UPDATED WITH CORRECTIONS & MORE INFO

The Tascam US-122L is an upgraded version of the US-122. The only real additions I can see are a sampling rate or 96KHz (upgraded from 48Khz max) at 24-bit, and USB 2.0 (upgraded from USB 1.1) to handle the extra bandwidth. From a physical perspective, the knobs are now wider and nearly flush with the unit, meaning you can't use your knuckles to turn them anymore. This was apparently to prevent in-bag damage to the flimsy-feeling knobs of the US-122, which contrasted strangely with its otherwise extremely solid feel. As for me, I prefer a raised knob I can get my fingers around, as I transported my US-122 in its original box anyway, which was pretty small and effective for transport. Aside from that, the differences appear just cosmetic.

You must read my review of the US-122 (not L) on this site before considering purchasing. The US-122 crashed or hung virtually all Windows PC's in Standby AND Hibernate mode. Tascam never specifically mentioned this in any of the literature I ever read or heard about. The farthest they went was a vague suggestion to "turn ACPI off" in the manual as a troubleshooting help if you're having problems. But the fact is, the US-122 just wasn't Windows-compatible, hanging or crashing virtually every Windows computer in the course of proper normal Windows use.

Don't believe any ignorant review which says it's normal. It's not. The US-122 was to my knowledge the only audio device ever to crash virtually every Windows PC on standby and hibernate. While it may be normal to tweak and compromise your system performance a little to have better/proper device performance, it's NOT normal to force Windows users to not be able to use major elements of their operating system.

Ever since the US-122 came out, Tascam's tech support has been worse than horrible. How is it worse? Well, if they don't even pick up the phone or reply to you, that's worse than horrible. I couldn't get through to a live person and left two separate voicemails to tech support, hearing a promise they'd call back promptly. Those calls never came. I then sent an email, which was also ignored.

Those who have gotten through to them report getting attitude, feeling rushed, and not having their problems resolved. To me this is indicative of extreme penny-pinching over at Tascam, designed to maximize profits short-term, which is how many companies are run now, as many CEO's don't stick around long, and do whatever it takes to make the company more money in the short-term, even if it means running the company into the ground long-term (they'll be gone by then).

Another note on the US-122L. Although what I'm about to say is likely to change, it is extremely difficult to get usable information on this product. The company just doesn't help the process along.


Here are 2 main questions about the US-122L you should have answered before you purchase:

1. Is 96 Khz/24-bit recording full-duplex (or only half-duplex, which would render the feature less useful and more in name only)?

2. Does the US-122L crash or hang Windows computers on Standby and/or Hibernate like the US-122 did?


The fact that I *can't even find out* despite my research for my reviews means to me that Tascam is not forthcoming about their product info. Sure, it can record at 96/24, but many audio interfaces today won't go at full duplex in their highest-bandwidth modes. But at least other companies are more honest about it or you can find information more easily. For instance, Edirol's UA-25, which competes directly with US-122L, can't do full-duplex in highest-bandwidth mode. This info is easy to find so you know what you're getting. Not so with Tascam (at the time of this writing).

Edirol's UA-25 is an upgrade of their venerable UA-5. The UA-5 competed directly against the US-122 and was similar in features and price, and the UA25 now does the same for the US-122L, with similar upgrades. The difference is that Edirol produces stable drivers which actually work and which are actually "compatible" with Windows. I found this means more than anything else on a device. Edirol simply has the best reputation for producing the best, most widely-compatible drivesr, but I (unfortunately) do not own an Edirol product, being a disenchanted Tascam buyer. Many companies produce stable drivers, and no company to my knowledge produced drivers as disastrous as Tascam's US-122.

Another question you should find an answer for (a question whose answer I've searched for but can't confirm) is whether the US-122L's included software is powerful enough to make full use of the device's features. I say this because the included Cubase LE of the US-122 could not record at the 24-bit/48Khz the US-122 was capable of, rendering the product useless for that purpose if you stuck with the included software (I find Cubase LE an otherwise incredibly bad bit of software anyway, by the way; your mileage may vary). Cubase LE, which also ships with the US-122L, may be upgraded by this time to be able to record at 24-bit/96Khz (and, for that matter, 24/48), but this is something you or someone ought to verify before purchase (if included software is important to you).

