Lossless SD-card player

Status
Not open for further replies.
Jump to Latest
#562
Onno.

Pretty good stuff. THX a lot.

For now Yuri Bakker Piano Prokofiev confuses me. Channel swap? I doesn't sound right. Piano comes from left and right

Misha Fomin Amsterdam Bach Piano sounds right. IMO a pretty good piano recording. Very nice.

Gruppmann Violin Bach sounds also pretty good. Lot's of details you can "see"
the bow on the strings.

I'll go get some more. 😀

Would be nice to know how you'd done it.

Cheers
 
#565
Onnosr said:
Hi Soundcheck,

Your ears are perfect.

Many thanks for your OK comment.

Pse check Yuri Bakker Prokofiev etc again. Now I hope it's ok.

Very Best regards,


Onno
Click to expand...

Yep. It's OK now.

BTW. We're hijacking that thread.

Can we split off this recording discussion and open another thread Dear Moderators?


Anyhow, I would like to give some feedback:

This is Germany calling - Here are our points:

Pays-Bas Douze Points! 😉

(The Europeans around know what I am talking about 😉 )



A little more meat on my last comments: ( I had to streamline the filenames - to get them properly integrated into my system)

01-Onno-Recordings-Doelen_Rotterdam_Jurriaanse_Zaal_Igor_Gruppman_violin_Bach_Chaconne.wav

Very detailed and 3D. Nice.

I had the feeling a bit less sharpness/aggressiveness would be perfect. I don't think that the violin is that sharp. Violins are extremely tricky I know.

Ideas:
How about hardwiring your mikes and using a battery supply for the mikes. Silver cables?
You might clean the contacts and use contact enhancer.

Could also be my system. Lets see. Perhaps others come back with a comment,

02-Onno-Recordings-Paleis_Het_Loo_YPF_concert_2008_Yuri_Bakker_Prokofiev_Sonata6.wav

Now it's better. Though the recording is IMO not coming close to 03 - see next.


03-Onno-Recordings-Concertgebouw_Amsterdam_Kleine_Zaal_Recital_Misha_Fomin_Piano_Busoni_Bach_Chaconne.wav


Sounds perfect to me. That's IMO the winner track from what I heard.

I hear the strings as they're 3D lined up in the room. High tones have a more left sided front attack. Deeper tomes come clearly from the back. Pretty consistent throughout the track.

Feels like being 3 m away and 50 cm above the piano rim.

Great stuff. Cool illusion.

04-Onno-Recordings-NKO_Muziekgebouw_May_25_2009.wav

Lot's of focus on the violin and flute ( fagot?) - IMO too much!

The orchestra in the back gets a bit mixed - separations and presence is not very well .

05-Onno-Recordings- NKO_Rehearsal_Beurs_van_Berlage_February_19_2010_III.wav


Orchestra well positioned in the room. from left to center-back to right.

Great. IMO the 2nd best on my list.

06-Onno-Recordings-Mara_Van_Pommeren_Shostakoviich_prel_3_4_Scriabin_Et.wav

Hmmh. Very dynamic. Great low tones, reverberation and details. Sounds like a nice piano ( I am not an expert 😉 )

However.

Compared to 03 the strings/hammers seem not to be consistently positioned in the room at the right position. It sounds like a two layer piano. Sounds appear on top of each other -- very weired. If you take 03 as benchmark. This rec sounds somewhat strange.


Finally I put on a new CD I bought the other day to compare your recordings.

Hideyo Harada - Schumann - HYBRID SACD. SQ 5* ( max) from Stereo in June 2010 magazine.

I played in 44/16.

I have to say. Your recordings do play a league higher then that. More air,
more real, better separation.


Well done. I do understand that it must be a hell of a job to catch the
music at these different locations.

Let see, what you or others say, about my comments. I am just guessing here.


Would be nice to hear about the original SPL inside the recording rooms during the sessions. Just to compare that to a similar SPL playback at home.

Did you compress the recordings?

Sorry for hijacking the thread!



Cheers
 
#568
Hi all,

I got some questions about the SD-player that I wanted to answer here.

Most of the onno classics tracks won't play on the SD-player, this is probably caused by incorrect file headers. The SD-player checks these headers and if contained information is incorrect it skips the file.

The file headers can be fixed by importing these files into iTunes (CD toast) using Apple Lossless format. Then create a play list and use ECDSD utility to convert and write the tracks to a directory or SD-card.


