Settings for Olympus LS-P1 voice recorder help (1 Viewer)

[John Hague]

Hi all, time to get my Olympus LS-P1 setup for recording calls. I’m completely new to this and have no idea where to start. I have a small external mic (with a dead cat).

What settings should I use to be ready for recording. I honestly have no idea so all help much appreciated.

Thanks.

 

[Jon.Bryant]

I don't know the Olympus LS-P1, but you want to record in Linear PCM format as this is uncompressed and better quality than MP3. The bit rate determines the dynamic range (the bigger the number the greater the difference between the loudest and quietest thing you can record). The sampling rate has a impact on the max frequency of the sound you can record - this is 1/2 the sampling rate. As most birds vocalise below 10kHz, and field recordings normally suffer from ambient noise, CD quality (44.1kHz at 16 bit) is generally OK. However, if you have plenty of memory, you can select anything up from CD quality - on your device up to 96kHz 24 bit. It would be worth trialing higher quality setting, just to see whether you can detect any change in the recordings.

When recording, I would try selecting the manual gain option, rather that the 'presets' or auto level. You will then need to manually alter the gain so that the recording levels (as shown in the meters on the device screen during recording) are good, but not too high (you don't want the levels to go over and for clipping to occur as this ruins the recording). You may well find that with a bit of trial and error the gain level can be left at one setting - although of course if something pops up ultra close and blasts its head off, you are certainly going to have to change the recording level - assuming the standard setting you have adopted is for more distant and quieter sound.

Good luck!

 

[John Hague]

Thanks John.

 

[Ruff-Leg]

I don't know the Olympus LS-P1, but you want to record in Linear PCM format as this is uncompressed and better quality than MP3. The bit rate determines the dynamic range (the bigger the number the greater the difference between the loudest and quietest thing you can record). The sampling rate has a impact on the max frequency of the sound you can record - this is 1/2 the sampling rate. As most birds vocalise below 10kHz, and field recordings normally suffer from ambient noise, CD quality (44.1kHz at 16 bit) is generally OK. However, if you have plenty of memory, you can select anything up from CD quality - on your device up to 96kHz 24 bit. It would be worth trialing higher quality setting, just to see whether you can detect any change in the recordings.

When recording, I would try selecting the manual gain option, rather that the 'presets' or auto level. You will then need to manually alter the gain so that the recording levels (as shown in the meters on the device screen during recording) are good, but not too high (you don't want the levels to go over and for clipping to occur as this ruins the recording). You may well find that with a bit of trial and error the gain level can be left at one setting - although of course if something pops up ultra close and blasts its head off, you are certainly going to have to change the recording level - assuming the standard setting you have adopted is for more distant and quieter sound.

Good luck!

Jon, I often contemplate using the 96 kHz sampling rate, but then I realize my microphone's listed frequency response range doesn't go up to 48 kHz, so I've always stuck with 48 kHz sampling.

 

[WRL]

Jon, I often contemplate using the 96 kHz sampling rate, but then I realize my microphone's listed frequency response range doesn't go up to 48 kHz, so I've always stuck with 48 kHz sampling.

Those two frequencies have a different meaning, however have some relation in terms of analog-digital conversion.

• Sampling rate defines the the number of instances per second in which the analog signal is converted into a (binary/digital) number.
• Microphone frequency range (let's assume from 20 to 20000 Hz) is what it says, the acoustical frequency range which the microphone can technically convert from sound into (analog) voltage.
• In order to properly convert analog (voltage) signals into digital numbers it needs at least double the convertion (sampling) frequency as the highest sound frequency (Nyquist or Shannon Theorem). So for a 20 kHkz sound 40 kHz sampling frequency is the minimum. 48 kHz sampling would always be ok for a "normal" microphone. However, a higher sampling rate provides a "finer digital representation " (signal over time) compared to a lower sampling frequency. This may be an advantage when later manipulating the (digital) sound file.

Werner

 

[Jon.Bryant]

Those two frequencies have a different meaning, however have some relation in terms of analog-digital conversion.

• Sampling rate defines the the number of instances per second in which the analog signal is converted into a (binary/digital) number.
• Microphone frequency range (let's assume from 20 to 20000 Hz) is what it says, the acoustical frequency range which the microphone can technically convert from sound into (analog) voltage.
• In order to properly convert analog (voltage) signals into digital numbers it needs at least double the convertion (sampling) frequency as the highest sound frequency (Nyquist or Shannon Theorem). So for a 20 kHkz sound 40 kHz sampling frequency is the minimum. 48 kHz sampling would always be ok for a "normal" microphone. However, a higher sampling rate provides a "finer digital representation " (signal over time) compared to a lower sampling frequency. This may be an advantage when later manipulating the (digital) sound file.

Werner

Yes agreed.

