Tuesday, November 10, 2009

Which Portable Digital Audio Recorder?

In this article I'll continue from my Summary of Portable Digital Audio Recorders. Specifically I'll consider those that work well for gathering environmental and nature sounds. Last time I collated information and provided a general overview of the field. Today I'll be making specific recommendations.

How to begin to compare 21 different recorders? Let's start with the assumption that you want to have an audio recorder on hand always. If you don't have one with you, you'll miss that intriguing sound when it happens. Thus the device needs to be small enough to fit in a pocket and add little to your daily load. That eliminates from consideration the Shoulder units, which work best with external microphones and a backpack full of gear. A more extensive setup has its place, but I recommend that first-time buyers get something easier to use. A hand-held recorder will get you started in field recording and will still be useful once you have a full rig.

N.B. This is only one way to start narrowing down the list of recorders. In my next article I'll consider a different perspective.

The First Cut

For ultimate convenience we'll consider only those units with built-in microphones. This is not to say that we will always want to use the built-in mics since they have several limitations: relatively high self-noise, limited or uneven frequency response and the inability to position them independently of the recorder. It is convenient to have them available, even if the ability to power external condenser mics is still desirable.

This eliminates from consideration models that we would have in any case found other reasons to forgo. For instance, two units do not record to flash media. Neither do they have a dedicated record level control or decent ergonomics. The Sony MZ-RH1 is limited to 1GB MiniDiscs. The Korg MR-1, more a proof of concept than a well conceived field device, records to a 20GB hard drive that will be vulnerable to knocks and spills.

And then there's the Marantz PMD620 which has a strange issue of headphone latency and only permits record levels to be set when it is paused.

Finally, the Samson Zoom H4 will not be considered in favour of its updated unit, which fixes many flaws including the lack of direct record level control.

Without too much blood-shed the contenders have thus been reduced from 21 to 11. Simplifying the chart by eliminating the dimension and mic columns gives the following, sorted by decreasing volume:
MODEL                VOL  MASS  PRICE  EIN  XLR
-------------------  ---  ----  -----  ---  ---
Sony     PCM-D50     365   366  $ 450  126    -
Tascam   DR-100      428   290  $ 380  113    +
Samson   Zoom H4n    382   280  $ 300  114?   +
Tascam   DR-1        256   208  $ 250  115    -
Samson   Zoom H2     230   172  $ 190   99    -
Tascam   DR-07       212   130  $ 170  113    -
Sony     PCM-M10     161   187  $ 300  122    -
Edirol   R-09HR      186   166  $ 300  118    -
Maudio   MicroTrack  174   192  $ 200  106    -
Olympus  LS-11       139   165  $ 400  122    -
Olympus  LS-10       139   165  $ 300  122    -


Now let's add columns corresponding to various characteristics we may be interested in. Questions will be answered with a simple yes ("+") or no ("-"), a rating, or a numeric value, as appropriate.

The first set of criteria concern ease of use in the field. We require a dedicated record button and level controls so we don't need to hunt through menus for such important features. (You'd think this would be obvious in a device used primarily to record, but no.) Controls should be designed for convenience and durability. It is important that the screen be legible in both daylight and dark environments. We need a clear level meter and peak warning so we know when there is a problem with our recording.

The overall build quality is important, since we are considering a piece of electronics designed for field use. Many less expensive units are made of flimsy plastic. Some have cheap surface-mounted sockets which fail after a limited amount of use. Even some of the larger recorders are plastic to save on weight and manufacturing costs. Don't take these to the Antarctic or the desert!

Additionally, it is handy to be able to mark positions within a recording while it is being made. This is especially true for long ambiances that might have occasional distinguishing events. Likewise, it's convenient to be able to name files immediately after recording. These features won't make or break a recorder, but it's useful to note them.

Many people may prefer to capture a single source with a single channel; this is quite common when using a shotgun mic, for example. In these cases it is more efficient to record in true mono but not all recorders support this, instead recording the single source to both channels of a stereo file. This doubles the disk space requirements and makes copying files more time consuming.