Some SD-cards with speed rating 2 ... 6 cause drop-outs. This is caused by SD-card access time (time required for the SD-card to respond to a command). This is problematic with real-time audio playback and finite buffer capacity.

I advise to use original (no imitation) major brands. I tested 32Gb Kingston (speed rating 4), with lifetime warranty, these can hold approx. 54 CDs (WAV) and work fine. This is also the max. supported card capacity by the (I)SD-player at this moment.

When connecting the SD-player to external equipment like (power) amplifiers, multiple ground loops are created that feed interference signals into the SD-player. Depending on mains pollution and properties of connected equipment this could lead to significant sound quality degradation.

The problem can be avoided by using external battery power supply for the SD-player. The SD-player requires 3 external 12V batteries. The +5V power supply draws up to 220mA (SD-transport, masterclock, reclockers, TDA1541A, JFET output buffers), the -5V (TDA1541A) and -15V (TDA1541A) up to 50mA.

12V / 17Ah for +5V and 12V / 4Ah for -5V and -15V will provide approx. 100 hours of operation, 12V / 8Ah for +5V and 12V / 2Ah for -5V and -15V will provide approx. 50 Hours of operation.

Batteries are connected using DC adapter plugs (center pin 2.1mm). center pin = plus, outer sleeve = minus. Fuse must be used on each battery for safety.

The (I)SD-player output signal is highly transparent and has high resolution (almost true 16-bit playback, depending on converter chip tolerances) this puts very high demands on connected equipment and interlinks (resolution and transparency). Equipment with limited resolution and / or transparency will lead to significant reduction of perceived sound quality.

Because of this high resolution / transparency, each and every flaw in connected equipment becomes clearly audible as they are no longer masked by interference signals. Sound quality now depends on the weakest link in the chain.

The elimination of SPDIF / USB interface, digital audio receivers related jitter (spectrum) and absence of asynchronous reclockers (ASRC) greatly improves clarity and resolution.

This also means that differences between different recordings will become very clearly audible now.

The clock feedback for slaving the SD-transport is very short (few cm). This easily outperforms external digital audio sources that are slaved using a much longer interlink for the clock or servo feedback signal. This way, both digital audio source and DAC chip receive very low jitter clocks.
 
#569
Highest sound quality for me

I 100% agree with John about what he says about the sound quality.
Just listen to the SD player quickly, and out of the box, these qualities are obvious. Before, I had a fanless PC cMP/cplay with I2S direct from ESI Julia soundcard to a Buffalo DAC, nice too be not close.
Thank you again John.
Alain
PS: battery plug should be long enough : 14mm work (10mm is too short)
 
#570
alain_online said:
I 100% agree with John about what he says about the sound quality.
Just listen to the SD player quickly, and out of the box, these qualities are obvious. Before, I had a fanless PC cMP/cplay with I2S direct from ESI Julia soundcard to a Buffalo DAC, nice too be not close.
Thank you again John.
Alain
PS: battery plug should be long enough : 14mm work (10mm is too short)
Click to expand...

1. It is a known fact - at least to me and a couple of other people - that
cmp2/Cplay is far away from being the optimum solution.
My own Linux setup on a FitPC outperforms cmp2/cplay easily.
Or just try a SB Touch and a Buffalo. At around 550$ you'll get a real
nice solution. I am pretty sure it'll be better sounding then that what you've got
before your SD-player.


2. We tested Johns SD-player against a well tweaked and powered SB Touch
solution (see my signature for modifications proposals, which were only partly applied at that time) and TP Buffalo 32
solution connected via SPDIF!
Honestly -- two of us (we've been 3 rather experienced people) -- the ones not
owning that device -- didn't get the impression that it would be worth
to switch to SD. There was no Day and Night difference.

The Sabre solution had more air&space, the SD solution on the other
hand a slightly better overtone spectrum. IMO these differences
were to be expected.
Of course to be fair - I don't think that we really fullfilled "all" the
conditions as mentioned by John, such as running 3*12V batteries,
avoiding all kind of ground loops ( on a single ended device 😉 ), etc..
This is IMO gonna be a real challenge for the customer to fulfill all this.

(Recently the Touch/Sabre solution has even been further modified and
improved.)

3. One should not to forget the volume control issue!! You'll need a pot in
front of the amp if you go the passive route. That one also needs a
careful implementation. We didn't have that problem on the Touch
solution.

