SMSL DO300 DAC Review

[hearone]

...
I am also very attracted to products using ES9039Pro and ES9038Pro. However, in recent DAC products, the difference in sound quality between high-end and low-end has become smaller, and the performance improvement every year is remarkable. So I haven't reached the point of buying a high-end product yet. I mostly use cheap products such as Topping E30ll and E50.

Yeah, the advantage of the ES9039PRO is that you can get it now for not much more money than the older ES9038PRO in the Toppng/SMSL lines. Agree, it is primary just ups in SINAD and power efficiency. I just ordered the SMSL DO300 for $352. It seems like a very good deal for a high performance/featured DAC.

I just recently entered again the home HF 2-ch market after several years absence. A lot has changed from only low quality lossy streaming options to actually streaming lossless audio without having to buy and mange all the high resolutions files myself. And for $9-12 a month, one has access to millions of tracks-sweet!

I found todays DACs are a surprising relatively mature market. Not so much for streamers or streamer/DACs units. They also over the map and very wide price range. I initially was strongly intersted in the new Eversolo DAC for $860. It was the hype of the day. But slowly while reading tons of ASR reviews, I realized I just need a bit perfect steamer and a good DAC. I ended up getting a WiiM PRO streamer/DO300 DAC combo. Heck, I could have even spent less if I dropped the bal outputs/remote/preamp functions and still got a transparent DAC like the SMSL SU-1 for $80 +$80 for a WiiM Mini. If you are willing to forgo an all-in-one device, for $250 it delivers as good as almost all of the bit perfect streamers used a digital transport with DAC for hundred or even thousands less. Even on this ASR site, I sill scratch my head when I head someone wanting to spend hundreds of $ for a streamer/DAC combo with its measured transparent DAC and then use it as digital transport only and now add another transparent DAC in the hundreds of $ on top of the already expensive steamer/DAC.

Good time to be be in the market.

 

[hearone]

FYI, here is an inquiry on the FW levels. I did not see a link to ask SMSL directly, so inquired with the seller. They responded quickly. The reply did not mentioned anything about a "sound" change, but did acknowledge it was around a BT issue.

I do't use Windows anymore, so when I can borrow a PC, I will just upgrade my DAC's FW to the latest FW. Meanwhile, the unit is working fine. It runs very cool with about 5.3 watts of power in typical use and about 2 watts in standby mode. The external build quality is very good and solid.

Here is the FW inquiry/response, if interested:

Hi, I got the SMSL DO300 DAC SN:421012122xxxx (via Amazon) today. Per a couple of forum entires on AudioScienceReview their seems to be some confusion on the FW level and the update on your site for V1.13. I am showing on my unit is at FW V1.0. They commented the latest FW iimproves the sound considerably. However, I only use a MAC ( MACOS not Windows), so can not update the FW apparently. Can you confirm for DO300 SN:421012xxxx that I need to update to the lasted FW or do I have the latest already? ? And what conditions did the V1.13 address? Thanks xxxx

Dear xxxx, Thank you very much for your contact. Regarding the firmware you mentioned, it is a fix for the issue where the Bluetooth automatically turns on and is recognized on boot. You do not need to upgrade the DO300. Please feel free to use it. Judy(Aoshida)

 

[Iron Man]

The FFT/SINAD measurement is at 1k. It doesn't reflect the performance across complete audio range. It cannot judge if there is low hum caused by PSU or GND, or if there are more harmonics at higher frequency which does happen in the field. A THD+N sweep is necessary in my opinion.

 

[hearone]

Amirm discuss how he does a "synthetic" sweep test using 32 muiltiones. He has a tutorial on UT and discuss how to read his DAC test. Look at 14:40 for this particular discussion.

 

[widemediaphotography]

HI,
I have this DAC and have two questions.

1) What about the behavior PCM filter - Filter off? If I set it the sound is higher and more open, but unfortunately it's very, very distorted. Do you have the same behavior?