To summarise, here's our first set of questions:
* Is the screen legible and clear?
* Is there a dedicated record level knob? (And not just buttons.)
* Can one mark tracks while recording?
* Can one rename/edit files after recording?
* Can one record in true mono?
* Ranking for build quality (lower number better).

There are nine further characteristics to consider. I'll describe each in a separate paragraph so you can decide for yourself which are important to you.

Manufacturers format their media using the FAT file system, so as to gain maximum compatibility. But FAT has one major limitation; it cannot accommodate files larger than 2GB. If you want to record for long durations, you will need a recorder that can seamlessly split files "on the fly", once the 2GB limit is reached. This is often an issue for concert tapers who want to keep their device concealed for the length of a show1. It will be especially important to you if you wish to record at high bit rates, since that chews through disk space quicker2. As it turns out, none of the recorders remaining on the list are stumped by this. (Though be sure you have the latest firmware for your device, as many companies fixed this problem some time after the initial release.)

There are many types of flash cards, but SD seems to have won the format war. I have several devices that share SD cards, so I am never without one. In a pinch I can pluck a spare card from my camera to use in my recorder, or vice versa. Plus, SD cards are cheaper than most of the alternatives. I prefer not to use a device that makes me commit to yet another media format.

A few recorders have built-in memory in addition to a flash card slot. This is handy and adds value, since the cost of the memory is included in the purchase. However, I prefer to use a card, since I can eject it and transfer data while the recorder is still in use. This also avoids the need to carry proprietary USB cables.

I discussed XLR inputs in the last article. They are crucial in order to use professional microphones with phantom power.

I prefer a device that can use AA batteries as they are easy to find anywhere in the world. I already have a good stock of rechargeables plus a professional charger. I'm not crazy about spending more money on yet another proprietary rechargeable cell.

It's nice to get a decent amount of recording time from a set of batteries, since swapping in the middle of a session is a pain. This can be difficult to measure and compare across recorders since it depends on various factors: the type of microphones, whether phantom power or plug-in power is being used, the recording format, how much the screen is used and so on. Nonetheless we'll try to get some sort of a rough measure for the sake of comparison.

If you are recording impromptu sounds, the device should boot up quickly so you don't miss a recording opportunity. Some units can take an extraordinary long time to become active -- not something the manufacturer will mention in their promotional literature.

Some recorders offer a pre-roll buffer. This allows you to leave the recorder in stand-by mode and hit record after you've heard the sound you want. It works like magic but is actually a simple application of a memory buffer, one that is continuously recording in the background. If you collect birdsong or other timely events, this feature can be a godsend.

All the recorders under consideration support 24 bit recording. This is very important since it allows greater leeway in adjusting record levels. But not all recorders provide up to 96 KHz sample rate. This is not a problem when using the built-in microphones, since their frequency range would be well handled by a lower rate. But if you use external microphones capable of capturing very high frequencies, you might want more than 44.1 KHz. As it turns out, the units with XLR inputs also support 96 KHz, so all is well. This characteristic will not impact our table.

In summary, these are the additional columns:
* Does the unit use SD cards?
* How many GB of built-in memory is provided?
* Are there XLR inputs?
* How many AA batteries does it use (0 means batteries are proprietary)?
* How long is the battery life in hours?
* What is the boot time in seconds?
* How long is the pre-roll buffer in seconds?

