ProAudio Review - August 2008 - (Page 13)

numerous live music and speech settings: loud rock clubs, loud arena shows, acoustic and amplified music in outdoor amphitheaters, full band rehearsals and writing sessions, spoken word, acoustic and amplified instruments in the studio, in-person interviews, and environmental sounds. I recorded at all settings between 16bit/44.1kHz and 24-bit/96kHz. The LS-10’s LCD displayed everything I needed to know: all current numbers; times — remaining and program; resolution; level; mode; recording media; folder and file name; battery level; Euphony and Reverb settings used, if any, and so on. In most cases, I used the LS-10’s built-in microphones, which I found to sound accurate and quite nice — especially the latter, considering the unit’s street price. In most cases — except when recording environmental sounds and one interview with a particularly soft-spoken individual — Mic Sensitivity stayed on Low, or, “standard level” according to Olympus. Rec Level offered a wide range of gain; I was always able to get an appropriate recording level. I auditioned all Reverb settings; they were OLYMPUS continues on page 15 ® On The Bench: Olympus LS-10 High Definition Audio Master Recorder Fig. 1 Frequency response at 44.1, 96.0 kHz Fs Fig. 2 THD+N as a function of frequency at 44.1 kHz Fs and measurement bandwidth. Red = 22 kHz, Magenta = 80 kHz Fig. 3 THD+N as a function of frequency at 90.0 kHz Fs and measurement bandwidth. Red = 22 kHz, Magenta = 80 kHz The Olympus LS-10 is one of the newest entries in the ever-increasing field of small solid-state memory recorders. This one having built-in stereo microphones, line and external mic inputs, and a USB connection for offloading recorded files to a computer. There is no digital output for directly playing files into an external device. Analog input/output measurements were made by inserting signals at the line or external mic inputs and taking signal output at the headphone output jack. Additionally, recorded files were transferred to my measurement computer for subsequent playback. Playback of the transferred files was done with the Foobar2000 audio player application and sent to my Audio Precision System 2722 from the S/PDIF output of a M-Audio Audiophile 24/96 sound card in the measurement computer. As a test of this playback setup, playing back Audio Precision made wave files of 24 bit data indicated near 24 bit resolution with distortion Plus noise magnitudes of about –139 dBFS Frequency response for the sampling frequen- cies of 44.1 & 96.0 kHz using the line inputs is shown plotted in Figure 1. The frequency range is down to 10 kHz and is linear to best show the high frequency rolloff shape. Including frequencies down to 10 Hz reveals a low frequency rolloff of 1 dB at 44.1 kHz Fs, and 2.8 dB at 96.0 kHz Fs. Microphone low end response seemed to be down 3 dB at about 80 Hz. With the low cut filter switched in, the 3 dB down point was about 400 Hz. Square wave response didn’t have the usual symmetrical shape of FIR low pass filtering. Instead, it had asymmetrical ringing suggesting IIR type non-linear phase filtering Total harmonic distortion plus noise in both a 22 kHz and 80 kHz measuring bandwidth as a function of signal frequency is plotted in Figure 2 for a sampling frequency of 44.1 kHz. As can be seen, increasing the measurement bandwidth to 80 kHz reveals that there are some out of band artifacts that do increase the THD+N within the audio bandwidth. The same measurement done at a sampling frequency of 96.0 kHz is shown in Figure 3. Results were similar for transferred and played back files – in effect, a measurement of the A/D converter. Since the amount of distortion is about 0.02% before it rises with frequency, it would appear that the distortion level is mostly generated in the A/D converter. This was confirmed by a measurement of known low distortion 24 bit Audio Precision generated wave files transferred to the LS-10 and then played back on the LS-10. Measuring its D/A converter in this manner revealed a THD+N of about 0.005% How THD+N for a 1 kHz test signal varies with decreasing input level is shown in Figure 4. The results were the same for both 44.1 & 96.0 sampling frequencies. Input/output linearity was quite good down to about -100 dBFS both sampling frequencies and deteriorated below that input level. This is plotted in Figure 5. Data is shown for both channels and at both sampling frequencies. Channel separation was found to be independent of sampling frequency and, unusually, of testing direction. Channel separation is plotted in Figure 6 for testing directions. — Bascom H. King - August 2008 Fig. 4 THD+N as a function of signal level down from full scale for a 1 kHz test tone. Fig. 5 Deviation from linearity of a 1 kHz test signal for both channels at a sampling frequency of 44.1 and 96.0 kHz. Red/Blue = Lch/Rch at 44.1 kHz Fs, Magenta/Cyan = Lch/Rch at 96.0 kHz Fs. Fig. 6 Channel separation vs. frequency in L > R direction, Red, and R > L, Magenta BENCH TEST continues on page 32 www.proaudioreview.com August 2008 | ProAudio Review | 13 http://www.proaudioreview.com

Table of Contents Feed for the Digital Edition of ProAudio Review - August 2008

ProAudio Review - August 2008 Contents Publisher's Page Broadcast News and New Products Yamaha Pocketrak 2G Olympus LS-10 Linear PCM Recorder Bench Test Marantz PMD620 Flash Recorder TASCAM DR-1 Portable Digital Recorder Tips From The Pros: Location/Production Audio Prosumer Portables in a Pro World Studio News and New Products Apogee Ensemble Solid State Logic Duende Mini FireWire DSP Live News and New Products Avril Lavigne's The Best Damn Thing World Tour Behringer Eurolive B412DSP and B415DSP Active Loudspeakers Gear for FOH Engineers — A PAR Contributors’ Wish List Contracting News and New Products Buyer's Guide Single Slice

ProAudio Review - August 2008

For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.