Hifi Forum TDA-1541A DAC Review

[Doodski]

You are expressing here one of the major doxas of audiophilia that appeared in the 1970s.

This "law" requires that the "musicality" of an electronic device be inversely proportional to its measured performance.

The background ideological on which was based, and still is, this doxa was that the race for performance of Japanese manufacturers had given analytical devices, dry that diverted attention from music...
Here we go again... but with Chinese

erm... Even tossing around inverse proportionality standards. It reads a bit reaching.

 

[pseudoid]

... Here we go again... but with Chinese

See? I told you so!
'China-metal' was definitely derogatory as it should have been back when most exports from that region was pot-metal.
Same could be said of the very derogatory 'Jap-car' (or a "Japanese rice burner") but again both were era-correct, if you keep it within their context.

Would you find it to be derogatory if someone made a reference to a 'Chinese-made' SMSL/Topping?

 

[Paz]

This is a review and detailed measurements of the "Finished Philips hifi forum tenth anniversary TDA1541 DAC Coaxial input L17-53." It was purchased on ebay by a member for $388.
View attachment 175243

Owner swapped out the fake TDA-1541A for a real one. This was a DAC design created for the tenth anniversary of hifi.net forum (?) and many vendors sell it online.

There are no markings anywhere on the device and only coax input is provided:
View attachment 175244

TDA-1541A Measurements
The DAC supports both 16 bit and 24-bit inputs even though the TDA-1541A only accepts 16 bits. Let's start with that input depth:

View attachment 175245

I reduced the input by 2 dB because the nominal output was 2.5 volt which is higher than 2 volt we are used to. It made a corresponding hit to SINAD as it is completely noise dominated. Distortion consists of 2nd harmonic at -102 dB or so which is very good for the era.

Switching to 24 bit input lowers the noise floor but raises distortion due to improper conversion to 16 bits (likely simple truncation):
View attachment 175246

Using the 16 bit input, the DAC barely escapes our poor category of DACs, ranking 260 out of 260+ DACs tested to date:
View attachment 175247

As noted, the problem is noise:

View attachment 175248

IMD test using 16 bit signal shows the same noise penalty but good distortion levels:

View attachment 175249

Jitter test shows many unwanted tones:

View attachment 175250

I did run the test with 16 bit signal and found similar level of interference.

Filter is rather slow with poor attenuation:

View attachment 175251

I could not my Multitone test even when I tried the 44.1 kHz sample rate one.

Conclusion
I was pleasantly surprised by the low level of the distortion in this TDA-1541A implementation. Everything else, especially noise level, is unremarkable to poor. I see nothing that indicates it would perform better than any modern DAC. Considering how much it costs, I can't think of a reason for such a device to exist other than audiophile myths around older DAC chips "sounding better."

Could I ask if the original codec was used as this does the correct 4 times oversampling algorithm which increases the the frequencies in the noise spectrum up by a factor of four, it also interpolates to produce perceived extra resolution. Could I also ask if the original external digital filter was used because this is also important in terms of audio quality. In fact the DAC itself does not correct jitter on the digital signal. Kind Regards Paz

 

[NTTY]

Hello everyone, sorry this is an old thread, but seeing Amir could not share a multitone test, I thought I'd share this one (16bits undithered) of an old Marantz CD-60 (1989) which runs the couple oversampling filter SAA7220P/B and DAC TDA1541A. It shows 96dB free distorsion. Not too bad.

Other measurements of this player are close to what Amir showed on the first page of this thread.

 

[AdrianusG]

I remember an engineer at Arcam telling me at the time the main contributer to the 'sound quality' of these dac chips was almost certainly in the digital filter these machines used as standard and NOT the 1541A (I was asking about the 'Crown' version which apparently had better low level linearity). Any 'more muscular' sonics I always put down to the analogue output stage used (there used to be a 'beefy, slightly dirty' quality to this generation I subjectively remember). I also remember sub 100khz noise coming out of such players which reviewer Paul Miller used to measure for I remember.

