Firmware Versions and Digital Sources

Elsewhere on the forum comparisons of firmware updates for DirectStream DACs range from disagreement on sound quality (subjective) to disagreement on update process success. I wish to intoduce the topic of the file source as some cause of these different sound quality veiwpoints. I can best do this with a brief summary of my own subjective experiences.

My system reference before PS was a fine CD player which as a redbook disc player and DAC always outperformed an OK streamer and computer quality NAS with FLAC files rippped by Dbpoweramp

A DS Junior and Bridge with Yale was introduced as streamer and DAC, and the differences between CD replay and FLAC files from NAS narrowed to evens. It was as good as the CD player DAC but presented differently with detail over fluidity (?).

Updated to Torreys (no problem) the reference sound was to listen to FLAC from the NAS, a moderate but distinct improvemnt over the digital feed from CD player, or from the CD player DAC.

Updated to Huron ( did not go well but rescued by Kevin, UK importer and Jeremy at PS Audio) and my first listens to FLAC files from the NAS seemed rather similiar - though some time had intervened and made comparison even more subjective. The surprise reference (and definite wow factor) came when I fed any red book disc from the CD player by coax to DSJr. This was better than anything before by some margin, in detail, image and tonally. Files from the NAS were indistinct and less dynamic by comparison.

Since then I have also used streamed FLAC files from Qobuz (no fancy internet connection) and this performs as well as NAS files but cannot match the CD player output. By the way the analogue output is through substantial balanced valve amplifiers and hybrid horn speakers which are dynamic,open and revealing by most listners standards. Analogue cables are higher end but digital connections are moderate quality items.

My question to the forum is which source(s) might, or might not, be favoured when I upgrade to Redbank ?

PS I have avoided being specific about my equipment as others will have there own taste naturally. The piece in common is the DS DAC.

I think you might still have too simple a model of the things that can affect the sound in your system. At times the way you route a cable will have more effect than the makeup of the cable itself. If you hear a difference between two different inputs the only way to get closer to what’s causing the difference is to unplug all other inputs (not just power off their sources) and at times move the sources to the same physical place with the same power cords plugged into the same outlets, etc. Needless to say this makes quick A/Bing more inconvenient.

I’m not trying to make anyone paranoid, I’m just trying to point out that extrapolations about what makes a difference in one system to another system is fraught with problems. Usually this is even more true when throwing in software updates in the DS.

There were some exceptions: one release quite a while ago made a significant difference for SOME people who were using AES/EBU or S/PDIF as their primary input (depending on the quality of the digital interconnect that they were using.)

Some releases have bigger differences in the source part of the FPGA code and other releases concentrate on the output code (this is the case for Redcloud and Huron.) Changes in the output side often affect DSD and PCM similarly. Changes in the upsampling may affect PCM more than DSD.

From a logical point of view I’d expect that the sensitivity of the sound quality on particular inputs to be the smallest between Huron and Redcloud than any other software upgrade. Most of the changes in FPGA code affecting inputs specifically were in the earlier releases.

Stephen, depending on the revealingness of your system you should stream from aif or wav files. They sound clearly better than flac and when comparing with physical CD you should use those formats.

I have a slightly different point of view: aif (and wav) support losssy compression so they are not a quality guarantee and wav files don’t support meta data nearly as well as .flac files. If you are concerned about the CPU power and possible noise generation when decompressing then use .flac with no compression - then the audio is stored in essentially the same format as wav files with just the metadata format being different.

Another feature of .flac is that it has a checksum for the audio so you can tell if something went wrong on your disc or backup by checking that checksum periodically (and yes, this does happen, even tho discs are incredibly reliable, they are now huge and you have to divide that reliably by the amount of data you store to get a rough idea of how often you’ll get an uncorrected data error.)

Ted Smith said

I think you might still have too simple a model of the things that can affect the sound in your system. .

Ted thanks for the comment. I agree entirely about the many seemingly superficial things which change a system and have spent 40yrs chasing lots of them around! My surprise was not that different inputs sounded different but that firmware updates alone changed my order of preference for those inputs. Is there an inherent reason why Huron should sound so good with the digital output of a CD player , when Torreys always sounded better from a NAS file?

Thanks for specific reply that Redcloud and Huron should be similiar in that respect - that certainly makes me want to update!

Ted - re: your Post 2 - this is the thing I keep coming back to more and more these days.

We have a tendency to assume equivalent/same/similar conditions, whereas I would hazard (goin’ out on a limb here) that pretty much ALL OF US have completely different chains - yet we discuss it as if we were all listening on identical Reference Systems in the Same Room.

I’m not (usually) Paranoid, but all this is indeed Fraught : )

badbeef - it’s even worse than us all having different sytems - we put them in different rooms and then listen with only ‘our’ ears and brains.

It’s amazing we ever have a sane conversation about any Hi Fi, let alone accept someone elses opinion. But we do (cautiously!)

Yeah - actually not much of it is sane. In the strictly tradional sense : )

That’s interesting Ted.

