Raspberry Pi 4 and Caiman SEG

[Labarum]

I have not posted here in ages and only recently returned to the Squeezebox forum.

I have restored Squeezeboxes to my home network and moved Logitech Media Server from an ageing PC to a Raspberry PI 4 running PiCorePlayer. It works extremely well.

In Southampton I have an old 7510 DAC which has no USB input. In Nicosia I have an original Caiman with Gator board fitted DIY. That too still works well.

My plan is, in due course, to get a new Caiman SEG and connect the Raspberry Pi 4 directly to it by USB.

Has anyone in these parts experience that combination? In the other place the opinion was offered that the USB out from a Raspberry Pi 3 was a bit rough, but from an rPi4 was excellent. The Topping E30 is known and common in the Squeezebox forum, and PiCorePlayer has an adjusted USB output option for it, but Standacs are virtually unknown.

Does anyone here use PiCorePlayer via USB to a Caiman? Can anyone shed light on the need to have various USB audio output configurations? Surely USB out is USB out!

Here is my Raspberry Pi 4 in a KODI case next to my router with a CD to show the scale. Tiny, tiny, tiny!

30079

My Flac library of ripped CDs is now on a 128GB button sized USB flash drive, and that's my complete server! It runs on less than 10W.

I am presently using a Google Chromecast Audio dongle as a virtual Squeezebox optically coupled to the 7510 DAC, but if I get the USB connection right I can lose that extra bit of hardware and put the Raspberry Pi behind the TV next to the DAC.

What can anyone tell me?

[Alp]

I used the USB output from a Pi3 to a Caiman SEG a few years ago and it worked fine but the Allo Digione Pro SPDIF was better. I replaced the SEG with a Topping D90 and the Pi3 with a Pi4 and spent quite some money on a reclocker and expensive USB cables only to conclude USB is not the way to go unless you need very high sampling rates or are willing to spend serious money. I now have a Pi2AES hat in my Pi4 feeding I2S to my DAC but SPDIF is almost as good. With the SEG, the Digione is worth a try.

Sent from my SM-G988B using Tapatalk

[Labarum]

Thanks Alp.

Stan is about to send to me one of his latest batch of his latest DAC. My 7510 is close to 20 years old!

Experts on https://forums.slimdevices.com/ offer the opinion that whereas the USB output from a Pi3 could be rough that from the Pi4 is extremely good.

Time will tell. Maybe Stan has some comments even now.

[Lawrence001]

Can't comment on the individual DAC in question, but from my experience older DACs have weak internal USB to SPDIF convertors and you can't judge USB as a medium by listening only to them. A decent external convertor will upgrade them and keep them sounding their best with any source.

I've read reviews of more recent DACs where the USB was said to be the best sounding input, so it depends on the individual DAC and implementation.

Sent from my BKL-L09 using Tapatalk

[MisterMoseley]

Can’t answer the specific question re Picore player to a Caiman but I have used a RPi4 for about 18 months now, with a variety of DACs.
I previously used the USB connection and about 6 months ago, bought a 2nd hand Allo Digione Signature, allowing me to connect to my DAC via co-ax.
Over Xmas, I put together another RPi4 streamer and whilst it was in my possession, couldn’t resist doing an A/B between the 2 RPis. Both playing exactly the same FLAC tracks, into the same DAC (one USB, the other co-ax), into the same amp and speakers.
I’m not going to say that I couldn’t hear a difference, but quite honestly it was so small a difference (in favour of the co-ax connection) that I’d be more than happy reverting to USB if I had to.
I suspect that part of this may be down to the quality of the USB connection on my DAC - a fairly new Schiit Modius.

[StanleyB]

There are so many variables to think of, I have started to get paranoid about the quality of the individual audio files that I have in my library. I'll probably do a write up on various technical aspects of things that I found out compared to what is generally talked about. But I might end up turning others paranoid as well...

The reason I typed the above is because the quality of the original file will affect the quality of the playback. Less forgiving decoding can quickly make a bad quality file sound bad, whilst a more forgiving decoding system can still make a poor audio file sound reasonably acceptable. Whether you use USB or optical is down to personal taste in any carefully assembled audio system. I myself prefer playing back some tunes via USB, whilst others sound nicer to me from optical. I don't so much use "better" as my reference, but "nicer".

[Labarum]

There is one issue using the SEG and a Raspbery Pi and PiCorePlayer via USB , the cause and remedy you will see in the link below.