People use the lightbulb analogy, but you could also consider the moon. If you sampled the lumens from the moon every 28 days, you would not detect much difference - you would think it was in a near permanent state. Sample every 14 days and you would detect a significant change and if you assumed it was a sinusoidal pattern you could reasonably accurately calculate the in between states. But if you sampled at a much higher frequency, you may start to detect finer details (due to distance to the sun, distance of the moon etc.).

In theory higher sampling rates create better digital approximations of analogue sound. That said, most bird vocals are below 10kHz, so even sampling at 44.1kHz the sound wave is being sampled a minimum of 4 times in each wave cycle…. And a lot of us can’t hear sound as high as 10kHz in any case, so perhaps you will just get a more accurate sonogram and not any audible difference.

 

[WRL]

Yes agreed.

People use the lightbulb analogy, but you could also consider the moon. If you sampled the lumens from the moon every 28 days, you would not detect much difference - you would think it was in a near permanent state. Sample every 14 days and you would detect a significant change and if you assumed it was a sinusoidal pattern you could reasonably accurately calculate the in between states. But if you sampled at a much higher frequency, you may start to detect finer details (due to distance to the sun, distance of the moon etc.).

In theory higher sampling rates create better digital approximations of analogue sound. That said, most bird vocals are below 10kHz, so even sampling at 44.1kHz the sound wave is being sampled a minimum of 4 times in each wave cycle…. And a lot of us can’t hear sound as high as 10kHz in any case, so perhaps you will just get a more accurate sonogram and not any audible difference.

Jon: That is a nice picture to explain it.

Unfortunately, if I may, it is not the whole story or reason for the Nyquist-Shannon condition.
The other problem lies in the way a digital audio signal is being reproduced, or better, should be "truely" reproduceable to represent the analog source signal correctly.
Simplified explanation: During the technical analog-digital sampling process not only the original frequency spectrum is unambiguously imaged or mapped into the digital domain but also many identical and mirrowed digital spectra as multiples of the sampling rate. In the case the sampling rate is lower than the maximum (analog) audio frequency these spectra will (partly) overlap. Now we don't have an unambiguous part in that part of the spectrum anymore (because of original and mirrowed spectra). It is easy to understand that the "retransformation" from digital to anlog will result in "false" analog signals in that overlapping frequency range (so-called aliases). In any case it needs (during digital-anlog transformation), in addition to abiding by the Nyquist-Shannon Theorem), always a "low-pass" filter (with half the sampling frequency) to exclude all those undesired digital spectra.

 

[Jon.Bryant]

aliases

Yes. I am not great on the theory, but I am aware on anti-aliasing.

Presumably if these false signals are multiples of the sampling rate, then they shouldn’t really have an impact on bird acoustics (unless we pick a very low sampling rate)? Presumably even at 44.1KHz these false artifacts will be well above our hearing threshold?

It is interesting that CD quality has generally been deemed good enough for music. There was talk of audio DVDs with higher sampling rates and bit rates, but they never really took off. Ultra high definition music only really became mainstream when streaming services introduced higher resolution (originally for a higher price). There has been a lot of chat since then about whether this a marketing gimmick to crank up the price, or if the audiophiles can really perceive the difference.

That said, if you master at a higher bit and sampling rates, you can always convert down later. I never seem to experience any issues when I convert audio to a lower resolution - although conversion to lossy MP3 (even at highest quality) noticeably impacts the sonogram (and presumably the sound quality).

 

[WRL]

[...] Presumably if these false signals are multiples of the sampling rate, [...]

Not really. In the case the sample frequency f'sample is lower than twice the maximum audio frequency f'max frequencies in the range between f'sample/2 and f'max (i.e. the "upper end" of the audio spectrum) will be mis-reconstructed and will appear as lower frequencies than the original ones and with different amplitudes. Thus the "aliases".

The "CD-sampling rate" of 44.1 kHz is more than 2x the maximum audio frequency of 20 kHz. However, 20 kHz is hardly something one could hear. The additional margin (2.205x instead of 2.0x) is required to lower the technical requirements for the mandatory low-pass filter. A safety margin, if you will.

There could be a lot more to (and more detailed) be written about this topic. But I think this would not be of general interest.

 

[besaide]

Hi,

I have the lsp1 and lsp5

Lsp1:

Rec menu
Rec level: manual
Rec mode: pcm -- 44.1khz mono if you have a mono external mic.

Once you push rec you will see PAUSE on the screen, you can set the gain pushing left and right buttons.

Day birding with external mic or parabolic dish I like rec level 6.

Nocmig with little noise in the area rec level 15.

Internal mics are not good for birding.

Device menu
Memory select - micro SD card (32gb max).
Internal memory is very slow if you record big files, it takes an eternity to download.