Topping E50 Review (Balanced DAC)

[TheWalkman]

Can someone share their thoughts on the three filter options?

These are all filtering frequencies far above what my tin ears can hear any more.

Are these designed to protect someone‘s super tweeters from working too hard?

What am I missing?

 

[linuxfan]

Hooked a SONY Blu-Ray via coax to the E50 and connected the DAC to a Marantz PM700N via RCA cables. Played a hybrid SACD. The E50 display reads 44.1 PCM. What does that mean? Am I right this should be reading the 24/96 SACD layer and outputting a hi-res signal?

Atomicdog, the only way to get raw DSD data output from an optical disk player is via DSD-over-HDMI.
First you need an SACD player or Blu-ray player which supports this feature. There's an unofficial list of compatible players here -
https://www.avsforum.com/threads/dsd-over-hdmi-players.1155206/

Then you also need a particular type of HDMI de-embedder device, such as this -
https://www.ebay.com/itm/353347272941
You can't just use a generic HDMI extractor.

More details in this thread -
https://www.audiosciencereview.com/...mi-audio-to-usb-conversion.11481/#post-328929

 

[raif71]

Hooked a SONY Blu-Ray via coax to the E50 and connected the DAC to a Marantz PM700N via RCA cables. Played a hybrid SACD. The E50 display reads 44.1 PCM. What does that mean? Am I right this should be reading the 24/96 SACD layer and outputting a hi-res signal?

Yeah, I too got my Gustard X16 to show 44.1khz when playing sacd using a Sony Bluray Player however I got the DAC to show DSD by using an hdmi via i2s converter board that connects the bluray via its hdmi out to Gustard X16 i2s input. The thread is shown here.

 

[JaccoW]

It may only output SACD digital audio via HDMI due to the copy protection on coax/optical outputs as @Gradius mentioned. Maybe the 44.1 downsampling is intentional as a SACD copy protection mechanism that retains the ability to output in CD format.

Some (but not all) SACD's have a so called Redbook 16-bit 44.1 kHz layer so regular CD players can play them at normal CD-quality. So that's probably what is going on here.

I asked a similar question a couple of pages ago but yeah virtually zero SACD players will output DSD over Coax/Toslink. That is down to the standard and limits of those technologies and the DRM at the time SACD came out.

To give people an overview here:

• CD: 1.4 Mbps for 2-channel. 16-bit 44.1 kHz
• SACD: 5.6 Mbps for 2-channel 1-bit 2.8224 MHz (2822.4kHz) DSD, more if it is a 6 channel (5.1) recording.
• Coax/S/PDIF: 3.1 Mbps 20-bit 48kHz but can go as high as 24-bit 192 kHz PCM (9.216 Mbps) if the receiver supports it. Theoretically no maximum bandwidth but the S/PDIF standard doesn't support it in most devices.
• TOSLINK: 3.1 Mbps. 20-bit 48 kHz max. Modern TOSLINK can go as high as 125Mbps but is rarely supported. Higher jitter. Based on S/PDIF.
• HDMI 1.0: 36.86 Mbps
• HDMI 2.0: 49.152 Mbps. At least 16-bit 44.1 2-channel but up to 8-channel 16/20/24-bit x 32/44.1/48/88.2/96/176.4/192 kHz or 8-channel 1-bit DSD (aka 22.4 Mbps)

It was unfortunately a standard that came into existence when companies were trying to figure out DRM and record companies were very apprehensive about giving consumers access to their Master Tape copies through hi-res audio formats. So every single older SACD player has a Sony approved internal DAC to decode the DSD signal or sends the encrypted data stream over HDMI to a Receiver that can safely decode DSD into analog before it leaves the device.

There are a couple of Blu-Ray players that can extract it to USB through a commandline tool but other than that you're out of luck. And I feel your pain.

 

[yanm]

Can someone share their thoughts on the three filter options?