2) On the SMSL DO300 support web page there is a firmware version 1.03... but, as already discussed on this forum, there seems to be no way to understand if it is the same one already installed on my DAC and what are the improvements compared to the version previous... Is there a way?

The SMSL support, which answered me by email in half a day, advised me to install a new firmware version only if my DAC has problems because our devices are not all the same...

What about my opinion on this DAC…
In my equipment the improvement is quite evident, now it seems I have removed a cover on the tower speakers. The USB port connection is better than the optical/coaxial one on my PI4 Streamer and Volumio. Hysolid player notebook windows sound better than Raspberry and Volumio. Even with the standard Windows player Windows wins, but DSD files cannot be compared. I think it is a Rasberry hw limit. In any case Volumio is complete, great and stable, it is not easy to find a complete quality alternative free solution like Volumio (its support is also very efficient.

wide

 

[hearone]

1) What about the behavior PCM filter - Filter off? If I set it the sound is higher and more open, but unfortunately it's very, very distorted. Do you have the same behavior?
...

I am by no means an expert in digital filters, just learning myself. But I came to the conclusion that Linear Fast is the overall best for me. Unfortuenly, apparently there is no perfect digital filter- each has a set of tradeoffs. Just the nature of the beast, so to say. And lots of just personal preferences like tube/vinyl. If you cannot hear above 10k or so, some of the other "slow" filters might be an options also. But I would NOT use NOS myself.

Below is an article I read just last night on filters. It has a lot of graphs of some of the effects of different filters including a NOS option. I also looked at the RME ADI-2 FS DAC manual on how it handled the NOS on the DAC and their comments did not seem very encouraging to make me want to use NOS on it either.

My conclusion on why Linear Fast is a good tradeoff is that is does have symmetric pre/post "ring" around the real impulse. This looks bad, of course on a scope, but it seems the "ring" is not audible (or a least at a very low level) and the filter does not cut into the 20-20K audible range and has good cutoff of spurious signals beyond hearing signal being fed into the post DAC electronic chain. Each other filter has its respective trade-off too.

Also consider this from the review of the DO300 by Amirm on which filter to use:

"I would personally use Fast Linear and be done with it. Note that "off" causes a large rise in output level so be careful when comparing that to other settings:"

"- Linear Fast the better one.

how-to-pick-the-best-filter-setting-for-your-dac

 

[widemediaphotography]

Thank you hearone.
I haven't trained my ear well to understand which PCM filter is better among those available and so at the moment I'm running the comparative tests with the default filter set at the DO300 reset. You are absolutely right when you say that there is no perfect digital filter, for the same reason there shouldn't exist, in a digital environment, sw readers that sound different or hardware that, by treating a digital signal, already colors the sound differently in their flat settings. I will try your suggestions

 

[Gerardo61]

Thank you hearone.
I haven't trained my ear well to understand which PCM filter is better among those available and so at the moment I'm running the comparative tests with the default filter set at the DO300 reset. You are absolutely right when you say that there is no perfect digital filter, for the same reason there shouldn't exist, in a digital environment, sw readers that sound different or hardware that, by treating a digital signal, already colors the sound differently in their flat settings. I will try your suggestions

I bought DO300 on March to replace a Wyred4sound dac2dsd.
I believe smsl do300 is a best buy for my set up.
I suggest you the follow:
PCM filter: Linear Fast (linear and neutral Response)
DSD filter: 47 KHz cut off
Volume: 0 dB
That's all
I have a jadis i50 ampli (tube 50+50w A Class) and I use only unbalanced connection but believe me this dac is great

P.S. Without filter the curve drop soon and output Level is +4,5 dB respect any other filters
Enjoy the music
Gerardo

 

[Blashyrkh]

HI,
I have this DAC and have two questions.

1) What about the behavior PCM filter - Filter off? If I set it the sound is higher and more open, but unfortunately it's very, very distorted. Do you have the same behavior?