I'd also add that the noise output may play a subliminal role as well as 'we' really do seem to like this aspect, even if it's as low as this one is.

My Micro Seiki CD player is from this era (Marantz CD94/Philips CD960 chassis) but the output stage recommended to be used is an add-on with transformer coupled balanced outs. No doubt the 'sound quality' benefit this player had over its competition back then (late 80's) is actually distortion... Gawd this machine is complex compared to a modern better performing dac...

View attachment 175261

View attachment 175262

View attachment 175263

Donor chassis CD94 below. TDA 1541 can be seen with the green film caps surrounding it as my machine has under the copper cladding...

Beautiful player, i had long time ago the Philips CD960, wish i still did.

 

[NTTY]

I remember an engineer at Arcam telling me at the time the main contributer to the 'sound quality' of these dac chips was almost certainly in the digital filter these machines used as standard and NOT the 1541A (I was asking about the 'Crown' version which apparently had better low level linearity). Any 'more muscular' sonics I always put down to the analogue output stage used (there used to be a 'beefy, slightly dirty' quality to this generation I subjectively remember). I also remember sub 100khz noise coming out of such players which reviewer Paul Miller used to measure for I remember.

I'd also add that the noise output may play a subliminal role as well as 'we' really do seem to like this aspect, even if it's as low as this one is.

Oh, I did not reply before, thanks AdrianusG for quoting DSJR.
I think too, from what I measured in my modest lab, and especially what I can hear, that the oversampling filter plays probably a more important role than the DAC itself. I just fail to demonstrate it (so far…).

I did not mention it before, but the Marantz CD-60 I talked about previously shows a very poor linearity (barely down to -70dB), it is using a standard TDA1541A chip, not the selected crowned versions, and that’s its main flaw. Yet, I could not hear a difference with a more recent Denon SACD player (of much better performances) when these two were playing a good old CD with decent DR from a respectable Japanese mastering studio (I know it sounds old and obsolete to listen to an audio-CD…).

So I recorded the outputs of the Marantz and the Denon with a Motu UltraLite Mk5 audio interface (tested by Amir, by the way) at 24bits/96kHz (to include stop-band content of the two players) and then played them back, using ABX comparator in Foobar, on my main system and using a headset too.
If I was able to easily identify A from B with a p-value of zero, 1) it was on a small detail which required me nearly 20min to find what to listen to (piano reverb more on the left side on the “A“), 2) I was unable to decide which one I preferred in the end.

That puzzled me.

So I decided to rip the CDA on my computer to compare with the two files recorded from the two players (Marantz with TDA1451/SAA7220 and Denon with DSD1792/AL24). Fact is the rip had a piano reverb between the Marantz and the Denon but closer to the Denon stereo’s representation. Other than that, nothing else was different while the three files generated should(?) have been very different.

One will argue that Audio-CD is obsolete low-res anyways and that’s why no difference can be heard. Well, knowing from the test of Archimago that differentiating a 24bit from 16bit dithered file is impossible for me (I got the p-value: 0.5982) as well as for the others, I don’t think it’s the point.

That being said, when I compare the output of the Marantz and Denon with whitenoise, this is what I see:

So I question myself, if the two filters show such a difference and if I can tell who’s who in an ABX test on a tiny difference, why can’t I decide which one is the best?

Then we can talk about speakers in the room and other devices which have a higher influence than the two players. Yes we can, but what about exactly same results (for me) when listening with an HD650, or DT Pro, or Oppo PM3 (yeah, I love this one) through the Motu? Oh yes, maybe I’m simply death, but ma wife and doctor say the opposite

 

[DSJR]

I love the status enhancing feel, physical weight and 'vibe' of the Micro Seiki player, but now worry how much 'nice' distortion the output transformers are putting in... Just thinkl, I suspect the eighty odd quid SMSL SU-1 dac would comfortably out perform it (I may yet buy one of these little dacs to see...)