Here is the new table with information gathered from a variety of reviews.
----------------    ---  ----  -----  ---  ------  --  --  ---  --  --  --  ---
Sony    PCM-D50     365   366  $ 450  126  -++?-1   -   4    -   4  12   8    5
Tascam  DR-100      428   290  $ 380  113  +++--2   +   0    +   2   4   6    0
Samson  Zoom H4n    382   280  $ 300  114? +-+--2   +   0    +   2   6  40    2
Tascam  DR-1        256   208  $ 250  115  ++---2   +   0    -   0   7   8    0
Samson  Zoom H2     230   172  $ 190   99  +----3   -   0    -   2   4  10    2
Tascam  DR-07       212   130  $ 170  113  ++---2   +   0    -   2   9   6    0
Edirol  R-09HR      186   166  $ 300  118  --?+-3   +   0    -   2   5  10    0
Sony    PCM-M10     161   187  $ 300  122  ++++-1   -   4    -   2  24   6    5
Maudio  MicroTrack  174   192  $ 200  106  +-+-+2   -   0    *   0   4  18    0
Olympus LS-11       139   165  $ 400  122  +++++1   +   8    -   2  23   2    0
Olympus LS-10       139   165  $ 300  122  ++---1   +   2    -   2  12   4    0


Now let's have a look at what all this might mean, taking note of some distinctive features of the recorders along the way.

The PCM-D50 is solid and sounds great, with a host of useful features, including a 5 second pre-record buffer. The built-in mic capsules can be moved between 90 and 120 degree positions, though they cannot be made coincident. The 12 hour battery life and 4GB internal memory are generous. Sony has implemented a very clever input limiter feature that deserves explanation. The unit records all incoming audio at both 0dB and -20dB. If the first stream peaks then the second is normalised and is used to fill in what would otherwise be a distorted signal. Brilliant!

On the down-side the screen may not be as readable as others in all lighting situations. Unfortunately Sony uses their proprietary memory stick format; it is the only recorder here that is not compatible with SD cards. But the main problem is that the PCM-D50 is by far the largest unit, and doesn't even have XLR inputs as consolation. With all the high quality features it really should have had XLR in, to take full advantage of the quality preamps.

In line with its professional placement, the unit has digital I/O.

The two Olympus models are identical in appearance except for their finish (the LS-10 is black and the newer LS-11 is silver). They are made of robust aluminum and have excellent usability. Reportedly the LS-10 rolls off frequencies below 70 Hz, even without the low-cut filter engaged. In practice this may not be much of a limitation, since most frequencies in the low bass are undesirable. I have not noticed it as a problem in my usage. (Olympus claims improved sound for the LS-11. But that unit is too new for third-party test results to be in.)

The LS-11 allows files to be divided and moved between internal and external memory. It has an incredible 8 GB of built-in memory. Likewise the battery life has somehow been optimised to 23 hours! Showing Olympus' heritage in voice recorders, the LS-11 can automatically start recording when triggered above a preset input level. I am not sure if that is unique among these models, but it could certainly be very useful3. The units come with a simple remote control in Europe, but it's not part of the package in North America.

The Tascam units are built well; I judge them as second to the Sony and Olympus models. All have similar audio performance but only the DR-100 can record at 96KHz. It better, since good mics attached to the XLR inputs would be ill-served otherwise. The AA batteries last 4 hours; while this is still a decent amount for a single session, it lags behind the competition. Tascam makes gear for musicians and that bias is evident. Features include a tuner, varispeed playback (for practicing), vocal cancellation and the ability to overdub -- all features useful in their market but irrelevant to field recording.

The DR-07 manages a decent 9 hour battery life. The DR-1 stands out of the crowd for using proprietary rechargeables. All of the Tascam units have low gain, which might hamper their use with low sensitivity or dynamic mics.

The Samson units are unique in acting as audio interfaces for a computer when connected via USB. The Zoom H4n is the only 4 channel recorder in this comparison, again betraying the musician-centric design. Build is more impressive than its predecessor, the H4, which was poor. Meanwhile the Zoom H2 is novel in offering four built-in microphones and a variety of recording modes including mid-side. Both units share an Achilles heel in the lack of a dedicated record level dial. Push buttons simply do not suffice for quick field adjustments. However, for the target market that is not really an issue. The Zooms are not made to be hand-held field recorders so much as they are designed to be mounted on stands to record band practices and the like.

The Zooms take a lot longer than other units to power on. Battery life is on the low end. On the plus side they have a pre-roll buffer. If you are thinking about using the H2 with external mics... don't. The EIN is much worse than any other unit in this comparison, and the sound is actually better with the built-in mics.