4. I mentioned that before. The coupling cap issue. There is IMO space for
improvement on the SD player too. Perhaps a nice transformer would do
a good job here. This way you'd also get rid of potential groundloops and
you'd be able to lift up the pretty low output voltage level.
Here a Sabre gives you more options such as differential out.


In the end you'll face compromises here and there, as with any solution.


Let see how things will develop. Enjoy your stuff. 😉

Cheers
 
#571
New vs old

Hi all,

I have the oportunity of listening a lot of sources and amps, etc, because my work. Lately, I've heard the Esoteric D-07 versus an own modified old Philips CD with the TDA1541A. I've put a superclock (my own design), I've bypassed the SAA7220B, I've made a discrete output based on ECD, named "1541IV concept" I believe. I let the CD player warm for 2 days, and I compared. The old converter has a very high resolution. At the moment it isn't as open as the Esoteric, but, it really sounds very very real, very natural. At home I have the same CD player, but with tubes output, very high quality capacitors, etc..., and the sound is still far better. On the other side, the ESS sabre it sounds too very fine and open, but IMHO, it lacks of some natural sound. I've heard the Copland CDA825, and it sounds very fine too, it has a pair of WM8741 dacs. I think that we still have to extract all the information and natural sound of the 16bits technology. We are far from 24 bits. Simply begining with the pcb, there is phisical limitations on implementing such a system. I believe that we have to eliminate at "quantum level", as much noise as possible. why then should we use a lot of digital filters?.


Best regards,
 
#572
Volume control

Klaus,
Very interesting information, and your DIY website is great.
I've listened with no volume control and no coupling capacitor. But just to stay in the WAF safe area, I quickly ordered a 5K Goldpoint stepped attenuator 😀
You and John have used attenuators between the power amplifier and the speakers; do you still consider it as a good solution ? And is it just a voltage divider with the correct power rating ?
Best Regards,
Alain
 
#574
filters

Hi, Sonic real one,

I kindly apreciate your opinion, of course, but I do not agree with you. After bypasing the upsampler x8 + digital filter chip + analog filtering, the sound you get is very clean, even not having yet implemented the DJA (by ECD), nor having treated the data signals with a reclock or something simmilar. Simply, the digital filtering is an invention that began with the TDA1540 because it has 14 bits, and just some time after, the industry changed the standard definition to 16 bits, and the Philips engineers had to adapt it to 16 bits. How?, just filtering and upsampling the frequencies. And after this, well, you know the history...filters and more filters..etc.... IMHO, we have first to reach the top level of the 16 bits system without filtering, and then, if necessary, add some kind of "filtering". It is like a kind of patch to correct a bad design, like the NFB.

Best regards,

PD: Did you've tried the PCM1704 with and without filtering?, and with/without upsampling?, you'll be very surprised.
 
#575
galeb said:
Simply, the digital filtering is an invention that began with the TDA1540 because it has 14 bits
Click to expand...

Your history of CD sounds just a little garbled. Digital filtering of course existed before CD, but you are right that Philips used 4X oversampling to get better than 14bit performance out of their TDA1540. That was their best dac at the time, whereas Sony (their partners in developing CD) had a 16bit part. I seem to recall though that Sony's dac was more expensive to produce. They only used one, timeshared, in their first CD player.

and just some time after, the industry changed the standard definition to 16 bits
Click to expand...

No, the CD standard always was 16bits, courtesy of Sony who thought they'd have a slight lead over Philips by insisting on this number as the standard. They knew Philip's dac design was 14bits apparently.

and the Philips engineers had to adapt it to 16 bits.
Click to expand...

Philips maintained their 4X oversampling approach even when their dac technology improved to 16bits (stereo even, the 1540 being mono) in the form of the TDA1541. I had one of their earliest players with this part, the CD-160.

IMHO, we have first to reach the top level of the 16 bits system without filtering, and then, if necessary, add some kind of "filtering". It is like a kind of patch to correct a bad design, like the NFB.
Click to expand...

Curious to know where you picked up this opinion from? If there's a 'bad design' its that the CD sampling frequency was set too close to the top of the audible range so anti-imaging filters (as used by Sony on their CDP-101) had to be very steep.
 