If you're a Windows user and having a mobile bus-powered device is important to you, you're going to have to stick with a USB interface, because to my knowledge there just aren't any Windows-based laptops shipping with full (powered 6-pin) firewire (ieee 1394). Laptops with onboard firewire are actually 4-pin (non-powered), and using a PCMCIA with 6-pin ports won't work because they won't be powered, despite having 6 pins, because the PCMCIA standard doesn't generate provide enough electricity to the device for bus power. PCMCIA cards which DO advertise "bus power" apparently are all ones which feature a DC input with which the bus can be powered. However, carrying around an external power supply defeats the purpose of bus-powering.


ALTERNATIVES
It's still surprisingly difficult USB Audio Interfaces which meet the following qualifications:
- Bus-powered
- Phantom Power
- Decent preamps with TWO XLR inputs
- 24-bit audio recording (and playback, for that matter--ahem, Sound Devices MobilePre)

So you can see why the US-122L is alluring. Don't do it. Below are alternative devices which meet the above qualifications:
----
- Edirol UA-25 . Right now this is the US-122L's main competitor. Replaces UA-20 like Tascam's US-122L replaces US-122. 40dB gain. Can't find really usable noise specs. Edirol claims -96dB at max gain with a 1 k ohm terminated dummy plug, but does not state if preamps are involved in that figure. Renown for the stable drivers. Tech support seems worthwhile from feedback I've read.
----
- Alesis iO-2. It is similar to the US-122L in every way except that it is 24-bit/48KHz and probably USB 1, although USB 1 shouldn't be seen as a negative. Considering the US-122L's 96KHz may not even be full duplex, and also considering the added bandwidth, ask yourself how much more useful is 96KHz than having better drivers and a better company (I can't speak to Alesis yet, but I feel anything would be better than dealing with Tascam).
----
- M-Audio Fast Track (FastTrack?) Pro. Can't find usable noise specs anywhere--not even in the manual. That should tell you a lot about this device. Has 40dB preamp gain. I've seen a lot of complaints about driver problems but the people like the hardware. Your mileage may vary.
----
- Lexicon Lambda. UPDATE: I decided on, and purchased the Lambda. Here is my review of it. Original text continues...This has become the most interesting card out there so far to me. Especially considering it used to be quite expensive and they've lowered the price about a third from introduction, making what once was out of reach for some now compete closely with what is (the US-122L) more of an entry-level device for those who care about sound. The Lambda seems so far from my research to have the quietest preamps from a USB-powered device. It also has the highest gain I can find with USB power.

What's super-cool is that Lexicon has the confidence and respect for us to post noise floor ratings at maximum gain, not minimum gain like most of the other guys. I'm not an expert, but it's obvious to me that mfr's post noise ratings at "min gain" because that's what looks best on paper, or is what's best for them, not you. Noise rating at maximum gain is what I'm concerned about.

UPDATE: Lexicon has some typos in its literature. I've emailed them and gotten word back from their engineers thanks to a helpful tech support agent in Salt Lake City. This review will now be updated with correct info. (By the way, after the Tascam experience I recommend emailing an audio interface's tech support with a non-FAQ question before purchase, to see how they behave.) Lexicon declares the Lambda preamps to have an EIN of -115dB A-weighted at 44dB ( 50dB gain specs you may find in literature are typos; also, an engineer was kind enough to test a production unit himself for me and got -117 and -118 dB respectively, A-weighted, on the L and R channels, which is even better than their claimed but intended -115dBA@44dB spec). Often, you will find preamps with EIN's of about -127dbA, but unless it's otherwise mentioned, you have to assume these are measured at 0 gain (terminated with a 150-ohm or so resistor dummy plug). You have to then add the max gain in dB to figure out what the EIN (Equivalent Input Noise) level (or noise floor) will be at max gain. For instance the E-Mu 0404USB which declares -127dBA EIN, but doesn't say at what gain level (it is 0dB--confirmed by their tech support, who actually showed me this trick). The 0404USB has 60dB max gain, so just add the 60dB gain to the -127dBA noise at min gain: -127 60 = -67 dBA noise floor at max gain of 60dB. To do a fair comparison with the Lexicon Lambda, take the E-mu's -127dbA at 0dB gain, and add 44dB gain (the Lambda's max gain), and the Emu 0404USB has -83dBA gain at 44dB gain (-127 44 = 83), the same level where Lexicon Lambda's peaks out... so Lexicon Lambda is actually much quieter at the same amount of gain! Now that doesn't mean the 0404USB is noisy. But the Lexicon is a lot quieter. The 0404USB has more gain, but I don't see the advantage in a noisier signal that's just amplified more (it probably has better A/D converters, though, for people who will use outboard preamps and line-level signals in an environment where external wall warts aren't an issue, and almost certainly has superior D/A converters). Despite having an outboard power supply, the 0404USB's preamps are still (from what I can conjure on paper) not as clean as the bus-powered Lambda's. Switch up to Lexicon's (externally powered) Omega, and you get an impressive -120dBA at max gain of 51dB, or -118dB unweighted (it's cool that Lexicon publishes an unweighted max gain preamp noise rating, isn't it?). So the Omega would be -127dBA at the Lambda's 44dB peak gain, or 15dB quieter than the already-quiet Lambda.

