There are multiple separate issues. Clicks and pops during format/sample rate, etc. changes and clicks and pops with very loud music.
Clicks and pops when going from DSD to PCM, PCM to DSD, changing sample rates (and the source clock starting and stopping.) are highly source dependent. Some of them were fixed going from Torreys to Huron, other’s got worse. In Redcloud I added a generic hack when I notice such discontinuities which helps most people at least a little. I might have a bug going from PCM to DSD (or vice versa) but all of the issues I’ve been able to reproduce here are the source sending “bad” data at the discontinuity. E.g. foobar2000 (at least the most current versions) goes to PCM in between DSD tracks and that’s one of the pops that people hear. Similarly with various versions of JRiver. In general the DAC can’t solve all of these programs (how long should it wait with no data before it decides that the source has stopped?) but I can adjust some parameters so that the fewest people encounter clicks and pops.
If you get clicks and pops at other times while music is playing on Torreys or Huron, you might have a bad cable (or too long of a USB cable), a network bandwidth issue, etc. that’s causing tiny dropouts. The above fix can make those sound a little worse in Redcloud than in previous releases (i.e. the DS might see the missing, corrupted or slightly late data as a format change and do a quick volume ramp down and back up.)
There’s also a problem with very loud music in Redcloud - perhaps just a very momentary crackle at peaks or in some cases continuous crackles with some especially dynamic loud (“peaky”) music - this happens a little more often in DSD since DSD can get about 4dB louder than PCM (but DSD shouldn’t stay that loud very long.) Each step the DS’s volume is turned down should lessen the probability of one of the crackles. They happen with loud tracks a lot more at volume 106 than volume 100 or 90. I have in hand some tracks that crackle every great once in a while and a few that crackle almost continuously. I also have FPGA code that fixes the problem for everyone that’s tried it (tho some people are encountering a different problem and it doesn’t help them.) There’s no sense asking for that code since it hasn’t had it’s audio quality graded and the particular version I have sounds worse than Redcloud on the average. There are some other issues that PS Audio has to run down, (e.g. the 24/16 sample width issue for supposed 24/192 MQA…) before they can plan a new release. To me it seems like forever since Redcloud was released but since PS Audio has been on vacation it’s been a lot less time for them
Another helpful explanation Ted. I will just stick to Redbook through the DMP running with Redcloud in the meantime… Listening to DSD with Huron although devoid of clicks and pops isn’t as musical as Redcloud… A few SACD’s click and pop so bad that they are unlistenable …
Ted, I often notice the “crackle” you’re referring to. It kind of sounds like clipping/distortion to me and I have always wondered if this was source related or if the DAC is doing it (or some other component). But now that you mention that you hear it too, it confirms this artifact is on the DAC. This is mostly with PCM music (as I have a pretty small DSD library), but it doesn’t happen on all music I listen to. Like you said, it’s only when things get "loud’, which is why I thought it was clipping. I usually listen with the volume set between 50-80 with the attenuator engaged. Is it normal in your opinion to be hearing this crackle fairly often? This is actually the second Directstream I’ve owned and I don’t remember hearing this issue on the previous Directstream, but as you point out, it’s probably because the old Direcstream I used to own had an older firmware on it not accentuating the artifact as much. I really wouldn’t mind trying out the other firmware to see if I prefer it to RedCloud, but if this is something you could fix in a future firmware while retaining redcloud’s performance, that would be awesome and I’ll just wait it out.
In general the DS (and before Redcloud) has about 2-1 headroom in the analog and about 2-1 headroom in the digital domain to handle issues like clipping. I.e. the DAC shouldn’t clip or max, or round at all at 100 or below. In fact no one has ever reported the slightly rounded tops at 106 so I’m very sure you aren’t hearing clipping.
In Redcloud the bug I fixed was a too small accumulator in the SDM - my experience was only that very high level (near 100) would cause it to happen very often, but with slightly quieter music the odds go down. It could sound like clipping but I heard it more as crackling - it lasted a little longer than just rounding the tops of some signals. I don’t expect that this bug happens very often at all below, say, 80 or 90. A sure test for this bug is that lowering the level a little (a few volume steps) immediately changes the character of the distortion, and lowering it more stops it from happening. To test I would lower the DAC’s volume while raising my preamp’s volume which helps keep loudness issues from clouding the matter. (Of course I can easily see it on the scope with test tones, but this doesn’t help others recognize the problem.)
I was afraid that if I mentioned the bug, a lot of people would get paranoid thinking that almost anything anomalous in their system (or their use of the DS) was the bug… So far, crackles at lower levels within tracks has only been reported with the bridge (and I’m not keeping track of those) or USB cables that are too long (or have way too much impedance), or trying to run TOSLink at 176.4 or 192k with too cheap of a cable, etc.