The Edirol R-09HR has weaker build with a below average screen (poor in bright light). It has no dedicated record level knob and insufficient PIP to drive many microphones. Those are fatal flaws. On the other hand it comes with a handy remote control with more functions than any other brand. They've fixed the bad solder connections on the predecessor and done away with the poorly designed battery compartment. A firmware update added a metronome, tuner and timer for recording.

Finally, there's the M-Audio MicroTrack II. M-Audio were one of the first companies to target this new market and I treated their debut recorder, the 2496, in detail4. It never lived up to expectations but the newer unit corrects many flaws, for example having correct 48V phantom power. It does not have XLR jacks, however, connecting using 1/4" TRS jacks. This will in most cases require a cable adapter. Though solidly built, the MicroTrack has a proprietary built-in battery that is not user serviceable. This immediately eliminates it from consideration in my eyes; what if I want more than four hours of operation away from a power supply?

Further limitations include an amplification stage that overloads fairly easily; some users recommend an external preamp, which rather defeats the purpose of a convenient all-in-one unit. It also has no low frequency cutoff filter. I note in passing that though it does not have a built-in mic, it comes with one. Technically it should have been eliminated from this round-up, but its popularity argued for inclusion. Finally, on the plus side, the MicroTrack II has an S/PDIF input, so it can be used as a bit-bucket if you wish to put a separate A/D converter in the signal chain. That sort of operation is outside the scope of this article. And why not just buy a unit with a better gain stage to begin with.

New Addition

With the new PCM-M10, Sony has shrunk their PCM-D50 while adding battery life and improving the microphones. It now has an internal speaker for quick reviewing. It can record MP3s, while its older brother can only play them back.

It is the first Sony unit to accept MicroSD cards (along with MMC). While I'd argue it should have been compatible with regular SD, it's still an improvement over their proprietary nonsense. Instead of requiring 4 AA batteries to get 14 hours of recording, it requires only 2 AAs and gets 24 hours. That's neck and neck with the LS-11. The unit is half the weight and significantly smaller than the PCM-D50. About the only significant limitations are the fixed microphones and the absence of digital I/O.

The remote control, previously an option, is now included in the price. And all this for $150 less than before. Incredible.

Decision Time

There are several units I cannot recommend due to the distinctive flaws discussed above. Here I will point out the best choices.

If you need XLR jacks then choose between the DR-100 and Zoom H4n. They are similar in size, price and audio quality. Although the latter has longer battery life and a buffer, I would choose the Tascam in order to get faster start-up performance and a nicer build, not to mention a proper record level control. However, I am not sure it makes sense to get an XLR recorder in this class, since the main idea was portability. Besides, the quality of the pre-amps will then become a limiting factor, and these units do not have the best EIN. (More about this in the next article.)

The Sony units have the best sound quality and some cool features. The main problem with the PCM-D50 is that it is simply too large, without even the benefit of XLR inputs to justify its size. If it had those connectors it'd be great; but as it is, it falls between two stools: a pro recorder that cannot avail of microphones of comparable quality.

The Sony PCM-M10 is too new for reviews to be available, but judging from the specs it is recommended for portable use, thanks to its long battery life, great features and emphasis on sound quality.

The Olympus LS-11 is smaller again while maintaining excellent sound and unsurpassed ergonomics. It has so much built-in memory you may never need to buy a card. And it is tied for longest battery life. Until I know more about the real-world use of the PCM-M10, the Olympus LS-11 is tied for best hand-held recorder. It's also the winner if you need the longest possible recording time at 96/24, managing 13.5 hours if you add a 32GB card.

Meanwhile, the Tascam DR-07 provides the biggest bang for the buck, even including a 2GB SD card in the package. This is the choice if on a limited budget.

In the next article I'll focus on sound quality, and pairing the right microphone to the right recorder.


1 I suggest readers interested in that application consult the experts at Tapers Section.

2 2GB will hold about 3 hours of sound at 44.1 KHz / 16 bits but just under an hour at 96 KHz / 24 bits.

