Qudelix-5K Performance Testing

We saw before how the Qudelix-5K uses the Qualcomm QCC512x SoC, with the company doing the engineering work itself to get LDAC up and running on it. There is native 32-bit audio processing courtesy the dual core Kalimba DSP, compared to the 24-bit single core Kalimba DSP of the Qualcomm CSR8675 used in the FiiO BTR5 we saw recently. This variant also prioritizes Bluetooth A2DP as the source even when the 5K is connected to another source in wired mode and allows for the 10-band EQ we saw before. LDAC on the QCC512x allows full 96 kHz/24-bit playback at 330/660/990 kbps bitrates, but know that most Android phones are set to upsample 44 kHz playback to 96 kHz, so you may want to manually set it to 44 kHz for better battery life without unnecessary upsampling.


We also saw how the Qudelix-5K supports 3.5 mm single-ended output at a maximum of 2 Vrms out and 4 Vrms with the 2.5 mm balanced output. Combined with the app, this allows for good matching with power-hungry headphones of higher impedance than you would generally think to play off the 5K, as well as high sensitivity IEMs you need to be careful about when turning up the power. Battery life is also influenced by the operating mode and set volume/gain, so there is little reason to push the Qudelix-5K unless needed. The ESS Sabre ES9218P chips work well even in USB mode, where you get 96 kHz/24-bit audio streaming. This is lossless in USB mode compared to the lossy LDAC codec; however, I will say that above 44 kHz with LDAC barely felt any different from USB mode, and Qudelix itself says that LDAC 96kHz/24-bit on 660 kbps or higher will be no different than USB mode in all practical scenarios. I have to say that I agree and would even go further by not even differentiating between wireless and wired performance with the Qudelix-5 since even the wired and wireless frequency measurements were within error margins on the same output. This is a testament to the Qudelix-5K in more ways than one since you really don't sacrifice much going the wireless route with LDAC, which no doubt will please potential customers.


To test the audio performance, I have randomly chosen a set of high-end IEMs with balanced and single-ended cables to pair with the Qudelix-5K Sparrow and my laptop. Testing was done similar to any other IEMs, which you can read up on in an example review. The earphones used here are extremely well-tuned to begin with, with lots of active cross-over, which helps easily discern any output/drive issues, and I know exactly how it should sound on my reference soundcard as well. The two graphs above first compare the output from the 3.5 mm and 2.5 mm balanced outputs on the EarMen Sparrow, which in turn was connected to my laptop. There are no first-party ASIO USB drivers, so choosing the Qudelix-5K USB device as the output is all you can really do, in addition to player-specific tuning/EQ. There are obvious improvements in comparison to the laptop's own Realtek ALC3266 chipset throughout the frequency range. Gone is the boosted bass that loses clarity on a whim, whatever that mids response is with poor imaging and soundstage, and gone also is thankfully the ridiculously bad treble roll-off that makes my ears ring at times. The Qudelix-5K cleans it all up and provides for a fantastic listening experience. Note that the 3.5 mm output shows a tiny boost to the mid-bass at the loss of some range in the mids, and this is also why I preferred the 2.5 mm output, which was truer to design in retaining the original signal integrity and just is an inherently better experience. This difference is even more obvious when comparing the 3.5 mm output on the Qudelix-5K to the 3.5 mm output on my reference soundcard, with a further boosted bass and sub-bass alike. There is also slightly more variability in the highs, but those could be measurement artifacts since I did not really perceive it in my own listening.

What this means is that the 3.5 mm output can potentially add the tiniest bit of warmth to the bass response of your IEMs or headphones, so keep that in mind if a pro or con to you. For highly neutral devices, such as the Audeze Euclid or Etymotic Evo, I did not even notice a change to where I also suspect some measurement artifacts at play. At the same time, the 2.5 mm output is truer to the actual signal and also provides more power. I personally would use it far more, but also lament it being used over 4.4 mm knowing that the 2.5 mm output is getting more uncommon daily. It does mean the Qudelix-5K will sip more battery from your phone, and the USB Type-C port without a passthrough means the phone cannot be charged at the same time.

Speaking of which, Qudelix has a dedicated page for battery life depending on how the 5K used, and my own experiences match it quite well. The battery log in the app is neat for checking battery usage over time and extrapolating it further, and getting a single cycle of use on par with, if not better than, high-end TWS earphones does add more context for a direct comparison with the competition. The integrated 500 mAh battery does not support quick charging, and the supported 5 V rail maximum means this is a 2.5 Wh battery. USB 2.0 charging at a maximum of 500 mA theoretically has charging fully take 1 hour, though it is slightly longer realistically. I know it is a side note, but the microphones aboard are really not what you want to get the Qudelix-5K for—get something better, in-line or completely separate, since the integrated microphones are mediocre and best left as a last-minute resort.