 

[LTig]

Oh, I did not reply before, thanks AdrianusG for quoting DSJR.
I think too, from what I measured in my modest lab, and especially what I can hear, that the oversampling filter plays probably a more important role than the DAC itself. I just fail to demonstrate it (so far…).

I did not mention it before, but the Marantz CD-60 I talked about previously shows a very poor linearity (barely down to -70dB), it is using a standard TDA1541A chip, not the selected crowned versions, and that’s its main flaw. Yet, I could not hear a difference with a more recent Denon SACD player (of much better performances) when these two were playing a good old CD with decent DR from a respectable Japanese mastering studio (I know it sounds old and obsolete to listen to an audio-CD…).

So I recorded the outputs of the Marantz and the Denon with a Motu UltraLite Mk5 audio interface (tested by Amir, by the way) at 24bits/96kHz (to include stop-band content of the two players) and then played them back, using ABX comparator in Foobar, on my main system and using a headset too.
If I was able to easily identify A from B with a p-value of zero, 1) it was on a small detail which required me nearly 20min to find what to listen to (piano reverb more on the left side on the “A“), 2) I was unable to decide which one I preferred in the end.

That puzzled me.

So I decided to rip the CDA on my computer to compare with the two files recorded from the two players (Marantz with TDA1451/SAA7220 and Denon with DSD1792/AL24). Fact is the rip had a piano reverb between the Marantz and the Denon but closer to the Denon stereo’s representation. Other than that, nothing else was different while the three files generated should(?) have been very different.

One will argue that Audio-CD is obsolete low-res anyways and that’s why no difference can be heard. Well, knowing from the test of Archimago that differentiating a 24bit from 16bit dithered file is impossible for me (I got the p-value: 0.5982) as well as for the others, I don’t think it’s the point.

That being said, when I compare the output of the Marantz and Denon with whitenoise, this is what I see:

View attachment 303094

So I question myself, if the two filters show such a difference and if I can tell who’s who in an ABX test on a tiny difference, why can’t I decide which one is the best?

Then we can talk about speakers in the room and other devices which have a higher influence than the two players. Yes we can, but what about exactly same results (for me) when listening with an HD650, or DT Pro, or Oppo PM3 (yeah, I love this one) through the Motu? Oh yes, maybe I’m simply death, but ma wife and doctor say the opposite

The SACD track is louder. That makes an audible difference when comparing it.

 

[NTTY]

The SACD track is louder. That makes an audible difference when comparing it.

That’s true, and it would make the one playing louder the preferred.

And thanks for your note, that reminds me I forgot to mention that the levels were adjusted between the two players when recording, but there was 0.3dB difference worst case between them, as they both have channels mismatch. It probably helped during ABX testing, but not to tell which one I preferred.

I obviously did not adjust the levels on the previous traces, so I updated the screenshot with new measurements between the two players.

 

[AudioX3]

For grins I just so happened to pull out my 1541a chip based Magnavox cdp 473 and I can tell you my opinion which I was not anticipating. It was awful. I then searched out what others felt and found this ASR review.

It is really easy to describe. It had those tones that a certain segment of nostalgic audiophiles crave that no studio engineer would be striving for. There are no highs or definition (it's musical liquid), it's muddy (tubelike) and one I don't think anyone is after which was the surprise it is strident in the vocals almost like an overdrive easily breaking into harshness. That is what was awful.

I then played through the coax digital out and it sounded very nice as a transport into a sigma based DAC. So mechanically it appeared fine though I could imagine a certain crowd informing me it was my RCA cables or needed rebreakin time for the electronics of say 100 hours.

Experiment conclusion for me, once again I have no idea what it is that the nastalgia based audiophiles hear as high fidelity sound.

 

[Veri]

For grins I just so happened to pull out my 1541a chip based Magnavox cdp 473 and I can tell you my opinion which I was not anticipating. It was awful. I then searched out what others felt and found this ASR review.