3 Turn the unit on at midnight and have it automatically start recording when the dawn chorus heats up in the morning -- while you sleep in!

4 See MicroTrack versus MiniDisc, MicroTrack Update, MicroTrack vs. MiniDisc Chart, MicroTrack driver released and MicroTrack Review, Tip, Firmware Update.



Anonymous said...

You've eliminated the Zoom H2 at the outset due to noise issues, but those noise issues relate to the external mic input, not the internal mics - and note that as it records two stereo pairs (front facing and rear facing) it can capture soundscapes in a way no other such recorder can. See for instance the example I posted in first post at http://www.2090.org/zoom/bbs/viewtopic.php?p=96859

robin said...

You are dead right. I am correcting the article, which needs a few other tweaks in any case. Thanks for taking the time to comment.

robin said...

Added the amazing Sony PCM-M10 to the list.

pipmics said...

"The Edirol R-09HR has weaker build with a below average screen. It has no dedicated record level knob and insufficient PIP to drive many microphones." - Sorry wrong, I can easily run 3 PIP mics in parallel from the R09HR's PIP supply - dont you mean the LS10 in this regard?

robin said...

picnet: I have done a bit more reading and both recorders you reference supply 3V PIP, which is less than the 5V spec but more than devices like MiniDisc. I will likely drop the reference, since I will cover microphone usage more in a future article.

robin said...

I've updated the power-on timings with information taken from the excellent Avisoft Field Recording Tutorial, which you can read as a supplementary approach. I have also added a column to indicate if one can record in true mono.

Anonymous said...

I record video with a Panasonic DMC-GH1 and now want to improve the audio quality. My primary use is recording martial arts at roughly 30ft and theatrical performances at roughly 20ft. The GH1 has automatic gain control, so I'm inclined towards a digital audio recorder. My choices are: (a) on-camera shotgun mic (eg., rode mic) plugged into the camera or (b) portable digital audio recorder mounted on the camera. Which will give me better sound?

robin said...

I am not intimately familiar with the audio abilities of the Panasonic DMC-GH1, but I am sure a dedicated recorder will be better. However, it's all down to the mic. If you need to capture sound from a good distance then you need either a parabolic dish or a boom mic placed close to the action.

There is no substitute for this! If you cannot get close to the sound you will be getting a lot of room sound. And likewise not enough direct sound. The recorder is likely NOT the limiting factor here.

Anonymous said...

Thanks, Robin. The Zoom H4n is fast becoming an industry standard in the digital video world but I am less than impressed by the recordings I have heard from its internal mics. What would you recommend for pure on-camera audio with the internal mics?

Anonymous said...

It looks like you took the device weights from the manufacturer's spec sheet. If so then those numbers aren't comparable because some manufacturers quote weights with batteries (Sony), some without batteries (Tascam, Samson), and some aren't clear (Olympus).

robin said...

Hey, anon, you are dead right about the weight. I tried to figure consistent weights with or without batteries, but gave up. Besides, weight also depends on accessories, like whether a strap is attached etc. Still, for smaller units 2 AA batteries makes little difference; I would not be too concerned about differences under 50g. (In fact weight was rarely a deciding factor for me within one of the three categories.)

robin said...

"The Zoom H4n is fast becoming an industry standard in the digital video world but I am less than impressed by the recordings I have heard from its internal mics."

I would say for video you always need some sort of a shotgun, and furthermore should get it off camera if at all possible. The mics on the portable recorders are either omni or cardioid and are simply not directional enough to omit off-axis sound.

That Zoom is the cheapest possible multi-channel recorder, which might be what makes it popular for that application. But bear in mind that video work is a different application from what I am focused on here. Ultimate low noise is not always so important. Anything important is dubbed in post anyway.

Anonymous said...

It mentions a third part to this article?

robin said...

That would be this one: Choosing An Audio Recorder For Ultimate Sound Quality

Unknown said...

