![]() |
Division of Registrar & Student Services
Disability Services Unit
|
Digital Recording DevicesWhat are they? How do they work? Which ones should I use?Trevor Allan, Manager,
Disability Services Unit, Australian National University & ContentsBackgroundIn the good old days, audio recording was very simple. The ubiquitous cassette player was the simple answer to most portable recording situations. The only major concerns were the type of recorder and whether you needed an external microphone. Controls were simple and intuitive, and everybody and his dog could learn to use one quite quickly. If you needed to record a lot of material, you simply needed to make sure you had enough cassettes to record on. However, for all its simplicity and ease of use, the humble cassette had its limitations. The most obvious is that associated with any analogue device – linear recording. The material is recorded in a continuous line, and to find any particular part of that line, the tape has to be physically moved forward or backwards to find the relevant section. This means that finding material is time-consuming, imprecise and cumbersome. Linear recording is fine if you simply want to start listening and then listen all the way through until the end. Anything else is problematic. Another problem with cassettes is that they are difficult to transfer to a computer or other device. The recording has to be played back in real time and a means of transferring it to the computer needs to be established. Copying tapes is time-consuming (even with high speed copiers), tapes get tangled and twisted in the machines, heads get dirty, complicated mechanical tape transport mechanisms break down and so on. It is not possible to use Voice Recognition to transcribe cassette tapes directly. Also, if a copy is needed, tape to tape copying degrades the quality of the signal. Enter the Digital Recorder! No tapes to tangle, degrade or break. Files can be easily transferred to computers and other devices such as MP3 players. Copying recordings is fast, easy and simple. Recordings can be located on the web, they can be bookmarked, edited and divided simply and quickly and so on. Digital Files can be named and filed in a manner that doesn't take as much physical space as tape. In most instances, it doesn't take any more physical space at all - just memory space on a computer drive. These, and many other advantages of digital audio recordings have seen the explosion of digital devices recently, with an associated explosion in the range, types, functions and quality of devices. To try to clear up some of the confusion around digital recording devices, The Australian National University and the University of Wollongong agreed to a joint research project to investigate the range, effectiveness, functions and applicability of different digital audio recording devices.
Types of Digital Audio RecordersThe types of Digital Audio recorders tested fall into the following categories:
Flash MP3 Recorders – these are compact, relatively small capacity recorders that use an inbuilt Flash disk to store recordings. They are cheap, readily available and popular for storing music in MP3 format, with audio recording a secondary function. They usually have:
Multifunctional Devices -- these are devices such as mobile phones, Cameras and Portable Data Assistants (PDAs), etc that have voice recording capability as a subsidiary function of other purposes. Because these devices only see recording as a minor subsidiary function, they tend not to have a high-quality recording capability, but with recent developments such as increased storage capacity and the convergence of different specialist technologies these devices were included in the tests with some surprising results. Figure 4: Palm LifeDrive 4 GB PDA
Devices Tested
|
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
yes - 3.5 plug |
no |
Yes 2 AAA |
Wav |
128Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 352Kbps |
Standard |
19 Hours 30 |
66 Hours 50 |
11.5 H recording 19.5 H |
Test results:
Default Sampling Rate - No Mic
These files were transferred over from the Panasonic with the default sampling rate
HQ-16 FQ-16 SP-8
| SP Meeting |
quite distorted, compression artifacts very obvious, can hear what is said (just) |
| HQ Meeting |
good recording, quite a bit of background noise |
| HQ Dictation |
good recording, little background noise |
| FQ Meeting |
quite distorted (less than SP meeting) can hear what is said, compression not so obvious |
Highest Sampling Rate - No Mic
These files were transferred over from the Panasonic with the highest sampling rate
HQ-22 FQ-11 SP-11
These files were marginally better than those transferred over at the lower sampling rate. The lower resolution recordings suffered the most from the difference in transfer resolution but the difference is hardly discernable when comparing the higher resolution transfers.
