-3.0 dB @ < 10 Hz, 71 kHz

TOTAL HARMONIC DISTORTION + NOISE

44.1 kHz Fs< 0.009% 20 Hz – 20 kHz
95 kHz Fs< 0.011% 20 Hz – 20 kHz
192 kHz Fs< 0.016% 20 Hz – 20 kHz

LINEARITY ERROR

44.1 kHz, 96 kHz Fs+/- 1.0 dB 0 to -105 dBFS
< +5 dB @ -120 dBFS

SIGNAL TO NOISE RATIO

44.1 kHz Fs, Lch/Rch
Fs/295.1/93.3 dB
A-weighted97.3/96.2 dB

96 kHz Fs, Lch/Rch
Fs/292.5/90.2 dB
A weighted97.4/96.9 dB

192 kHz Fs, Lch/Rch
Fs/280/78 dB
A-weighted97/95.4 dB

DYNAMIC RANGE

44.1 kHz- 192 kHz97.5 dB

QUANTIZATION NOISE

44.1 kHz - 192 kHz Fs-96 dB FS

CHANNEL SEPARATION

44.1 kHz - 192 kHz Fs

Ch 1 > 2, Ch 2 > 1> 75 dB 20 Hz – 4.0 kHz
> 65 dB @ 20 kHz

Note: Most measurements were made at 44.1 kHz, 96 kHz, and 192 kHz and with bit depth of 24 bits unless otherwise noted. The input volume controls were set for the onset of clipping for both channels for a 1V input.
(click thumbnail)

(click thumbnail)

(click thumbnail)

(click thumbnail)

(click thumbnail)

(click thumbnail)

(click thumbnail) The TASCAM HD-P2 is another fascinating two-channel solid state memory recorder of the new generation of machines utilizing solid state memory for storage unstead of a HDD, tape, or optical disc. This unit came with a 1 GB card installed which provides about 14.5 minutes of recording at the highest sample rate of 192 kHz and 94 minutes at 44.1 kHz.

Since the unit has analog and digital I/O, it would be possible to measure the unit in all four of the I/O modes. However, since the majority of use for the machine is likely to be analog I/O and analog in to digital out, these are the two modes that measurements were made in. To further simplify and reduce what would be an unwieldy amount of data, sample rates of 44.1 kHz, 96 kHz and 192 kHz were used along with a constant 24 bit word length. The input level control was set for an output condition just shy of visual clipping for a 1V input level. This corresponded closely to 0 dB FS at the digital output.

Frequency response for the sampling frequencies of 44.1 kHz, 96 kHz and 192 kHz are shown plotted in figure 1 for the analog I/O using the line in/out RCA jacks. As can be seen, there is a severe rolloff at the two higher sampling frequencies with the 192 kHz sample rate response being down a bit more than 20 dB at the usual band limit point of some 96 kHz for that sample rate. Most of this is in the D/A converter as the response for the digital output for the same analog input as shown in figure 2 holds up quite well at the 96 kHz Fs but still smoothly rolls off above 50 kHz at the 102 kHz Fs. One nice consequence of the nature of the HF rolloff at the 192 kHz sample rate is the great reproduction of a 10 kHz square wave. It is almost ring and overshoot free as opposed to the symmetrical ringing that occurs with the usual FIR low pass response. The digital output has just a little symmetrical overshoot. Including frequencies down to 10 Hz shows a small rolloff of 0.25 dB With a 600 ohm load, 10 Hz was down about 3.7 dB.

Total harmonic distortion plus noise in a 22 kHz measuring bandwidth as a function of signal frequency and sampling frequency was virtually identical for all sampling frequencies. A typical result is shown in figure 3 for the analog output. Sampling frequency for the figure is 44.1 kHz. With the measurement bandwidth opened up to 80 kHz, The distortion from 20 Hz to 20 kHz was dominated by noise and measured about 0.15% for 44.1 kHz and about 0.2% for 192 kHz. Distortion vs. frequency and sampling frequency was somewhat better in the A/D mode and is shown in figure 4 for all three sample frequencies and for the left channel.

Here, the distortion at the highest sample rate is quite a bit better than in the A/A mode.

THD+N of a 1 kHz test signal as a function of level down from a few tenths of a dB below full scale is shown plotted in figure 5 for the analog output. Results are essentially the same for all three sampling frequencies. The results for the A/D mode were virtually the same.

Deviation from linearity was essentially the same for all sampling frequencies and is shown in figure 6 for the analog output and a sampling frequency of 44.1 kHz. For the A/D mode, the deviation from linearity was somewhat better holding down to zero to about –115 dBFS and being up about 5 dB at –120 dB Fs.

Channel separation was essentially independent of sampling frequency and direction. A plot of separation for the analog output is graphed in figure 7. The channel separation in the A/D mode is essentially the same.