You might also add the Edirol R-44 to this comparison. Four-track, on-board mics, XLRs with phantom power, limiter and more.

Unknown said...

You wrote:

"Though solidly built, the MicroTrack has a proprietary built-in battery that is not user serviceable. This immediately eliminates it from consideration in my eyes; what if I want more than four hours of operation away from a power supply?"

Add a simple, cheap and small USB battery pack. It plugs into the MTII's USB port.

Paul said...

A couple corrections regarding the Marantz PMD620...

1. It does have internal microphones.

2. Recording levels *can* be adjusted during recording. It does not need to be paused.

tom said...

loving your site, so much good info!
i'm in need to jump ship from my mz-m10 minidisc recorder for timeconsuming import issues on my mac, so i've been looking into the ls-11, the pcm m10, both which you analyze thoroughly.

i also came across the Tascam DR2D Portable Digital Recorder, any thoughts on that unit? thanks again for a great resource!
cheers, tom

robin said...

Tom: I have no experience with the Tascam units, of which there have been several. They are more oriented towards musicians, with features like built-in metronomes and the option to overdub audio. I believe the DR2D can record from two sources at the same time, which might certainly be useful to some. But as for audio quality... I do not know.

tom said...

no worries, thanks robin! it's difficult to compare all these products with their strength and weaknesses differing, but finding your articles has def made it easier! i'll just try and check out the tascam at a store...peace, tom

Unknown said...

I want to record a 3 - 4 piece band and get good quailty- Which recorder would be best.

robin said...

Time: Either / both / none depending on how the band sounds live, whether you are planning on buying external mics, if you are taking a feed from a sound board, etc.

Anonymous said...

Hi. Has anybody else had major sync problems with the LS-10 and the Panasonic GH-1??? We bought an LS-10 (guided in large measure by the wonderful info here) and have found that the (audio from the) camera and the LS-10 disagree by something like 11 (PAL) video frames after only something like 25 minutes. To me that seems unbelievably huge. Any advice would be appreciated.


robin said...

Sam: That's only a 0.02% difference which seems incredibly good synchronisation for two devices running free. In any NLE it's easy enough to either chop the audio to insert tiny pauses, or to stretch to fit.

Anonymous said...

Hi Robin,

Thanks for your answer, but it seems large to me. After all, according to Wikipedia typically a "quartz wristwatch will gain or lose less than a half second per day..." and I remember reading some time ago that crystal oscillators have accuracy to something like 6 parts per million. Or even if it was 60 parts per million that's still actually some 5 times better than we're getting. Huh. I dunno. Have you tested any of these recorders for their clock accuracy by any chance, and if so are our results typical???

Oh and by the way what we've done to correct the error is to use the wonderful open source utility SoX ("the swiss army knife of audio") to resample the files. Works very nicely, and it can be done from the command line as a batch. I could post the command if you want (though it'll have to be in a couple of weeks as at the moment I'm out of town and won't be back to my computer till then).

Thanks again for a very nice web site.

robin said...

Thank you Sam; I am glad to be of help.

While I haven't tested, I know from experience that even much shorter lengths of audio require fiddling to match lengths. There's a reason people use time code synchronisation even for short shoots!

Anonymous said...

These tables would be 10x more legible if you knew how to code a simple HTML table. Why bother presenting the data if the column headings don't line up!?

robin said...

Thanks for so kindly and politely pointing out the editing error that crept in the last time I updated this article. As for tables, I have had numerous issues with Blogger interpreting them incorrectly and applying spurious formatting. Besides which, it is much faster to compile, create and edit simple tables in the format used here.

I note that only the headers were in error and return the favour you so pleasantly bestowed upon me: If the columns truly do not line up it is because a) you are incorrectly applying your own CSS to the PRE tag, or b) you need to see an optometrist.

robin said...

For those reading this older article: time moves on and I would now most heartily recommend the Olympus LS-5. It's the same as the LS-11 in audio terms, but has less memory and packaged accessories. No recorder has better features and audio quality for the money.

Post a Comment