I formerly stated in this review that E-mu's 0404USB had roughly the same noise specs as the M-Audio DMP3 stereo preamp--I take that back now. The M-Audio DMP3 has an EIN of -128dBm (I don't know what the "m" in dBm stands for--apparently "milliwatt", but I don't know how it compares to dB and dBA) at its MAX gain of 66dB. Strictly by E-Mu's own numbers, the 0404USB should have an EIN of -67dBA at is max gain of 60dB, and if could even go up to 66dB gain,that would make it -61dBA at that level.

Preamps with gain in the 40's dB is "medium" in terms of gain boost for a dedicated preamp, but it's high for a USB bus-powered device. Firewire bus-powered devices, however, often go into the 60's, which is about as high as you'll see any dedicated preamp gain. Tascam won't actually SAY what their gain ratings are for the US-1xx family (because they don't want you to know, because it's so low), but methinks it's around 35dB or so, considered a "low gain" preamp, as is so common on underpowered bus-powered devices.

UPDATE: Tascam doesn't tell you straightforwardly what the gain is on their US-122L and US-122 preamps, so you have to figure it out from their cryptic specs in the manual: "Nominal input levels: MIC IN L and R (XLR balanced): -54 dBu (TRIM=max) to -16 dBu (TRIM=min)". What? Okay, I finally deciphered it: take the -16dBu at TRIM=min (which, somewhat counterintuively or cryptically, is max gain), and subtract from it -54dBu at "TRIM=max" (minimum gain)... Comes out to 38dB of difference between minimum and maximum gain (trim being gain attenuation; apparently the preamps of the US-122(L) are always at max gain, and just get resisted or trimmed down by the knobs). This is validated by the next spec, "MIC/LINE-GUITAR L and R in", the two values being -34dBu to 4dBu. Again, note 38dB of difference. So 38dB gain. As far as noise floor, they don't give you even cryptic but decipherable information. I never found signal-to-noise ratios useful, as boxes with significantly different noise floors often have similar SNR's (usually 100 to 107 dB).

In the real world, I measured the noise floor of the US-122 preamps in the at -72 to -69.5 dB (unweighted) with phantom power off, and -70 to -67 dB with phantom power on, at max gain ( 38dB). (This is unterminated, i.e. with no connector, and I don't know how a terminated dummy connector would affect things.) This might look bad but it's actually pretty good for human ears, but if you turn the volume up, you can hear the white noise, albeit it is quiet. I took a max-gain recording of ambient room noise, then US-122 self-noise (no connectors), then generated silence to compare, but I can't provide URL's here.

Some people think high gain will get them away from the noise floor, which isn't true. The preamps' noise comes up as you amplify. However, what proper gain will provide is more "bits" or more information--more of a signal--to be processed into little 1's and 0's. So a recording at proper gain has higher resolution than one at anemic gain, as more of the dynamic range is used, or fewer of the bits in the data stream are wasted. When all things are said and done, including running that anemic signal through multiple chains, then yes then a recording with insufficient gain can translate into more noise. But honestly it's not as much of a problem as people would make it out to be; far more noise is coming from the microphones themselves.

That said, the US-122 preamps are almost universally considered underpowered, even for high-output condenser mics. People who sing pretty loudly won't have a problem (esp w/condensers), but spoken word and ambient sound, especially with dynamic mics, works only thanks to low-noise electronics of today. Most situations will have the gain maxed all the time. Better options are available. Still, a worst-case recording in the US-1xx family will sound extremely sophisticated when contrasted with the kind of gear most of us had in high school. It's not any noise or lack of gain which is my problem with the US-122.

Preamps sporting max gain in the 60's range are considered "high gain", suitable for ribbon mics. So with a claimed 44dB max gain (not 50dB erroneously in some of their literature), the Lexicon Lambda is medium gain for a standalone preamp, but the highest I've seen on a USB-powered brick (please comment if incorrect), and it is the quietest usb-powered one I've been able to verify in my travels.