It is really easy to describe. It had those tones that a certain segment of nostalgic audiophiles crave that no studio engineer would be striving for. There are no highs or definition (it's musical liquid), it's muddy (tubelike) and one I don't think anyone is after which was the surprise it is strident in the vocals almost like an overdrive easily breaking into harshness. That is what was awful.

To be honest, can't the circuit have some faulty caps or other malfunctioning? The reviews I saw of 'vintage' Philips 1541A DACs can do between 90-96dB SINAD: close to a full 16 bits of resolution. None of what you describe should be so very obvious/apparent in a device that in theory should be reasonably transparent.

I own an Audial DAC with the chip and it sounds, well, fine. Most of the roll-off/"no highs or definition" would stem from intentionally disabling the oversampling chip and feeding it 44kHz sampling data. If no oversampling is happening, you should feed it 88/96kHz or ideally even higher; this to prevent imaging/aliasing in the audible band.

 

[DHT 845]

I owned Philips CD473 as my first CD Player in my life. Was is so bad? At the time, no. Better than some technics for sure.

 

[Ralf Stocker]

I owned Philips CD473 as my first CD Player in my life. Was is so bad? At the time, no. Better than some technics for sure.

There is no audible difference between CD players. Of course, DACs have gotten better over the 40 years. 20dB less THD etc. But it is not audible.

 

[audio_tony]

There is no audible difference between CD players. Of course, DACs have gotten better over the 40 years. 20dB less THD etc. But it is not audible.

That's quite a statement to make.

I still have a Sony CDP-227ESD (dual TDA1541 dacs), a Marantz CD17, a Cambridge 640C, Aiwa XC300 (and a few others).

They all sound different, always in the top end of the audio range.

Additionally, modern delta sigma DACs sound better than all the above CD players.

 

[DHT 845]

There is no audible difference between CD players.

You must be joking, the differences between CD players in sound were and still are significant, which results not only from differences in the chips used in the digital section and various treatments in this section, but primarily from differences in the analog section. Every CD player includes an analog amplification section. The thesis that there are no differences between the sound of CD players is as foolish (or more) as the thesis that there are no differences between the sound of audio amplifiers.

 

[Ralf Stocker]

You must be joking, the differences between CD players in sound were and still are significant, which results not only from differences in the chips used in the digital section and various treatments in this section, but primarily from differences in the analog section. Every CD player includes an analog amplification section. The thesis that there are no differences between the sound of CD players is as foolish (or more) as the thesis that there are no differences between the sound of audio amplifiers.

The frequency response is linear +-0.5dB. Noise is at -96dB. THD at 0.0x%. This is all beyond the limits of perception since 1980.

 

[Haskil]

You must be joking, the differences between CD players in sound were and still are significant, which results not only from differences in the chips used in the digital section and various treatments in this section, but primarily from differences in the analog section. Every CD player includes an analog amplification section. The thesis that there are no differences between the sound of CD players is as foolish (or more) as the thesis that there are no differences between the sound of audio amplifiers.

This thesis does not exist... We must rather understand that there are no audible differences if all the measurements show that the defects are all so weak that they are in the inaudible domain...

And for amps, of course, that they are used with speakers that do not require more power than they can provide.

Now, a tube amplifier may have more than one measurement that shows it has defects that are in the audible range... but what is a defect to one person may be a quality to someone else. other...

 

[DHT 845]

The frequency response is linear +-0.5dB. Noise is at -96dB. THD at 0.0x%. This is all beyond the limits of perception since 1980.

I have no intention of arguing with silly "scientific" reductionism.

 

[voodooless]

I have no intention of arguing with silly "scientific" reductionism.

And we have no intention of arguing with silly anecdotes. So please bring something of more substance.

 

[DHT 845]

And we have no intention of arguing with silly anecdotes. So please bring something of more substance.

Why I should convince anybody to sth. I know from experience? NO WAY.