Filter off is intended to be used with an external upsampling, it has already been pinpointed here somewhere, but don't want to search
It's also stated in the official documentation for 9038 pro (but is the same for 39 pro), here at page 9

It suggests to oversample at least 8x the original sample, so for 44.1khz files you need to set oversampling at 352khz in the player you are using.
This way it will work perfectly without distortion (tested myself)

Measurements without external upsampling doesn't make any sense and don't know why Amir did it

 

[widemediaphotography]

Thanks Blashyrkh, I would have never found it!

 

[widemediaphotography]

It suggests to oversample at least 8x the original sample, so for 44.1khz files you need to set oversampling at 352khz in the player you are using.
This way it will work perfectly without distortion (tested myself)

In your opinion, is this filter better than linear fast? In my opinion, avoid the distortion, the sound the sound look likes more open

 

[Blashyrkh]

In your opinion, is this filter better than linear fast? In my opinion, avoid the distortion, the sound the sound look likes more open

I played a bit at the beginning with filters, but didn't notice many differences so kept linear fast and never changed.

Consider that which should be the best with higher fidelity is the linear fast, for frequency response, ringing pattern, and removal of ultrasonic noise.

Nos (filter off) is always the worst, rolling frequency much before other filters, and having a lot of ultrasonic noise.


Ringing and ultrasonic noise should not be audible anyway, but rolloff in high frequency may be....moreover I don't want to have the oversamples done by software I don't know much about (for sure windows one is quite bad, at least on windows 10)


That said I don't see the need to chose something else than linear fast that is objectively the best filter on paper, since I want fidelity

 

[hearone]

On the setting of Gerardo, I use exactly the same. I did tweaked the DPLL setting to min for the lowest jitter, but how low you can go on this DPPL setting is limited by your equipment. Lower is better.

FWIW, I was curious to so see what RME (ADI-2 FS DAC) said about NOS filter.

From RME Owner's manual V3.1page 57:
"NOS (Non-Oversampling, SuperSlow)
The DAC includes another filter which is called Super Slow in its data sheet. The impulse response looks perfect, but checking the output signal with an Oscilloscope reveals steps that are more typical for so called Non-OverSampling (NOS) devices, so we renamed it NOS within the DAC filter menu. Note that there is no audible distortion, the steps equal high frequency harmonics that are mostly higher than 20 kHz. Please also note that Slow and NOS filters cause much more aliasing into the audio band and out-of-band noise than Sharp filters."

 

[Blashyrkh]

Please also note that Slow and NOS filters cause much more aliasing into the audio band and out-of-band noise than Sharp filters."

but aliasing in audio band means in audible range ? Because I don’t think that’s the case

 

[hearone]

but aliasing in audio band means in audible range ? Because I don’t think that’s the case

Good question. That was just a clip from RME on their NOS filters. I assume you mean aliasing fold back should be only in A-D not D-A? Regarding RME clip, could be German to English translation issue or just some slop in terms? The whole wording is kind of odd.

I did find this on ASR on NOS audible in-band distortions (see below). Not sure if they came to a solid conclusion what they were seeing in scopes. They also mentioned some sloppy in term like imagine vs aliasing. But I could not track the discussion very well. Amirm had some comments also.

If interested, here is the ASR NOS DAC in-band distortion
index.php

We are probably getting off-track on DA300 discussion, however.

 

[Blashyrkh]

We are probably getting off-track on DA300 discussion, however.

We are probably getting off-track on DA300 discussion, however.

don't know, since it's related to filters contained in the DAC, so I will continue

I read the thread, and from my knowledge, it's more or less how do I know things, but let me explain after I underline one important thing:

- Aliasing is what we get from the digital source after the A to D conversion (as you mentioned), it happens if things above the nyquist frequency is thrown in the mix.
That's why a low-pass filter (also called Anti-aliasing filter) is used to remove all the unwanted frequencies.