I can only say that I compared compressed flac to uncompressd flac to aif to wav after Paul gave the hint. I did this very carefully (and it’s easy to do) because it meant a noticable increase of harddisk space and I definitely heard quite obvious differences and did so in a previous comparison of flac to wav…I just didn’t want to pass on meta data, so I chose aif.

Regarding the technical background I don’t know much, just which has metadata in it and which not and that flac (uncompressed or compressed) isn’t recognized as PCM native format by the streamer/library software while aif and flac are. The technical explanation I knew so far was, that there’s even an add. CPU consuming process going on with uncompressed flac, that isn’t there with aif or wav.

Im not clear about what you said that aif and wav support lossy compression and it gives no quality guarantee, can you explain?

If you are the one generating the wav or aif files you can make sure they’re internally in 16, 24, or 32 bit PCM. But both formats also support lossy compression (some really bad) and those files aren’t great.

Everyone should do the file format checking themselves with the equipment they are going to use.

But, for example, if you are using the bridge you have two places you could be unwrapping the .wav or .flac (and perhaps the .aif, I don’t know) and they probably sound different and you might have skipped some tests and are really talking about what happens if you unwrap in the host computer vs if you unwrap in the bridge, etc.

As another example making sure that the files in question are occupying the same places on the source disk (or at the very least have the same disk fragmentation) isn’t easy. Often people have files that have been defragged by the OS as their source and then they convert to, say, flac and end up with a highly fragmented file (or the editing of the meta data causes a repositioning of the file and you end up with differing fragmentation). In any case you might be comparing how fragmented files sound vs. how unfragmented sound with your system. (Also you need to check if the files themselves are being compressed on the disk by the OS…)

I’ve just seen supposed experiments in otherwise respected magazines, etc. that purport to show one file format does or doesn’t sound different from another and they aren’t really comparing apples to apples.

There are some systems out there that definitely sound better with (losslessly) compressed files because their disk drives make a lot of electrical noise when seeking and doing fewer seeks is better.

Thanks for these thoughts Ted, I understand now,

For my part, I can ensure, the disk itself doesn’t compress either file and I compared both versions unwrapped by the bridge (at least I don’t know how I could have unwrapped them by Jriver in advance. If so, it would have been for both of them). I also made the tests with several examples of songs and partly copied them to the disc at the same time newly, so at least some of them probably should have had no different fragmentation.

So for my part I can just say it should have been apples/apples and I’m sure I heard the difference, which was no matter of 10 times forth and back comparison, but noticeable by one try with a verification. As I mentioned elsewhere, especially increased top end openness was quite easy to hear, bass improvement a little harder and ambiance improvement was just a little bit better with aifmwhen comparing uncompressed flac to aif. The easiest to hear immediately was compressed flac compared to the clearly better uncompressed flac and especially aif.

I can imagine that i.e. with a slow NAS, the bigger file sizes of the uncompressed PCM formats vs. flac could have different results or apples/oranges comparisons as you mentioned.

for me with all firlmware i choose flac files on NAS, minimserver and minimstreamler transcoding in wav24, read by dedicated pc with jlaystreamer and usb inpout of DS. the bridge never sound as defined and musical, but i have not tried redcloud, maybe this time the bridge will sound good for me, and justify i bought it…

Good grief, you guys are making my head spin. I just play the ones and zeros exactly as they are etched into the CD plastic, or exactly as they come out of the Spotify pipeline. Nothing much else to think about. FLAC, AIF, WAV, NAS, TRANSCODING and all these other words in the foreign language of which you speak is hurting my brain. I know when I’m reading it I can’t hear as well, it consumes too much of the CPU in my brain. I like simple. Connect the CD player to the DAC. Connect the ethernet cable to the DAC. All done. Sit back, listen and enjoy.

I now have Redcloud up and running on my DSJr and can (maybe) answer the question I posed at the top of this post.

First impression with Redcloud: there ae now less differences between the same CD played from a disc player straight to DAC, or a FLAC file from my NAS, or a FLAC file streamed over internet from Qobuz. (The last 2 both coming through the Bridge)

Good news is all sources are very untiring to listen to, but I bet I will end up with a ‘favourite’ source after a lot of listening. Having said that Ted has me worried about file fragmentation on the NAS. Should I check that and run defragmentation software??? Is this the road to paranoia?

PS the Bridge update which comes with Redcloud does let MConnect play from Qobuz and VTuner radio with no glitches. This is easier than work arounds I used to have to do to play them with Bubble. I dont think the control software makes a jot of difference to SQ so just use whatever suits your needs.

Sorry about that - my point wasn’t that file fragmentation is that big of deal, but that there are a lot of things to consider before making categorical statements about the audio quality of things like audio storage formats. They should at least address the possible effects of the storage device…

I know of multiple claims of copying a wave file and hearing a difference (in a blind test) - would I worry about it? Not at all. There are plenty of other things to do in your system that will make a bigger difference.