When connecting by USB the SEG has to be running before the rPi is started up (or at least before the player component "Squeezelite" is started.) If not the DAC cannot see the Pi.

https://docs.picoreplayer.org/projects/autostart-squeezelite-from-usb-dac/

The remedy requires command line adjustments. I will leave it as it is - my Linux knowledge is not great and I might break something I cannot mend! I hope some future upgrade of PiCorePlayer will have the remedy written in. In the meantime I will remember to turn the DAC on first - though both SEG and Pi I leave on 24/7

This issue occurs with at least some other USB DACs.

[Lawrence001]

There are so many variables to think of, I have started to get paranoid about the quality of the individual audio files that I have in my library. I'll probably do a write up on various technical aspects of things that I found out compared to what is generally talked about. But I might end up turning others paranoid as well...

The reason I typed the above is because the quality of the original file will affect the quality of the playback. Less forgiving decoding can quickly make a bad quality file sound bad, whilst a more forgiving decoding system can still make a poor audio file sound reasonably acceptable. Whether you use USB or optical is down to personal taste in any carefully assembled audio system. I myself prefer playing back some tunes via USB, whilst others sound nicer to me from optical. I don't so much use "better" as my reference, but "nicer".I'd be interested to know your thoughts on this. I only discovered recently when playing around with the config file on my Zen that there are various compression rates for Flac (1-9?), and the default is usually 5. I reduced mine to 1 and thought it sounded marginally better.

There's also an optional volume equaliser setting (and I'm talking about the music signal, not metadata) which would be even worse I imagine.

The former caused some controversy on the Wam when I posted (I think I was in a minority of 1 in thinking there's an issue). I don't think anyone believed me about the latter, every time I tried to explain they kept saying don't worry it's just metadata (but that has a separate setting of its own in the config file).

Sent from my BKL-L09 using Tapatalk

[StanleyB]

I mentioned numerous times in the past that there is an audible difference in the compression levels of FLAC. That's why I don't use any compression if I am the one creating the FLAC file from for example a CD.
The volume level thing is often referred to as Replay Gain. It is stored in the metadata. But not all replay gain setting programs use the same method to calculate that value. And some even modifies the actual audio file instead of just adding the replay gain value to the metadata. The best known program for that out there is MP4Gain - https://mp4gain.com/mp4gain
I prefer to use the replay gain function in MediaMonkey, which doesn't modify the audio file.
But "correcting" the replay gain doesn't mean that you get an improved signal output. So any clipped signal peaks will still be clipped after the replay gain is set. However, it is a useful metadata marker. Many ICE (in car entertainment) systems can now read that replay gain value and adjust the sound output of each track so that you don't have to fight with the volume control to keep the perceived audio level at more or less the same strength.

[Labarum]

I mentioned numerous times in the past that there is an audible difference in the compression levels of FLAC. That's why I don't use any compression if I am the one creating the FLAC file from for example a CD.

I don't understand the point of creating a FLAC file with Zero compression. Why not just leave it as a WAW file (or similar uncompressed format) if the file is not shrunk for easier storage? Maybe there are systems that will not play WAW or whatever, or maybe it's a question of attaching metadata.

There is of course confusion in some people's minds about the various sorts of compression:

1. FLAC and similar formats that are like ZIP files unzipping back to where they were;

2. the lossy formats like MP3 that decrease musical resolution to save storage space;

3. and dynamic compression that (by analogue or digital process) reduces the loud to soft ratio permitted in the file and on replay.

It seems to me that dynamic compression sucks the life out of the music faster than the lossy compression used in MP3 and similar. Radio stations apply different levels of dynamic compression - some like Radio Swiss Classic - are only MP3 128kb/s but maintain an open sound when playing CDs, presumably because they are gentle with dynamic compression.

[StanleyB]

WAV files are not so good with metadata. And they can be so large that they cause drop outs if you are streaming them wireless. My 192kHz WAV files cause a lot of buffering on my network. But the fLAC versions play fie in real time. And you can forget about trying to steam 384kHz WAV files. Its a constant buffering playback. Those ones I have to play back via a wired connection.
The uncompressed FLAC files are also smaller then their WAV equivalent. But isn't so much about size really. It's not so easy to get music in WAV format unless you buy a CD.

[struth]

ive never had an issue with flac files, but wav as said are not so happy with metadata and can be pretty big. when youve a big collection, it takes up a lot of room. I think if your obsessive you can make an initial wav copy, as a master and then copy over a flac one as the user. Ive tried using the different rip levels years ago but found the one the program selected was fine and no audible difference. i used to use fancy rippers, but found no difference in 99.9% of discs.