- Imaging is what is produced by DACs when they have to recostruct a digital source. It's actually the production of an "image" of the original source, but translated into ultrasonic frequencies (that is why are called "ultrasonic images", they are copies of the original signal but played inthe ultrasonic domain)
To remove them, we use a reconstruction filter, that is also called Anti-Image filter.

That said, these are the standard names that should be considered, and
now we are talking about the reconstruction filters, that have to handle the Image problem.


The issue here is that even if these images are in the ultrasonic domain, and we cannot hear them, they are reproduced anyway and can interfere in two ways:
- "strain" on the the speakers/headphones transducers that may struggle when need to play lots of frequencies. For example in speakers we have cones that reproduce different frequencies to have them reproduced the best way possible, think of a tweeter that also needs to reproduces from 5khz to 100khz, even if 100khz cannot be heard, the 5khz-20khz part may be affected by the reproduction of the rest of the bandwidth.
- IMD (Intermodulation Distorsion) that throws distortion into the audible range

This is why DAC use the filters, to lower the ultrasonic signal that may interfere one way or the other with the music reproduction

That said, NOS is the worst case scenario here, since it let everything go through

but there is also the Ringing issue

I won't go further since it's being to much of a post, but let me breifly summarize with a couple of filter's pro and cons:

NOS:
pro: less ringing among all the filters
cons: all the ultrasonic goes through

FAST:
pro: no or few ultrasonic
cons: lots of post and pre ringing

SLOW:
pro: low ringing
cons: some ultrasonic goes through

and obviously there are all other filters in between that are a tradeoff of ultrasonic cut, high frequency cut, ringing and also delay (for minimum filters)

The best one is probably the one that has the minimum of each issue, that in my case was the Linear Fast, but probably in the future I would go with one of minimum slow to avoid delay and pre-ringing, upsamplig a bit the original source to avoid ultrasonic noise enter the audible range

 

[hearone]

There is one ESS filter tha I might try on the DO300- APODIZING. I don't think it allow much ultrasonic (if any) and has no pre-ring. So more natural without the pre-ring? Probably only in that area, for what I have seen the post ring amplitude is twice as high. It may be even more audible the the two small peaks of pre/post rings. No free lunch in physics, as usually.

On the NOS, as you mentioned, I don't want to even try the NOS with all its ultrasonics possible issues. My old school design McIntosh amp probably would not have an issues or instability. But the tweeter damage/damage (even possible coil burn-out) would be a major concern. Botton line: It just seems like too much risk.

I should know more about digital filters, since I had purchased the worlds first CD player with its brick wall filter- the Sony CDP-101 in 1882. But until just recently, I thought the filter "problem" has been solved after 40+years. LOL I was aware generally of the flurry of initial activity in all the over-sampling designs to improve the DA conversion. And later, the BB ladder DACs were the one to get. I never had even heard of Delta/Sigma DACs, but since I had at one time the Super Audio Sony 1 bit CD player, I could make a guess since all these DACs supported DSD. But, silly me, I just assume in 2023 it was perfect by now. And probably my bad, but reading all the comments on DACs reviews, it seems the DAC problem was solved. Probably true for the DACs with good SINAD, but the which filter is more "subjective" it seems, if you have that choice. Gosh, say that word, but in quotes.

Also until I dug a little deeper, I had gotten the impression that all filters were pretty much similar in sound and not any possible ultrasonics affects on the post DAC audio chain. Just chose one you like- that seemed to be the simple message. With the ultrasonic issue, I would want to take that into consideration espically on the NOS filter.

 

[anome]

Sony CDP-101 in 1882

You were early to the game.

 

[hearone]

You were early to the game.

Just a tad...

 

[DeepSpace57]

HI,
I have this DAC and have two questions.

1) What about the behavior PCM filter - Filter off? If I set it the sound is higher and more open, but unfortunately it's very, very distorted. Do you have the same behavior?

Yes i have the same issue, moreover some color options results in broken speaker effect, buzzing