There are quieter bricks which offer more gain than the Lexicon Lambda, but you have to go with an external power supply (Lexicon Omega) or Firewire if you want bus power (having a hard time finding an verifiable example). There is a lot of bad information out there, and some of it is put out accidentally by the manufacturer... I mentioned Lexicon above, now I'm going to mention Focusrite with its Saffire and Saffire LE. Literature on the net claims an EIN of -120dB at 60dB gain--seems very impressive at first. But look closer. The mic gain range is actually 13dB to 60dB. The thing only has 47 dB of gain--not 60 decibels as implied by specs on resellers' pages! That -120dB EIN figure is "measured" at 60dB gain, but the signal is only being boosted by 47db because they start measuring from nothing at 13dB somehow. The 47dB gain is confirmed the same way as with the US-122 when you look at "instrument gain" now for the Saffire: exactly the same (13dB to 60dB). So compared to the bus-powered firewire Saffire, at 44dB gain (the max gain of the Lexicon Lambda) the Saffire's EIN is -117dB, or identical to the Lexicon Lambda... interesting. Now these are useful figures.

Also remember that bus-powerable Firewire drivers are going to tend towards catering to Macintoshes. And even some newer Mac laptops don't have powered firewire on them, part of Mac's new subtle de-emphasis of Firewire. If firewire ever dies without a superior replacement, it'll be because of all those 4-pin laptops providing conflicting standards of what firewire really is (Thanks Sony, for ruining another great technology for all of us!).
----
- To be fair, I'll mention the Digidesign Mbox 2, which costs more than 3 times twice as much. It is assumed that it was redesigned due to the poor & noisy entry-level Focusrite preamps in the original, now upgraded but still reportedly not what you'd expect (beware of blind-faith Focusrite and Digidesign fans, too, whose ears don't hear past their wallet). The redesign is still USB 1, but a frank and revealing letter (available on the internet) from Digidesign's senior director of product marketing states it was deliberate and that USB 1 is actually better for streaming audio and has lower latencies in most cases. I can't confirm that but I would say that choosing an audio device based on its USB speed is going down the wrong road unless you have a critical USB device which refuses to get along with any USB1 device, which is rare now. I've found Digidesign's stuff to be exceedingly Mac-oriented and that even in Windows, the hardware is designed to work as an interface to Pro Tools, not necessarily to other software(!). It seems to me that Digidesign's support of other audio recording software is at best an afterthought, and at worst nonfunctional--or in my estimation, even a lie (I won't go into that further here but I will never purchase Digidesign again, either). I also find Pro Tools to be extremely unintuitive and complicated for most people using 2-input devices, who will get more done using far more intuitive software like Adobe Audition. With the Mbox, much of the purchase price is for Pro Tools, and you have to have the device plugged in for Pro Tools to work. But you can get a free version of it with 8 tracks anyway straight from Digidesign (which I exhort you to download and try to record on before buying an Mbox), so it doesn't make sense to me. If I were going to spend that kind of money on an interface, I'd consider a MOTU UltraLite for about fifty bucks more, which is a serious professional bus-powerable device at half the price of a Metric Halo ULN-2, but has annoying AC-in, not DC if you're not bus-powering (it would be better to get a PCMCIA firewire card with a DC in and battery-power it), and whose drivers and compatibility lean heavily towards the Mac side.


There are a number of other 'similar' devices which don't meet my criteria above--such as having only 16-bit recording, or (more commonly) featuring only one XLR input. This should narrow down the research for you in your search. If you find another qualifying product, please record it as a comment.


Footnote: I've found a Windows-based laptop/notebook series which has full bus power 6-pin firewire: the Acer Travelmate series (verify before you buy). So it's not a licensing restriction issue, but probably mfr's not wanting to pay the whopping one dollar worth of royalties to provide a "real" firewire port. There may be other reasons (please comment if you know) but it's obviously not a battery issue. Unfortunately, the Travelmate series is very expensive (around $1700) and usually have small 12-inch screen computers, which will have smaller batteries and thus will be less-suitable for bus-powering anyway. Ironically, it's the larger laptops with bigger batteries are more suitable for 6-pin firewire. For that kind of money, it would be worth looking at Macs anyway, since they can now run or dual-boot Windows (just no right-click button!).

Recommended: No

Amount Paid (US$): n/a