[Edward]

IME (as that is all I have to go on) using a low powered CPU to decode a FLAC file (when compression is used) leads to audible degradation, using a sufficient CPU leads to zero degradation (as compared to say WAV files).

There are numerous other variables of course (such as what software is used for decoding).

:)

[Labarum]

I was not suggesting the use of WAW, but questioning Stan's words "That's why I don't use any compression if I am the one creating the FLAC file from for example a CD."

A detailed technical explanation seems to be here, and all flac file generation involves compression

https://sound.stackexchange.com/questions/41964/flac-compression-level-comparison-efficiency-analysis

"FLAC compression levels are (only) a trade off between encoding time and file size. "

So I remain at a loss to know what Stan was trying to say.

[StanleyB]

I was not suggesting the use of WAW, but questioning Stan's words "That's why I don't use any compression if I am the one creating the FLAC file from for example a CD."

A detailed technical explanation seems to be here, and all flac file generation involves compression

https://sound.stackexchange.com/questions/41964/flac-compression-level-comparison-efficiency-analysis

"FLAC compression levels are (only) a trade off between encoding time and file size. "

So I remain at a loss to know what Stan was trying to say.
I am saying that I know better. Of course anyone is free to believe what they read on any other site. But please don't assume that they are smarter and know better than me. I made a living from knowing my digital stuff since the early eighties.

[Labarum]

I am saying that I know better. Of course anyone is free to believe . . .

That is not a particularly helpful or edifying posture, Stan.

Doing some more reading I note that as well as FLAC compressions 0-8 there is also an "Uncompressed FLAC" option, and that I did not know. I am guessing that is what you are referring to when you say "That's why I don't use any compression if I am the one creating the FLAC file from for example a CD."

Now I need to learn some more. I am supposing that an Uncompressed FLAC is something akin to a WAW but wrapped up in such a way that metadata may be conveniently attached; and I am supposing that such an Uncompressed FLAC would stress the playback device less than a highly compressed format that has to be decoded on the fly.

I would value more elucidation. Is anyone offering?

[Lawrence001]

I'm with Stan on this one, from experimenting I prefer lower compression rates in Flac files and nobody will convince me otherwise.

I experimented with a setting of 0 but the file size was the same as the default (5) so I put it down to a quirk of the Vortexbox software and use 1.

Sent from my BKL-L09 using Tapatalk

[Lawrence001]

Anyway nice to know in not I'm a minority of 1 any more, and it's someone with the reputation of Stan who agrees with me :)

Sent from my BKL-L09 using Tapatalk

[Labarum]

I must rip a CD to uncompressed FLAC and have a listen myself, but I cannot reliably differentiate between a good quality MP3 (or AAC) stream and a FLAC stream. The best Internet Radio Stations that use lossy streams are as good as the few that stream FLAC.

On other fronts I did notice an improvement on the recent move for my old 7510 DAC to a new Caiman SEG. One of the reasons for getting the SEG was its USB input. My PiCorePlayer running in a Raspberry Pi4 now drives the DAC directly rather than via a Google Chromecast Audio. There is a workaround to allow that marriage. A plugin called Castbridge runs in Logitech Media Server/PiCorePlayer and it enables a Google Chromecast Audio to masquerade as a Squeezebox with caveats. The music can be sent track by track to the Audiocast, but without gapless replay - there the FLAC is decoded in the Audiocast. For gapless replay the whole ripped CD or playlist must be decoded from FLAC in the server and sent to the Audiocast as a straightforward single PCM stream. That is stable even with the few 24/96 FLACs I have. Either way I hear no difference when listening via the Caiman SEG driving my Sennheiser HD25 headphones; I did, however hear some improvement when removing the Google Chromecast Audio dongle from the chain. (I was using it only as a transport to feed the Caiman SEG optically.)

And all of this of course against the observation that the Caiman SEG does not measure well by modern standards. I knew that when I bought it, but did so because it has pre-amp features that cannot be found elsewhere at the price, and it is well and robustly made. The ruling principle was again "I won't hear the difference" and in general - in engineering "good enough is good enough."

I'm not wasting my life looking for the pot of gold at the end of the rainbow. I just listen to the music.

[Lawrence001]

I did 2 rips, one on setting 1 and the other on 9 to compare. If Flac is lossless then there should have been no difference but I'm sure the latter was slightly less open and airy and lacked the subtle nuances of things like guitar strums and breath of vocals.

This was not blind and I accept there is potential unconscious bias, but as storage space is not a limit I don't see any reason to increase compression rates beyond the minimum. These subtle improvements can make the difference between a short and long listening session into the night.

Sent from my BKL-L09 using Tapatalk