Random Link: An amplifier or preamplifier will therefore sound only as good as its worst capacitors (or resistors). Metal film resistors should be used, though again, steel by itself can lag dynamics something awful (steel is a much worse conductor than copper or silver), so larger resistors ought to be avoided where possible, and small SMD resistors are always better than long-tail radial ones. Audio-grade electrolytic capacitors should be used (Nichicon preferred) where electrolytics cannot be avoided, and silver mica, polypropylene and teflon or even dry tantalium where affordable.
Texas Instruments TPA3316D-2
A class D amplifier, this one is tricky. It can sound passably good, but this depends a lot on output (and perhaps input?) filtering. Filters have to be optimised for a particular speaker impedance rating. An amp board configured for 4-ohm speakers will sound fogged with 8-ohm speakers, this also depends on a particular driver's impedance curve, impedance rising steeply in treble usually makes speakers fatiguing. Fatigue is another issue with this amp chip, even 4-ohm speakers can get stressing/tiring after a while. It's a subtle effect, but usually the more tired the listener is, the worse it gets.
Class D. Highly sensitive to interference, very sweet sound for a class D amp. Sparkling treble, especially after the TPA3116, which is darkish. It must be powered either by a very filtered power supply or a battery (12V to 14V DC). Most cheap boards on sale have filters optimised for 4-ohm apparently, but it also sounds good with 8. Yamaha recommends 1-microfarad bypass capacitors on input, which can also boost bass in addition to cleaning line noise. Apparently those capacitors also affect character; dry tantalium added a lot of bass as an example, but also sounded a bit harsh, giving that class D fatigue similar to TPA3116. Silver mica or polypropylene might be better? The sneaky suspicion is that shielding speaker wire might reduce interference noise which it absolutely loves to inject in bass/low midrange, even in a Bluetooth amp with no line input.
Very sweet-sounding chip, approaching class D transient/drum detail without the fatigue. Possibly one of the best mid-power class AB chips. Its only flaw is it gets a bit anemic (not enough bass/low midrange) without enough power. Very nice treble. Sounds absolutely great with 1-microfarad silver mica bypass capacitors on input and Nichicon electrolytics.
Radio Company of Japan JRC4558
As the name states, and as one might suspect, this chip was created for cheap transistor radios. It sure sounds the part; forward midrange/bass, next to no treble extension, slow slew rate. Something curious about it is that it does image treble extension with DSD (there's space definition) but not as much with PCM output, not just CDs but also 96/24 and even 192/24 PCM. Talk about PCM fake reverb.
Guitar effect tweakers/makers call it a "warm" opamp. It might help electric guitars which have a practical midrange output (your average guitar amp cuts off around 5000-6000 Hz), and it does exaggerate bass/midrange detail, but at the cost of speed and treble. Great in a distortion pedal for sure.
Dynamics aren't quite as good on this one, and again, it tends to sound like an old radio when playing anything hi-fi (like DSD). Curiously, vinyl doesn't sound as bad through this chip in a preamp though. A bit darkish but passable.
Asian manufacturers love sticking it into amp boards, like the TA2020 class D little amps. There it tends to sound a bit cold/anemic. They also stuff it into anything from TDA7265 and TDA7294 to triode valve amps. Which is daft as a good valve amp's main value is quick dynamics and treble extension, which the JRC4558 happily mangles.
One reason for its popularity is that it "dresses up" low-res PCM-based formats like CD audio and MP3/AAC files, hiding the ugliness of slow transients and lack of treble detail by slowing it all down into a crawl (this is a 1.6-2V/usec chip). The drawback is of course that this backfires on vinyl and DSD and tape, which have much better treble/dynamics and real space separation/imaging.
So, in summary, warmish/slow (yet punchy) midrange and bass, slow and poor treble/space imaging. It might be good on bass/midrange drivers, and be combined with a faster opamp (OPA2134?) in a tweeter preamp/amp in a biamplified design. It's a very '60s-coloured sound though.
This is one of the nicest opamps out there, supposedly slightly worse than the OPA2132. It's not the cheapest design, but thanks to its faster slew rate (20 V/usec) it is quite suitable for music. In a good setup with fast, high-quality capacitors and metal film resistors it sounds very lively, almost 3D in imaging, while also making drums big and having a kind of a mellow ambience presence. Cymbals might be slightly sluggish compared to no preamp. Great in a headphone amp, especially with older 600-ohm headphones like AKG K-140 which sound very realistic even in a simple CMoy setup (just give them a good shielded cable). This is an obviously coloured amp, but in a pleasant way, with even harmonics giving it a cosy reaction/ambience, its slew rate still giving it enough treble/dynamics/space definition.
Similar to OPA2134, a slightly coloured, yellowy-warm sound. 20V/microsecond. Less candy-coloured than the OPA2134, more on the slower side as well. Some people call this a "neutral" opamp, well it's still coloured, only less so than slower opamps. It just isn't a "fun-ride" opamp like OPA2134. Cymbals can be less realistic, lacking sparkle compared to unamplified or a faster opamp, somewhat more towards "copper/brass" fundamental tone rather than "trebley sparkle". Overall its character is "stable", It can add some 3D definition that might be lacking without it or on a more neutral opamp to records that lack it (such as those with '80s samplers). Drums are a bit on the slower side, but weighty and defined. Bass can be a bit heavyish. Everything gets somewhat more punchy and strings and synthstrings have a more palpable, textured presence as a plus. Guitars are fine, a bit slow and lacking excitement, but it transmits the roar/distortion texture. A bit mellow/thick in the upper midrange, a bit heavy/stiff in bass. In author's opinion, LM6172 might be a better choice for a preamp because of its speed and transparency. LM4562 might be a good replacement for the NE5532, it's still vaguely neutral/analytical and a bit slowish/static; it is "contoured" in a word. Some say it's "vaguely cold"; it isn't, it's more like "slightly static" at times, especially in bass.
It's that '80s CD player sound. Glassy, nothing really stands out. No impact. Slow, but not as slow as to kill space altogether like JRC4558 does. "Dull" is a good description, some people call it "neutral". It used to be stuffed a lot in SSL consoles and Adam Audio monitor speakers. Also some TV sets and VCRs. It can be harsh with certain DACs (48 KHz and less sampling rates; it's mostly fine with 96 KHz and up) where it creates an ugly odd-harmonic treble fog. Too boring/slow for fast modern speaker drivers. IMHO it's one of the first candidates for replacement. "Mediocre" is another good description. It might be useful in monitor speakers to create a lively mix (rule of the opposites), but it's not good enough for any device that is meant to play lively/realistic. Drums tend towards "splats" and cymbals lack sparkle and metallicity, sounding more like tin midrange (or misaligned-head tape deck). Everything like sizzling analogue synth strings/pads tends to be mellowed down with a dullish temper. It can hide some flaws/low-res PCM harshness in the same way JRC4558 does, by not being detailed. Funny bit is, JRC4558 sounds more interesting thanks to its thick, contrasting (if slowish) midrange/bass. Replace with LM6172 for transparency or some of the OPA series for colour. Don't bother using it in a new design unless it's dull on purpose, like mixing/mastering gear.