Meeting
| Meeting Internal mic |
background noise, difficult to hear |
| Meeting Cheaper Sony mic |
higher tone than above but more distorted, background noise |
| Meeting Expensive Sony mic |
a bit clearer than above but still difficult to hear |
Dictation - Default transfer resolution
| Dictation Internal mic |
less noise than meeting setting but a bit difficult to hear |
| Dictation Cheaper Sony mic |
some noise but soft recording, difficult to hear |
| Dictation Expensive Sony mic |
a bit clear than above, can discern better |
Dictation - Highest transfer resolution
| Dictation Internal mic |
clearer than that transferred at default resolution |
| Dictation Cheaper Sony mic |
higher tone recording a bit more listenable than above, slightly quieter recording |
| Dictation Expensive Sony mic |
soft recording but easier to listen to, some distortion noise but less than others |
Voice Recognition Transcription Panasonic US395 (Carl Runeson)
Using the poem The Puddin' Thief (Norman Lindsey) as a reference I created a user file with the Panasonic. Below are the results from using Dragon naturallySpeaking Professional Version 8 to transcribe from these recordings.
A dictation test with the internal microphone gave an accuracy of about 15%
A dictation test with a cheap microphone gave an accuracy of about 25%
A dictation test with the expensive Sony microphone gave an accuracy of about 35%
Using the poem The Raven by Edgar Allen Poe
A dictation test with the expensive Sony microphone gave an accuracy a little over 80%.
Summary.
The upshot of testing the Panasonic RR-US395 is that you really need
to speak clearly and loudly and very close to the microphone to be
able to use the voice recognition software with it at all. This finding
seems to be supported by the literature on the internet. The best
response was with using an external (Sony) microphone.
It may be that in the future with higher capacity devices the bit
rate quality may be able to be increased leading to better voice recognition.
As it stands I would personally not use a digital recorder with Dragon
- I would use a USB microphone plugged into the PC running the software
in preference.
The recordings were performed in Room 203. Room 203 is a medium sized room suitable to seat about 20 to 30 people in a small lecture/tutorial type setting.
The recorders were all placed on a table and then turned on and set to start recording. The recordings consisted of me reading out "The Raven" by Edgar Allen Poe while walking from the recorders around the room and behind the recorders for the length of the recital of the some of the poem. An exception to this was the Toshiba GigaBeat which had lost its battery power over the three days since it was fully charged and was done at a later stage with condition similar to the other tests.
DEVICE |
RESULTS |
|---|---|
| Creative N200 Internal mic | very soft recording, not really discernable, not a lot of background noise though |
| iRiver H10 Internal mic | good recording (bit soft) but a high pitched background noise |
| iRiver T30 Internal mic | softer recording (than above) a little background noise (not a problem though) |
| Toshiba GigaBeat Internal mic | OK recording but an annoying buzzing noise |
| iRiver H340 Internal mic | good recording (stereo), some background noise |
| iRiver H340 Cheaper Sony mic | good recording, less noise than above |
| iRiver H340 Expensive Sony mic | very good recording, crisper than above, a little noise (not annoying) |
| SanDisk internal mic | dull record, indiscernible at a distance of more than 1m |
| TechWorks 256 MP3 Player | reasonable good recording for the resolution |
| Ricoh Caplio R4 Internal mic | a lot of noise, dull recording but listenable |
| Sony Ericsson k750i internal mic | soft recording, no good at a distance of more than 1m |
| Palm LifeDrive | noisy recording, voice barley discernable |
Note: Sony Ericsson recordings converted using Sony Ericsson conversion software which converts (up) to 128kbps.
| DEVICE | BIT RATES |
|---|---|
| Creative N200 | 32kbps (there doesn't seem to be any way of changing this setting) |
| iRiver H10 Internal | 128kbps |
| iRiver H340 Internal | 128kbps (lowest setting for H340 is 40kbps) |
| iRiver H340 External | 320kbps |
| iRiver T30 Internal | 128kbps |
| Toshiba GigaBeat Internal | 192kbps |
| SanDisk internal mic | 32kbps (can’t seem to change this either) |
| T echWorks MP3 256 Player | 32Kbps |
| Ricoh Caplio Internal | l 64kbps |
| Palm LifeDrive | 33Kbps |
| Panasonic Digital Recorder | Highest possible transferred recording is 352kbps HQ setting
(transferred at 22Khz with packaged voice editing software) Conversion with a 16Khz settings gives a 256Kbps recording. |
A test was conducted using a H340 as the digital recorder using its internal microphone recording at 128Kbps (44Khz, 16 bit). The H340 records in MP3 format so the files had to be converted to a mono wav file. The Creative Audio Converter (that comes free with the purchase of a Audigy USB sound “card”) was used for conversion. Several attempts were made at creating a voice file for setting up Dragon but all attempts created files with an inappropriate signal to noise ratio usually between 12 and 14. bearing in mind that the H340 is not designed for this kind of work. The accuracy was about 65%. The setup voice files were recorded in environments with some background noise in one case a noisy air conditioner and in the another case a very noisy laptop. This could have contributed to the unsatisfactory signal to noise ratios.