These are all filtering frequencies far above what my tin ears can hear any more.

Are these designed to protect someone‘s super tweeters from working too hard?

What am I missing?

F1 is the “mathematically correct” reconstruction filter for sampled signals, F3 is similar but minimal phase. F2 seems to be an attempt at a NOS-type of filter. I don't hear any differences between any of those to be honest.

Some reading if interested (I particularly like the second link):

Linear and minimum phase

A never-ending debate is that of linear versus minimum phase filters with proponents of each often deriding the other as unnatural or even unlistenable. We explore the meaning of these terms and ho…

troll-audio.com

How to pick the best filter setting for your DAC

Does your DAC let you choose the output filter? What do they do. How do the different kinds differ from each one? Which is the best? Which is the worst? We'll help you sort it out.

addictedtoaudio.com.au

All in all, I would have preferred if F1 had a lower cutoff frequency, now it is a bit hot and may show some slight aliasing artefacts. From that perspective F3 is better with an attenuation of about 80 dB at 22 kHz (only about 10 dB at 22 kHz for F1). TBH, I don’t fully understand why F1 and F3 do not have the same cutoff frequencies as for example in the the E30 DAC (may be due to difference in the chipsets, AKM for E30 and ESS for E50). Interestingly enough, the default settings are “minimum phase” for the E30 but “linear phase” for the E50.

 

[Gradius]

v5.30, finally:

V5.30 driver for most of TOPPING DACs-TOPPING

 

[JaccoW]

v5.30, finally:

V5.30 driver for most of TOPPING DACs-TOPPING

Any link to the changelog? I'm curious what they changed.

 

[Gradius]

Any link to the changelog? I'm curious what they changed.

I mentioned before more than once I think:

https://www.thesycon.de/usbaudio/TUSBAudio_history.txt

 

[Hee]

Dear Professors.
The name of this part? Or the purchase site?
This is estimated to be a relay. (3710 2007U)
I want to get it.

 

[audiofun]

Dear Professors.
The name of this part? Or the purchase site?
This is estimated to be a relay. (3710 2007U)
I want to get it.
View attachment 181900

SGMicro SGM3710
cheap dual channel analog switch from a Chinese IC producer.

 

[JaccoW]

Some (but not all) SACD's have a so called Redbook 16-bit 44.1 kHz layer so regular CD players can play them at normal CD-quality. So that's probably what is going on here.
[...]
To give people an overview here:

• CD: 1.4 Mbps for 2-channel. 16-bit 44.1 kHz
• SACD: 5.6 Mbps for 2-channel 1-bit 2.8224 MHz (2822.4kHz) DSD, more if it is a 6 channel (5.1) recording.
• Coax/S/PDIF: 3.1 Mbps 20-bit 48kHz but can go as high as 24-bit 192 kHz PCM (9.216 Mbps) if the receiver supports it. Theoretically no maximum bandwidth but the S/PDIF standard doesn't support it in most devices.
• TOSLINK: 3.1 Mbps. 20-bit 48 kHz max. Modern TOSLINK can go as high as 125Mbps but is rarely supported. Higher jitter. Based on S/PDIF.
• HDMI 1.0: 36.86 Mbps
• HDMI 2.0: 49.152 Mbps. At least 16-bit 44.1 2-channel but up to 8-channel 16/20/24-bit x 32/44.1/48/88.2/96/176.4/192 kHz or 8-channel 1-bit DSD (aka 22.4 Mbps)

To quote my own post and to add some more things I've since learned about SACD playback on the E50.

Traditional S/PDIF coax or TOSLINK might be limited but on the E50 these two connections support:

• 24-bit 192kHz PCM
• Dop DSD64
  • Packages the RAW DSD stream in a (24-bit 176.4 kHz) PCM package that gets unpacked back into DSD in the DAC.