I don’t think there’s another (sound quality related) bug in the FPGA and as I said I’ve had more than a handful of people test my fix for me - and I’ve got weeks of listening to it.
I’m currently only using I²S inputs. I’ll try out a few things to see if I can remove as many variables. It may very well be an issue elsewhere or specifically with this newly acquired Directstream (I’ve only had it for about a month now). I just don’t remember this issue with my previous Directstream.
Yes, certainly try the other I2S input. I know of at least one unit out there that crackled on one I2S input but not the other (AFAIK they either chose to not care or they were taken care of by PS Audio support.)
As a new DS owner (2 weeks) I thought I would chime in.
I hear one click when listening to DSD and switching between PCM and DSD.
I have my music on a NAS and run Minimserver with Bubbleupnp server used for OpenHome emulation and Lumin App.
But, here’s the thing…I had the same click occurring in the exact same way when I had my Sonore Rendu renderer feeding my W4S DSDse DAC. So when the DS arrived and started doing the same thing, I thought all was working as it should. LOL…
Thank you to @Ted Smith for explaining the reason for this anomaly.
Yep, those clicks between DSD and PCM are really annoying - I didn’t used to get them with foobar2000 - I don’t know when they crept in. I do know that we keep delivering an older set of USB drivers because newer ones click more between DSD and PCM. I had a Sonore Rendu in hand here to look at them, but the clicks came from deep within the device so I couldn’t patch around them in the DS without making the problem worse for other devices/software.
The transitional clicks do not bother me much as I know what causes them and I switch only between albums. But they could easily be annoying if one has a playlist which switches between tracks, etc.
With my Rockna WDN switching from PCM to DSD produce a shorter (one-off) transient with Redcloud than with Huron, but it is still present and make you jump in your seat.
I am still getting some clicks and pops during dsd files (and also between tracks) on my dsd dac. All pcm files are fine.
With the usb input, they happen when the attenuator is on, even if I lower the volume to 80. If I switch off the attenuator and turn the volume down to 60, they disappear. This is the workaround I currently use.
Ditto for sacds from the dmp.
With bridge 2, I still get some faint clicks, even without attenuator and volume at 60. So I use the bridge only with pcm files.
My local dealer has replaced my bridge, and I have downloaded the 3.5.1 bridge update. I have gone back to Torreys and then back to Redcloud several times.
One thing that’s confusing is that the only difference between having the attenuator in or out is the state of it’s relay. All the attenuator is is a relay controlled shunt to ground on the output. None of the USB processor, the FPGA or the Bridge have any idea that it exists. So I have no idea how having the shunt resistor in the circuit or not can affect clicks.
Matt (from PS Audio) is working with the bridge software vendor on the bridge w/ DSD clicks.
Clicks during DSD files with a lower volume setting from any input but the Bridge aren’t a known problem.
Clicks between DSD tracks depends greatly on your source, they are better in Redcloud (at least for most sources) but I don’t have a universal patch for sources that stop sending data, or fart when they change sample rates or that go back to PCM garbage between DSD tracks.
Thank you, and you are right: the clicks for usb and i2s were with the attenuator on and the volume at 100. When I lower the volume to 90, the clicks are gone, with and without the attenuator. I guess I will use the attenuator and the volume at 90, rather than no attenuator and the volume around 60. According to one of your earlier posts, this gives a better S/N ratio, right?
Bridge 2 is another story : clicks at volumes below 100 (even down to 60, although very faint), with or without the attenuator.
Ted said: “Clicks between DSD tracks depends greatly on your source, they are better in Redcloud (at least for most sources) but I don’t have a universal patch for sources that stop sending data, or fart when they change sample rates or that go back to PCM garbage between DSD tracks.”
I agree that clicks between tracks is a source problem and not a Bridge II or DS Dac problem. I heard the same clicks on my Sonore Rendu / W4S Dac for four years as I do with my new DS Dac / Bridge II combination. I have gotten used to them and they don’t bother me. On gapless playback they provide a rather low key, non-irritating signal that the track changed.
I would rather they weren’t there, but I don’t want to spend the time trying to figure out what part of the “source” is the cause. Is it the NAS? The original flac or aiff files? The AUI software that I use for upsampling? I don’t have the desire to figure it out, I’d rather just listen to the music. Coming from the vinyl era, this is nothing.
I’ve mostly been doing other things. One of the ideas didn’t pan out. But my subconscious is ruminating on others. An interesting idea I’ve had since seems to work on paper, but does nothing in the FPGA - I’m still not sure if it’s because I have a bug in the FPGA or a bug in the simulation. All in all, about normal progress…