#576
TDAs history

Hi Abraxalito,

You're right respect to the TDA1540 and 41 history. But I do not agree with you about the filtering. For me, any kind of filtering is any kind of "hide" the natural and real sound. Certainly, the base design theory of the CD is not optimal, of course, but I've made a lot of tests, with a lot of dacs, I've heard a lot of product too, like Esoteric, Wadia, Theta, Krell, Audioresearch, Copland, Electrocompaniet, etc, and after this, I still think that the pure conversion, if well done, is the best. Any kind of "masking" technique like filter, is not good, at least, for me.

Kind regards,
 
#577
galeb said:
Hi Abraxalito,

You're right respect to the TDA1540 and 41 history. But I do not agree with you about the filtering. For me, any kind of filtering is any kind of "hide" the natural and real sound.
Click to expand...

Well the digital recording process itself is a kind of filtering because it only works with a band-limited signal. So I guess you'll just prefer analogue? Yet even that has filtering because it can't be infinite bandwidth.

Certainly, the base design theory of the CD is not optimal, of course, but I've made a lot of tests, with a lot of dacs, I've heard a lot of product too, like Esoteric, Wadia, Theta, Krell, Audioresearch, Copland, Electrocompaniet, etc, and after this, I still think that the pure conversion, if well done, is the best. Any kind of "masking" technique like filter, is not good, at least, for me.
Click to expand...

That's what you hear then I'm not going to argue. But I'm myself not sure that the "problems" with filtering are really problems, I've also heard NOS and haven't noticed any superiority over OS. But that could just be because the dac I heard (TDA1543) isn't a particularly good example - it would probably sound unimpressive if oversampled. I'll just stick with OS until I hear an audible improvement in a NOS dac.😀
 
#578
... when they produce CD-s, they downsample the 192khz or 96khz/32bit material onto 44.1/16.

-they use dither AND ,a steep filter with pre-ringing to do this. If you dont use a phase linear (pre ringing alike ) DF to reconstruct this data, but a simple zero order hold alike NOS dac , what do you get? -let me help : staircase pre ringing 😉 ...
Arguing about decade old philips dac and DF has little point in this discussion.
 
#579
filters

I'm not talking about the recording process, only the listening results. The recording process is another entire world. In the end, the thing I want to transmit is that the NOS dac, well implemented, has a lot to do in the 21 st century, and the last dacs, if well implemented too, they sound good too, but the NOS topology, for me, sounds still more "natural". It it's difficult to hear a system for 2 hours or more, is because it isn't natural. I've herad the Esoteric 01 and 03 series, with atomic clock, transport, dac, etc. Opus cables, Wilson Alexandria MKIII, Audioresearch REF5&REF610T, KRELL EVO1&2, EVO900, EVO600, too many different cables,etc, (I can do this because of my work), and I have to say, that the definition is superb, it has a lot of detail, good bass, etc, but, always is a "but", for me, and only for me, there is a lack of "natural" sound.

Kind regards,
 
#580
alain_online said:
Klaus,
Very interesting information, and your DIY website is great.
I've listened with no volume control and no coupling capacitor. But just to stay in the WAF safe area, I quickly ordered a 5K Goldpoint stepped attenuator 😀
You and John have used attenuators between the power amplifier and the speakers; do you still consider it as a good solution ? And is it just a voltage divider with the correct power rating ?
Best Regards,
Alain
Click to expand...


It doesn't make any sense to drive down your analog signal with a pot into the noise floor. With the 1541 you'll be working on a pretty low voltage level already. In any case you should make sure that you run 0db attenuation on your low-level recordings at your preferred listening level. The same logic applies if you do digital volume control. My guess >95% of all DAC/Amp setups out there are less then optimal dimensioned.

Don't forget that modern recordings come with levels up to >8-12db difference compared to traditional recordings.

That means you'd have to apply a quarter of the original signal ( approx. 0,125V) only to cope with the differences related to the recording level.
I doubt that many pots sit on the 0db setting if you play low level recordings. That's why things get even worse in real live situation.


2nd you'll have changing impedances on the connection. ( This discussion we
had during times at times where transformer pots were hot topics)

But this we've discussed before. John came up with volume control in front of the speakers at that time. Of course this you can't sell.


What you can try. Lower the volume digitally ( using e.g. Sox) and save those files on the SD card. Make sure you avoid automatic dither.
Then match your amp-gain/speaker-spl/prefered listening SPL.


Cheers
 
Status
Not open for further replies.
Top Bottom