SACD is DSD64 (!), so if you were able send your DSD stream over HDMI you could use a DSD De-embedder to send it over Coax to the Topping E50. Several people have reported succes with a Topping D90 but apparently it depends on the HDMI output of your player.

More info in this thread: Converting a Blu-ray/SACD player into a digital DSD transport

Follow the links to the box on eBay. No link yet since I'm trying to figure out which one it is since they range from $40 to $110 and I'm not sure which one does what.

To be continued.

 

[bogi]

Not really. MQA is not lossy in the same way that MP3 is lossy, and it contains corrections in the analog domain.

Of course, MP3 and MQA are lossy in different ways.

1) Folding/unfolding process of MQA:

HQ Player

audiophilestyle.com

I found curious that people who deny to perform software upsampling are fans of MQA down and upsampling DSP.

2) Ultrasonic noise added by MQA filters - see Archimago's measurements comparing Topping D90SE built-in filter and MQA filters:
"As you can see, I've created a 96kHz file with content starting at 17.5kHz rolling-off by 45kHz. Obviously music should not have high levels out to 40+kHz, but you never know and I'm trying to be illustrative here. The shape is intentional so we can identify the pattern across the frequencies."

It is easier to trust propaganda words than to try to understand things.

 

[bogi]

Dop DSD64

• Converts the RAW DSD stream to 24-bit 176.4 kHz PCM

DoP DSD64 does not convert DSD stream to 24 / 176.4. DoP is used as a way how to transfer DSD bitstream through PCM only capable interface into DAC or other digital device. The PCM envelope is taken out on DAC side and original DSD bitstream is processed by DAC in the same way as if the same bitstream would be sent through DSD capable interface.

 

[JaccoW]

DoP DSD64 does not convert DSD stream to 24 / 176.4. DoP is used as a way how to transfer DSD bitstream through PCM only capable interface into DAC or other digital device. The PCM envelope is taken out on DAC side and original DSD bitstream is processed by DAC in the same way as if the same bitstream would be sent through DSD capable interface.

Changed it to

• Converts the RAW DSD stream to a 24-bit 176.4 kHz PCM package that gets converted back into DSD in the DAC.

 

[bogi]

No, that's too incorrect. There is no DSD -> PCM -> DSD conversion in the case of DoP. You would need to understand the difference between binary coded PCM samples and unary coded DSD bitstream (every bit in DSD bitstream has the same weight). DoP is only a way how to transfer the original DSD bitstream, without any processing done on that bitstream (no change in transferred bits), to other device. Imagine DoP only as a PCM envelope to transfer unmodified DSD bits. Those DSD bits in PCM envelope are not PCM - you could not play them to get real audio content. DSD to PCM conversion is quite complicated mathematical process involving so called decimation and is lossy.

 

[Grooved]

...
Traditional S/PDIF coax or TOSLINK might be limited but on the E50 these two connections support:

• 24-bit 192kHz PCM

I imagine it's for oldest devices, because even the motherboard from one of my computers (10 years old) can send 24/192 over Coax and Optical (confirmed by the DAC linked to it).
It also gives me idea to check an old DVD-Audio player that I still have, I only used it with it's analog outputs, will see what there's on the Optical output with a 24/192 disc as source.

 

[JaccoW]

No, that's too incorrect. There is no DSD -> PCM -> DSD conversion in the case of DoP. You would need to understand the difference between binary coded PCM samples and unary coded DSD bitstream (every bit in DSD bitstream has the same weight). DoP is only a way how to transfer the original DSD bitstream, without any processing done on that bitstream (no change in transferred bits), to other device. Imagine DoP only as a PCM envelope to transfer unmodified DSD bits. Those DSD bits in PCM envelope are not PCM - you could not play them to get real audio content. DSD to PCM conversion is quite complicated mathematical process involving so called decimation and is lossy.