Note: The different devices record at different bitrates.
Phone quality is considered about 64kbps at 8Khz sampling, a cell phone is 13kbps at same sampling rate
The more expensive Sony microphone is definitely better than the cheaper model.
The line-in on the GigaBeat and iRiver could not be used with an external microphone as one can on the H340. A powered microphone is needed. Using one negates the advantages of portability the MP3 player affords.
Microphone recording wasn't even tried on the Creative n200 as the connector is the smaller 2.5mm which I have no connections for.
"I thought I would just explain the possibly obscure choice of reading material. As I was going to be reading this stuff out over and over again I thought I would choose something that would present a bit of a challenge to the voice recognition software but was also going to be something I would enjoy reading, hence the choices. It really makes little difference apart from highlighting that some dated language will be a challenge to voice recognition as the accuracy was measured against the same texts.
A group of programmers have hacked some of the iRiver, Archos and iPod devices to come up with an alternative User Interface (UI) which may be of use to people unable to use the physical controls, such as a voice user interface (I haven’t tried this yet but may do it soon with my iRiver H340)
More information at: http://www.rockbox.org/twiki/bin/view/Main/FeatureComparison
iPods were not considered in this review as voice recording is not standard on these devices. Voice recording can be facilitated on these devices with the help of external hardware. Adding external hardware does wander away from the benefits of ease of use and portability and using the one integrated unit."
iRiver T30 Flash MP3 Recorder 512 MB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
no Line In |
Yes |
Yes 1AAA |
MP3 |
40Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 128Kbps |
Standard |
|
6 Hours 20 |
Up to 20h |
Toshiba GigaBeat Flash MP3 Recorder 512 MB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
no Line In |
Yes |
Rechargeable |
MP3 |
32Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 192Kbps |
Standard |
|
|
Up to 14h |
Techworks MP3 Recorder 256 MB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
no |
Yes |
Yes 1 AAA |
MP3 |
32Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 32Kbps |
Standard |
17 Hrs 30 |
17 Hours 30 |
Up to 10h |
SanDisk Flash SDMX1 -512 Disk MP3 Recorder 512 MB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
no Line In |
Yes |
yes 1 AAA |
MP3 |
32Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 32Kbps |
Standard |
25Hrs 35 |
35 Hours 35 |
Up to 15h |
Creative Muvo N200 Flash Disk MP3 Recorder 512 MB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
no Line In |
Yes |
yes 1 AAA |
MP3 |
32Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 32Kbps 192(LI) |
Standard |
|
|
Up to 15h |
River H10 hard Disk MP3 Recorder 20 GB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
no Line In |
Yes |
Rechargeable |
MP3 |
32Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 128Kbps |
Standard |
337 Hrs 12 |
1348 Hours 50 |
Up to 16h |
iRiver H340 hard Disk MP3 Recorder 40 GB
| Internal Mic |
External Mic |
FM Radio |
Batteries |
Recording Format |
Min Rec Rate |
| yes |
yes - Line In |
Yes |
Rechargeable |
MP3 |
40Kbps |
| Max Rec Rate |
USB Interface |
Min Rec Time |
Max Rec Time |
Battery Life |
| 128Kbps 320 (Line In) |
Standard |
|
|
Up to 16h |
Digital Audio Device
|
Page last updated: 31 October 2006 Please direct all enquiries to: Student Business Solutions Page authorised by: Registrar |
| The Australian National University — CRICOS Provider Number 00120C |