So as I understand it now:

"(DSD Over PCM) DoP is a method for transporting DSD audio over USB ports that do not have a DSD driver. Each consecutive set of 16 DSD bits is stored as PCM bits in the lower 16 bits of a 176.4/24 sampling rate. An 8-bit DoP header is added to each sample. In order to reassemble the PCM bits into a continuous DSD stream, the DAC at the playback end must be DoP compliant." (PCmag)

To summarize DSD-guide.com:

The normal DSD64 sample rate of 1-bit at 2.8224MHz has a data rate of 2.8224Mbits/sec. This is equivalent to 16-bit PCM at 176.4kHz but 8-bits are used to mark the PCM package as DSD so 24-bit 176.4kHz. The DAC requires 2x 16-bit streams, to be able to form a stereo stream and as such some compare it to a 16-bit 88.2kHz PCM file, though the differences in technology mean there are some heated opinions on which one is better and if it matters. There are some advantages to a continuous stream over blocks that need to be put together.
There is some built in safety there that if the package gets partially dropped or accidentally interpreted as PCM it will produce a low-volume, high-frequency 88kHz tone that is inaudible and won't harm most hardware. Though most will filter this out.

• Packages the RAW DSD stream in a (24-bit 176.4 kHz) PCM package that gets unpacked (but not converted!) back into DSD in the DAC.

 

[Hee]

SGMicro SGM3710
cheap dual channel analog switch from a Chinese IC producer.

Thank you sir.
But I can't find it on Google.

 

[Blew]

I found curious that people who deny to perform software upsampling are fans of MQA down and upsampling DSP.

I don't claim to be any type of expert here, but as I understand it, upsampling does not correct errors in the time domain like MQA does. Upsampling can have advantages for the filtering process but is not a panacea for reconstruction of 44.1KHz recorded digital audio. My subjective testing in listening to MQA encoded 44.1KHz audio vs straight PCM is MQA definitely sounds better. You could argue that a better DAC would rectify this for the PCM but the point of MQA is to enable this without requiring an expensive DAC with more sophisticated reconstruction filters.

2) Ultrasonic noise added by MQA filters - see Archimago's measurements comparing Topping D90SE built-in filter and MQA filters:
"As you can see, I've created a 96kHz file with content starting at 17.5kHz rolling-off by 45kHz. Obviously music should not have high levels out to 40+kHz, but you never know and I'm trying to be illustrative here. The shape is intentional so we can identify the pattern across the frequencies."

Archimago is known, and freely admits, to being biased against MQA due to the lack of transparency and general business practices. The measurements shown are of audio "content" from 17.5kHz to 45kHz. MQA is not designed to encode this, as it's designed to encode music within the audible spectrum. So this test appears designed to make MQA look bad due to his personal bias against it. I wouldn't place too much importance on it.

It is easier to trust propaganda words than to try to understand things.

There appears to be propaganda on both sides of the argument, as demonstrated above.

 

[Jobblin]

I don't claim to be any type of expert here, but as I understand it, upsampling does not correct errors in the time domain like MQA does. Upsampling can have advantages for the filtering process but is not a panacea for reconstruction of 44.1KHz recorded digital audio. My subjective testing in listening to MQA encoded 44.1KHz audio vs straight PCM is MQA definitely sounds better. You could argue that a better DAC would rectify this for the PCM but the point of MQA is to enable this without requiring an expensive DAC with more sophisticated reconstruction filters.

Archimago is known, and freely admits, to being biased against MQA due to the lack of transparency and general business practices. The measurements shown are of audio "content" from 17.5kHz to 45kHz. MQA is not designed to encode this, as it's designed to encode music within the audible spectrum. So this test appears designed to make MQA look bad due to his personal bias against it. I wouldn't place too much importance on it.

There appears to be propaganda on both sides of the argument, as demonstrated above.

So, MQA is "correcting errors in the time domain", deciding what part of an audio signal constitutes "music" and somehow subjectively sounds "better" (definitely)?
I thought this was Audio Science Review and not Snake Oil Corner?